Hagerman Technology WebLog 2008.  Comments?

Aug 18

Oh my, time flies.  Took some time off from development work.  Been thinking a
lot about new products.  Oh, what to do.  RMAF coming up, almost forgot.  Need 
to get some machines ready and shipped out.

Bought a piano.  A Yamaha Clavinova.  Its for the kids, so they can practice.
I should find a way to get a line level output so I can plug it into my real
amps.  The built-in digital amp sounds kinda thin and you get listening
fatigue pretty quick.  

Still looking for La Scalas.  Wanna hear them on my PA-10s.

Bought a PIC NIC, but now the computer isn't hanging anymore.  I think it's
really messing with my head.  

Aug 01

Ok, here are the scans.  I marked up the PA-10 design for use with 300B output 
and a WE437 driver.  I aimed for a 300B bias of 320V at 70mA and -60V.  The
driver stays at 175V, 20mA, and -3V.  Might take up to 2V input to reach full

To get there, B+ has to be +400V, it takes a 1000Vct power transformer at
160mA.  Sticking to AES catalog, the Hammond 282X will work.  This drives a
5u4G rectifier.  A single choke input will suffice and I use the 4H 159S.
Load capacitance has to be 400uF minimum.  With load a 450V rating will work,
however if you power up without output tubes, the B+ will fly unloaded up to
almost 600V.  So rate these caps high!  The filament trannys can be 166L5.  I
added 75 ohm resistors to anode of rectifier to tune down the LC input tank.
Basically, the ESR of the 159S is not high enough.  You can lower the 75 ohm
value by whatever the ESR is of the 500V secondary.  So maybe it cancels out
down to roughly zero, in which case you don't have to worry about it.  Very
important to get the power supply tuned right, or it ends up responding to
line and load, thereby imparting a sonic signature.

The WE437 is almost plug and play, changing the 160 ohm resistors to 150.  It
should bias well and easily drive the interstage tranny.  Best to use a high
quality 1:1:1 here.  The 124B works, but this is a good place to spend more
money.  Maybe a Tango or Audio Note?  The output tranny needs to handle 20W
and I like the 8k:8 ohm ratio.  The Hammond 1650F is close enough.  Anyway,
you should get a solid 12W to 15W out of this without feedback.  Ideal for
horns or Lowther type drivers.  Output impedance might be about 4 ohms on the
8 ohm tap.  

300B push/pull amplifier page 1
300B push/pull amplifier page 2

Jul 30

By accident I was on the dell.com site and noticed a user comment that said 
all of their 15 optiplex computers locked up because of the onboard network
cards.  Fixed when they put in a extra one.

Meanwhile, trying to solve a hum issue.  This is the first time I've been
unable to get rid of buzz.  Well heck!  It's because there is no earth ground
in this house.  Two prong plugs only.  Try running your phonostage on that.
So installing a reference earth wire from the water pipes.  Need a long drill
bit to get through the house.

Jul 29

So a customer, no actually just some guy in Europe, likes my PA-10 amplifier
concept and topology and wondered what it would take to change it to 300B 
output tubes.  He was going to pay me.  But I'm too busy at the moment to take
up a full-blown analysis, something I could charge for.  So instead, I thought
about it, used the WE437 driver tube he liked, and came up with a solution.  I
simply changed the values of trannys and Rs and Cs in order to scale up to a
300B.  Looks like it should work pretty easily.  I won't fit in the existing
chassis, but for P2P wiring it should be ok.  Into an 11:1 tranny it should
deliver 20 watts.  Or did I do that calculation wrong?  I didn't consider the
relatively high output impedance, maybe 4 ohms.  But it should be at least
double the PA-10.  We set the 300B bias to 320V at 70mA.  That requires a B+
of 400V.  The WE437 looks to be a fabulous tube, running at the same 20mA and
170V that the 6E5P did.  I'll draw it up tomorrow and scan in.  Power tranny
needs 500-0-500 on secondary at 160mA.  Choke is 4H.

Jul 19

I ran a full week without a computer problem.  Then it finally froze again.  I
was starting to thing the new ethernet cable did the trick.  Not.  Took me
about 8 or 9 reboots (over two days) to get far enough along to type this.

Anyway, someone reminded me that Greg already uses the SKA franchise, so it
would be wrong for me to use the SKP.  And SKD.  So I'll come up with some
other name.  

Jul 08

New ethernet cable is doing much better.  Have not crashed in 3 days.  I'm not
sure anymore, but I think it crashed even with the new cable.  But anyway,
stable enough to get work done.  Definately an issue with ethernet.  If not
cable, then hardware.  So I'm ok again.  If not satisfied with current
reliability I can just switch to a separate card.

Thinking maybe I can do a new series, add an SKL.  A simple killer linestage.
All of these would have that new low-cost form factor.  The idea is to push
prices back down to $1000.  That is the new sweet spot in the marketplace.  My
TEN series not fairing so well, as they got pretty expensive.  Too far upscale,
perhaps.  A marketing blunder?  Who can tell.  Combine the economic turndown
with a divorce and things change.  Mostly, if I can do a tube phono and a
chime-like tube DAC for $1000, they'll sell in today's marketplace.  Decided it
is easy enough to add a linestage to the mix.  After all, I still need a
linestage myself!  Using an HA-10 at the moment.  Works well with the PA-10s.  

Jul 05

Oh my, all of my computer troubles might be tracked down to a bad ethernet
cable.  We'll see.  I have a new one in there now.  So far so good.  You'd 
think I could have thought of this earlier?

Jul 04

Ok, the consensus is VMWARE for mac.  I'll buy one of them and try it out.


Had an idea this afternoon.  The SKD.  Simple Killer Dac.  Seems to me the
DA-10 is overpriced for today's market.  I pushed performance too much.  Maybe
if I scale back and focus on cost instead.  That's what I did with the SKP.  I
could use same chassis for doing a DAC.  Of course, I never finished the SKP
layout.  Been so busy with other stuff, and this damn computer flakiness.  
Tomorrow I will sketch out the SKD circuit.  Of course, it will use a HAGDAC as
the enginer - there ain't no better solution out there.  

Yep, today's economy came at the wrong time for my TEN series.  A little more
than most people want to spend.  The magic price point for a tubed phono or dac
is $1k and slightly over.  I aim for that.

Meanwhile, I've been playing the Frankenvandys on my PA-10s, and it seems a
little thin.  Not getting the punch until I crank it up.  Want to get that same
dynamics and slam at lower volumes.  Not happening.  Running off of the 4 ohm
taps helps a little, bass has better definition, but doesn't seem as loud.
Have not tried speaker placement, went only with the golden ratio dimensions.
Soundstage is fine, only that I feel the sound is a bit thin.  The good news is
that all of my TEN series equipment is performing without a glitch.  Thinking I
need to get into some horns.  

Jul 03

Yeah, maybe my computer crashing is network related?  Not USB.  I remember they
did not plug in ethernet at the geek shop.  Unfortunately, it is built into the
motherboard, so I'm not sure what my options are.  Got rid of everything else I
could in terms of software.  Still hanging regularly.  Not quite ready to pour
too much more money into it.  

But hey, here's the question I really want to ask.  What is the name of that 
Mac program that allows you to run PC (XP) software?  Is it an emulator?  Or
something like vmware, where I need to actually load a copy of XP onto the
machine?  Please let me know!  My oldest daughter can't run all her favorite
games on the iMac I just bought.

Jun 28

Crud.  Computer hung during reading email off the server.  Copy was not saved 
locally and is now wiped off server.  This is a first.  Anyone who sent me an
email or order this morning - it's gone!  I never saw it.  Not sure what to
try next.  Did the fan thing, no help.  I was pretty sure it was not thermal,
as most hangs occur cold.  Usually at the end of boot, when there is some
network activity going on.  I am running zonealarm firewall and the icon shows
I/O.  Machine does not stay alive long enough for me to type or pull up the
task manager, see what is processing.  Sometimes the machine runs for hours.

My next step might be just to buy a new box.  Move the drive.  Maybe the 
problem moves with it?  Or maybe monitor the power line?  I have other PCS on
the same circuit, nothing else hangs.  Hate this.  I just keep throwing money
at it.  Must have this machine running.  It is the only way to charge credit
cards, do invoices, print manuals, make schematics and layouts.  Really do not
want to build a new machine or transfer over to the imac.  Driving me nuts.

Jun 24

Still working computer issue.  Took it back to shop, runs fine there.  Only
seems to hang at my place.  Maybe I can use the Dell warranty.  It's only 3
months old.  Just can't run a business when your computer (with everything on
it) keeps hanging at random.  I don't even know if I can finish this sentence.
Ok, I did.  I'm almost wondering if it is heat related, hangs on USB driver
only when hot.  Wonder if there is a way to kick in the internal fans to a
faster setting.

Meanwhile, I have a production FRYDADDY 80% complete, just sitting there
waiting for me.

Jun 21

PC is now more stable, but still hanging.  Can't find anything else to play
with on the inside.  Using new keyboard and mouse, new printer cable.  Not
sure what to try next.  I did delete all printer drivers, then re-install the
one I need.  No change.  Had the cover off the PC for awhile, no change.

But I have good news.  Discovered there is more money in my bank account than I
thought (bad bookeeping?).  So I spent it on that iMac I've been dying to get 
for about a year now.  It will be turned into my music server.  I can use it
for demos and shows.  Beautiful machine.  Got it home, fired up ok, did the
username and password thing.  Got onto safari.  Then I noticed a dead pixel.
Crud!  One black pixel in the middle of a white screen.  Gotta take it back.
Hey, for this kinda money I expect better.  We'll see how they are about
customer service.  The original buying experience was superb.  Easy, fast.
Had the machine in about 5 minutes, walked out of store with one box.


Another tip.  Hated my old file cabinet, so I just tossed it.  Was planning on
getting a nice wooden one, but then I saw these Sterlite plastic crates (look
like for milk cartons) that stack, yet have hanging file folder rails.  $6 a 
pop and they work fantastic.  I just load 'em and stack 'em.  My mostly used
files on top.  Super cheap and way better than that old steel hulk.  

Jun 20

Ok, so far so good.  Opened up the PC, reseated everything.  Trying out a new
USB cable for the printer.  Possibly I had TWO problems.  Messed up USB
keyboard and a bad printer USB cable.  Both items seem to contribute to the

Meanwhile, Patent #7,388,196 was issued.

Jun 19

Sorry for no updates lately.  I am so hamstrung by this computer glitch.  If I
can get the machine to boot at all, I get somewhere up to an hour of normal
operation before it hangs again.  Very counterproductive.  It is all I can do 
to process orders and answer a few emails.  

Jun 16

Ok, computer finally back.  In the shop running diagnostics all weekend.  They
found no problem.  I plugged it in and it hung within seconds.  All I had 
plugged in was keyboard and mouse.  So ran out and bought new ones!  New mouse
didn't help.  New keyboard fixed it!  Gads.  I fried my hard drive and shut 
down the business for a week because of a flackey keyboard?  Cost me $450 by 
the time I was done with it.  Aye.

Meanwhile, building up the production FRYDADDY.  So far so good, but I can't
find where I put the power transformers.  Which box are they in?

Oh, and new post office box.  My new company address is:

PO Box 61911
Honolulu, HI 96839


Aw crud.  The computer hung again.  Same problem.  No idea what it can be.

Jun 12

Wow, what a long delay!  My business computer again suffered a hard drive
crash.  Really slows me down.  You don't wanna know what it cost me to fix.  I
still don't have the printer running yet.  My laptop covered me for most emails
at least.  But I am back.

FRYDADDY boards came in.  


Did another good race on Sunday.  Came in 3rd.  Really good run!  6:27 per mile

NS triathlon results


Crud, the computer hangs a lot.  Have to try several reboots to get it going
again.  Back to shop.  Ugh.

Jun 03

TRUMPET burned in, ready to go.  I can't believe sometimes how damn fabulous
this thing sounds.  The new version is so quiet.  With the STEPUP for a front
end, it is scary quiet.  I have to crank it sooo loud just to hear some hiss.

And then the FRYDADDY panels came in.  Will start putting one together.  But 
will wait for production boards.  All looks really good so far.  I'm happy with
the product.

Meanwhile, no progress on the SKP.  Need to get that layout complete.

May 30

Dang, postman did not bring my digikey parts.  Have to wait until tomorrow.
Trying to finish up that TRUMPET to send out for review.  Everything else done,
just needs the electrolytics.  Maybe I do it tomorrow night.  Gee, fun.

Meanwhile, making good progress on the SPK layout.  Very tight.  Parts on both
sides, overlapping everything.  Doing just the left channel for now.  It don't
leave much room for the power supply.  No idea where that B+ regulator is gonna
go.  Perhaps middle top.  Too many damn caps.  Already I have the cinemags
underneath and overlapping the input tube sockets and gain selector.  Small
standoffs maintain a 3D clearance.  Wires come out top and are short looped
down to holes in board.  Makes layout easy in that section.  One problem are
the large 1uF 400V caps.  The audiophile polyprop films.  No way the huge
electrocubes are gonna fit so I use the ppfmx from the TRUMPET.  Worked out
well before.  I have them mounted to bottom of board all in a row.  I need at
least 3 of these per channel.  The B+ and G2 decoupling for input stage, and
then the output coupling cap.  Output stage decoupling not as critical because
it is fed with CCS.  Super linear performance and super high headroom expected
in this design.  I think this stuff can be made to fit.  Another issue is
running the output signal trace down next to input tube.  As the design is 
noninverting, have to watch out for parasitic instability due to positive
feedback at HF.  Note I have almost all the ground returns down a path through
the center of the board.  That is, sort of single-point ground, all currents
return to roughly same spot.  I will split this up from the input, but overall
the concept tends to keep noise floor down, especially if the power transformer
can induce eddy currents in a ground plane.  If this design works, it will be
one very compact, very solid, and very heavy phonostage.

SKP layout

May 28

Ok, got a preliminary SKP layout done, trying to place the major components
(tubes, caps) to see if everything will fit.  Made some changes in construction
so power tranny will fit.  Luckily, the box is 4" deep, so there is plenty of
room.  The problem I see is that the power tranny is way too close to the input
tubes and input trannys.  Spacing is only 5" to tube and 6" to step-up tranny.
Have the coils oriented for minimal coupling, but from experience I have a 
feeling this ain't gonna fly.  It might be fine for MM.  I'd rather get MC
capability out of this product directly.  Maybe I'll play some more moving
input tubes farther away.  I could get another inch perhaps.

May 27

Wrote preliminary FRYDADDY manual today.  Uploaded.  Cleaning up shop.  Tossed
out a lot of old stuff.  Think I'll work on SKP layout tonight.

May 25

Took my kids to the Indy movie today.  They got scared and grossed out at all
the bugs and skeletons.  Oh well.  Otherwise not too bad.  They lived up to the
franchise hype.


Finished the SKP schematics.  Now ready to see if I can make a layout fit into
this tight box.  Really worried that it is too small.  Too much magnetic
interaction with tranny inputs or input wiring.  If it ends up being just an
MM stage I might not turn it into a product.  At least it is a nice design

SKP power
SKP phono

Hmmm.  Just realized I did not leave enough headroom in the regulator.  It can
only handle a 5% line drop.  I need to cover 10%.

May 24

Mostly taking the weekend off.  But I did get a few hours in late tonight after
we had gone out to dinner (gyros).  Put the SKP schematic into CAD.  Decided to
add a GAIN control.  Barely squeeze it in.  This selects the 3 primary taps of
the step-up tranny.  Three gains for three types of MC carts.  LO, MED, HI.
The MM version will have no gain control.  

May 23

Got another DA-10 out the door.  Discovered I forgot to order 450V 100uF caps
for the TRUMPET.  Dang.  It won't ship until next week.  One thing I really
like here is that I solder in all the parts on the DA-10.  Fire it up and
no problems at all.  It just works perfectly.  That means not only is the
design robust, but so is my construction.  Zero errors.  Turn it on and go.


Then I had this flash of an idea.  I don't know where it came from.  I call it
the SKP.  Simple Killer Phono.  For some reason, I was looking at a STEUP and
thinking what a nice little box that would make for a tube phono.  And way
cheaper than the expensive TEN series chassis.  But it would have to be bigger.
So I started thinking, maybe I can do a quickie version of the CA-10 phono
section, the new pentode input topology.  The schematic is done, maybe I just
implement part of it into a cost-reduced box.  Heck, make the whole design
cheaper, but without sacrificing performance.  Easier said than done.  So a
box the size of the LINE FILTER or bigger (I found one at 8.7" x 5.7") that
will fit all the parts.  So I take parts of the CORNET, parts from STEPUP 
and other older projects, come up with a very simple design.  Went in circles
for awhile and came up with a good starting point.  My goal is to put this
together for $999 retail.  Or maybe I do $999 for MM and $1499 for MC?

I had a 6X4 rectifier in there for awhile (because it is cool), but it made me
go with ac on the heaters.  I really didn't want to do that for phono.  So I
made the filamant supply just like a CORNET.  The B+ supply, in order to cut
ripple and reduce the number of big caps, I decided to insert a discrete linear
regulator.  An IRF830, two ZTX458, and one LM358.  All borrowed from other
projects.  It will provide a more stable B+ (regulated), plus cut a lot of
ripple.  Good DC supplies will reduce hum.  However, the one problem I face
here is the super close proximity between the power tranny, a 269JX, and the
tubes.  And the input trannys!  I think I have a smart arrangement where the
fields cancel.  Tube-wise, it is a pair of 6688 and one 6CG7.  That should give
great sonics and tons of gain.  Yes, a simple killer phono.  Looks like a big
PICCOLO.  Sounds like a TRUMPET.  It may not be beautiful, but at $1499 the MC
version will likely trounce just about anything out there.

It's a good idea.  Not many people pay for an assembled CORNET anymore.  They
want something new and better.  This is it.  A part CA-10.  Ready for RMAF? 
Best thing I can do now is come up with a layout next week.  Then build a proto
in June.


Ok, penciled a schematic, came up with values, workable circuit.  Simple little
feedback linear regulator.  Not the ultimate in performance, but rather a
simple and hopefully sonically benign stage that gets out of the way, merely
removing ripple.  Chenged it to use a 6.2V zener.  Heaters are rectified and
filtered DC, just like the CORNET & CLARINET.  No regulation.  But I think the
regulation on B+ might end up being beneficial.  The basic phono circuit is
from the CA-10 design.  Step-Up tranny into pentode stage, followed by a simple
passive RC EQ, then a triode output with a CCS plate load.  Should make for a
very linear circuit.  No feedback in the amplifier.  High gain phono using just
two tubes!  For stereo, it requires 3 tube envelopes.  Drew up a top panel 
tonight.  Overall size is a little under 6" x 9" x 4".  Trannys inside, tubes
stick out top.  Heavy.  Probably $1k for MM, $1.5k for MC.  Can be factory 
wired for gain and loading.

SKP panel

May 22

Got the FLUGELHORN shipped off to Harry Pearson.  We'll see what he comes up
with for a review.  Or at least his research into the different equalization
settings.  This could be fantastic exposure!  Is he not tops in the field?  I
took a chance and also sent him a new FRYBABY.  Maybe he'll use it to finish
the break-in the the phonostage.  Hey, one sentence in TAS about it's existence
is worth a lot more than what it cost to build.  


Got some pics back from the triathlon.  So excited about it, I might do another
one on June 8th.  We'll see.  I wasn't fully aero on the bike for this shot as
it was right after a speed bump.  Got into the bars about two seconds later.
And I know it looks like I am barely moving on the run, but that was the last
1/4 mile of the run and I had picked it up to a 6:30 pace.  Easily.  The photo
doesn't do me justice.  I blew by that guy in front of me.

Bike photo
Run photo


Made more changes to FLUGELHORN boards.  These are now ready for production.
Same as the one HP got, but with some minor manufactuability improvements.
Extra wire solder holes, mounting holes in the corners for easier assembly
(fits into jig), I added LEDs so they connect direct to front panel like the
PICCOLO does, fixed up placement of EQ caps. 


May 20

Added BLANK ADAPTER card to the HAGLABs order page.  They will be $39 each 
(with connector).  Working on a bio page.  Sort of a few paragraphs about the
designer and his audio philosophy.  


Finally fired up my pair of PA-10s.  One of them has been sitting all built for
about 3 weeks.  Now they are both operational.  Last measurement needed was SNR
which came out to 85dBA below 1 watt.  Or 0.15mV.  That's quite good, perhaps
ten times better than what most people get out of a DHT.  Excellent.  Now they
go into my listening room, which is being built.  

Auctioning off the FRYDADDY proto.  Figured it came out quite good, and it is
fully operational.  So rather than pull it apart for scraps, I'll sell it!

FRYDADDY auction

May 19

Ok, got the FRYDADDY design completed.  Tweaked on circuit, made changes to 
both layout and top panels.  Had to move the mounting screw holes 1/4" away 
from the RCA jacks.  This allows me to use the existing circuit boards.  Also
placing all red connectors (hot) on upper half, black on lower.  This will make
more sense to the user.  Anyway, bought parts, ordered production boards, and
bought production panels.  Oops, forgot the darn cases.  Anyway, added this
product to the order form.  This product is now officially for sale.  Lead
time for first units is 4 weeks.  After that, I'll speed it up to maybe 2 weeks
by always keeping panels in stock.  This is going to be a killer product!  
However, when putting an accurate parts list together, the factory-direct 
retail pricing came out to $649.  So this is $50 more than I originally
estimated.  That's ok, nothing else out there compares.

Need to get me a better photo, then do product manual.  But hey, this is a big
project finish.  Nice when you finally order production parts and can take real
orders.  That is such an incredible milestone.  


Just spent several hours updating the HAGLABS product pages with testimonials
and sections cut out from the professional reviews (sometimes I think the
brevity makes for a better and more convincing presentation).  Sort of cheated
on the DA-10, as I took the customer testimonials from the Chime.  Ok, same
HAGDAC, similar concept.  I'm thinking if the Chime was that good, then they'll
go nuts over the DA-10, which is clearly better.  Better power supply, better
tube gain section, better components, and better chassis.

May 18

Had me a good race.  All of the really good athletes did the long course today,
so I placed high in the short course.  Did well, a good tactical race from
start to finish, but with the races mixed together there was no way to know what 
you place was.  I ended up winning my sge group, getting 12th overall.  They
gave me the award for 3rd place "masters" instead.  I guess that is anyone over
40.  So a good day.

Honolulu Sprint Triathlon Results


And then I carved up the covers for a TRUMPET.  Maybe I will paint them

May 16

Got the rest of the FRYDADDY up and running.  Some trouble with amplifier 
stability and low frequency feedback through power supply rails.  Ok now.  The
transformer output works spectacularly.  Fabulous solution.  Output voltage is
not so great on RMS basis, but I am getting 20V peak-to-peak.  That's about 
five or ten times a FRYBABY.  Big powerful signal with a lot more current
horsepower behind it.  And perfectly balanced, too.  Never a danger of failure
causing a dc output voltage.  Checked the attenuator in SPICE, yeah, I got it 
right.  Will add a small cap to input of power amplifier to cut some of the
bandwidth beyond 150kHz.  Ok for cables, but too much for phono.  Carved up the
corners for proper radius and everything fit perfectly, just as I had measured. 
Even the tall binding posts fit with nothing to spare.  Hey, as long as the lid
closes!  I have some changes to make.  The circuit changes are small, and I can
use the boards I have.  It is the panel.  I think it needs some engraving lines
or arrows to show how to connect a cable.  Something intuitively obvious.


Also made a nice photo of the LINEFILTER.  It is ready for sale now.


May 15

Borrowed some caps out of a RIPPER and put them in the FRYDADDY.  Came right up
but with some issues with the waveform generator.  Odd, as it is quite similar
to the FRYBABY.  But with supply voltage changes and such, I had to tweak 
around some resistor values to pull things into play.  Then I optimized some
more to get exactly the waveform desired.  So that part of circuit is fine now.
Having a little trouble with the attenuator.  Maybe.  Could be ok, but it looks
like by the time I push it through the power amplifier the levels for phono
equalized outputs will be too high.  Hard to tell, as there is so much high
frequency content, that it appears 20dB higher than it actually is on an
oscilloscope.  Maybe I should do some SPICE simulations at 1kHz to double
check.  Tomorrow I'll get the power amplifier installed.  See if there is any
smoke.  Then check the transformer output.  The top panel came in today, so if
all goes well I can button it up and take photos.  And they'll be ready for
sale!  This is going to be a beautiful product.

Also got in a new LINE FILTER panel.  One without the power switch.  This will
be the final design.  Two more panel orders in limbo.  The FLUGELHORN for HP, a
couple of DA-10s, and then a pair of TRUMPETS.  Hopefully I can get all of 
these machines completed by the end of the month (all depends on when parts get

Finished cranking through about 50 HAGUSB boxes.  I hate doing this.  Have to
run each one through the CNC machine, then add holes to the other panel, pull
it apart and clean it (plastic dust shavings everywhere), put it back together
and into storage.  Takes long time to get through a batch.  Also carved up all
the old FRYBABY cases I had.  Made up a batch of HAGUSB XLR boxes.  Should be
set just in case I have to take down the CNC machine for awhile.  


Started putting the FRYDADDY panel together.  Oops, wrong corner radius, does
not fit into case.  Also, might have a problem with the binding posts being too
tall, top won't shut.  Damn, that's a big on.  We'll see once I carve corners.
Also discovered the teflon insulators for the RCA jacks are too big to allow
for the #4 standoffs and screw between them.  I use the standoff to help mount
the board.  Maybe they are not needed.  Anyway, here's what it looks like at 
the moment.


May 14

Spent most of the day pounding through manuals.  Yep, I did all of the HAL
manuals.  That's ten or eleven of them.  They are simple, but still take a lot
of time to generate.  Hope you like the look.  Uploading them all right now.

May 13

Dang, I can't fire up thie FRYDADDY because I don't have the 4700uF caps for
the power supply section.  Swear I had a pair sitting around.  Can't find them.
So now it sits until I put in yet another digikey order.

May 12

Getting a LOT of interest in the DA-10.  However, some folks want the old SPDIF
input.  Therefore, I will make it an option.  You can have either.  For the 
same price you can buy the SPDIF version.  I can then upgrade it to USB at a
later date.  What I don't want to do is have a select switch and toggle between
two inputs, as that will mess up the 75 ohm termination.  This is a top-shelf
DAC and cannot make that compromise in performance.  Yeah, it is only $2299, 
but you need to compare it to anything, including a $20k meitner.  The DA-10
ended up being a pretty much no-holds-barred high performance design.  Anyway,
I can see I need to get more parts on order, as orders are starting to

Got a call from SonicFlare today.  They want to do a review on the new Trumpet.
Ok, I guess I can spare one.  Reviews seem to be one of the best ways to do


Getting excited about my first (and probably only) triathlon race of the year.
It is next Sunday.  I am doing the sprint distance at the Honolulu Triathlon.
Great course.  Tons of people.  Super fun.  Yes, I do like doing this.  And I 
go all out, pushing as hard as I can.  A little bit of a competitive streak.
My biking and running lately are getting to a good level of fitness, and my
speed/endurance is up.  Could be a good race and I think a good possibility to

May 09

Added a new option for the HAGTECH FRYBABY kit/2.  For an extra $10, I'll add a
plastic case.  Normally these cost $13.  I bought too many, and so will sell at
discount to get rid of them.  I will even drill the holes for you, so they are
ready to go.  These are the translucent grey ones.  Assembled FRYBABYs are now
red and sold on the HAGLABS website.  It just kinda makes sense to do that,
group all of the burn-in stuff together.  Soon I will be adding the FRYDADDY to
the arsenal.  


FRYDADDY boards just came in.  Will start building it up tomorrow.

May 08

I've resumed my google adwords campaign.  Turning if off and then on again to
try and measure its effectiveness.  Hey, at $5 a day, maybe it brings in some
good sales.  A lot of people who find me do it from searches.  So I try and get
as high as possible in the engines, then suppliment with the sponsored ads on
the side.  

Meanwhile, just auctioned off some units on Audiogon.  Hopefully that catches a
few eyeballs.  Trying to get the news out regarding the new HAL brand.  One of
my blog readers suggested something very interesting.  Well, I'll post a link 
here and let you figure it out.  Personally, I think it is very clever.  And I
am starting to figure out how to do such a thing myself.  My first thought is
to write a series of papers on how to design a tube usb dac.  I would use the
awesome DA-10 as my reference platform.  Basically, I would give away my design
secrets.  Maybe I could get these published on 6moons or ETM or PFO?

Long marketing video on white papers

May 07

Finally starting to make some noise out there.  These press releases I think
can be very useful.  I'm wondering where else to put my advertising dollars.
Renewed my audio asylum sponsorship last night.  I like that place.  Not been
spending too much time on forums lately.  

ETM news

What other websites are popular?  Banners are expensive.  Maybe a magazine ad?
What rag would be most effective?  I'm thinking of an add in the back pages of
stereophile, where a little 1" by 1" add is kinda hidden.  But they are cheap.
Does anyone look there?  Really, all I have to do is get the word out about
HAGLABS, get potential customers to the website.  Maybe I just need some more
reviews.  What product should I have reviewed next?  I'm thinking maybe the HAL
PICCOLO.  I can spare one of them.  So who would want to review an active MC
headamp?  6moons, ETM, Dagogo?  It is really quite superb with the new
capacitors installed.  I can't imagine the Elevator at twice the price can come
close to touching it. 

Indeed, I am finally starting to think logical strategy again.


Took some more photos.  Did the HA-10 and DA-10.  But I forgot to measure the
noise floor of this DA-10, which seemed to be quiter than my personal unit.  So
something came out right.  Sounded great with the new caps.  My most
analog-like converter yet.  

DA-10 photos

May 06

Got a note from HP.  He's interested in borrowing a FLUGELHORN to do some LP
EQ research.  Ok by me.  I'll get one built right away.


I finally got to making the new manuals for HAGLABS.  Started with the FRYBABY,
as it was the easiest.  I am discontinuing the HAGTECH version, replacing it
with the red HAL one.  Same machine, but without all of the adapter connectors.
I ran into a problem buying the binding posts.  Crap.  So can no longer offer
the full kit.  So I'll let AA and MD run out of stock.  It was a nice attempt
to break into catalog sales, but too many variables in the marketing strategy
that real life got in the way.  There is no point selling a machine at cost!
And that is what I did selling it OEM.  Anyway, this move thus brings all of
the burn-in devices under the HAGLABS banner.  Assembled, anyway.  I will 
continue to offer the FRYBABY kit/2 at HAGTECH.  Here is the new manual style
and format.

FRYBABY manual

May 05

So it has been about 8 years now I've made my living solely from HAGTECH.  Not
quite the big 10.  But was reflecting a bit on how many products I've come up
with in this time.  Certainly, it has been by far the most productive time of 
my life.  Tried to make a list here:

07) OBOE*
08) UFO
22) HAL DA-10
23) HAL PA-10

*  discontinued
** new

I also had a few projects that never saw the light of day.

01) HUSH, an ultrasonic sinewave balanced drive for DHT heaters.
02) DACPRO, a low cost opamp alternative to the CHIME.
03) JIMDAC, a TDA1543 version of the HAGDAC.
04) MOTORPOD, a variable speed motor controller for Galibier, with built-in
recharger for batteries and a strobe.
05) COMPRESSOR, a microphone preamp and opto-compressor

And then there were several attempts at getting DoD grant money (SBIR) for a
novel approach I invented for real-time imaging SONAR.  I think I must have
written that up four times without success.  Then two more proposals I can't
even remember.

Let's not forget the consulting projects on the side.  In the early days of
HAGTECH before things really got going, I supplemented my income by doing odd
projects for local companies and the university.  Mostly electronics, but some
were software intensive, with C code, Z80 code, and even Windows GUI front

01) RFMS, four generations of PC104 electronics to control, drive, and read the
mass spectrometer.  With CPU, power supplies, electrometer, frequency
synthesizer, and RF output power amplifiers.  With software.
02) BITCANNON, a lasercomm tx and rx board with adaptive optics beam steering 
and peltier thermal control of lasers.  With software.
03) CONTROLLER, a cpu and tx and rx motherboard for a microwave radiocomm link
running at 3Gbps.
04) NRZI, 3 versions of FPGA-based nrzi codecs for radiocomm data.
05) PRBS, an SFP paddle board with prbs generator for observing eye patterns.
06) FLYSPEC, a motherboard gluing together various pieces of an air sampling
instrument.  With software.
07) SEAMODEM, a compact and low power dual-band FSK modem for use on a thousand
mile long underwater cable.  With auxiliary AM voice channel.  

Last and not least, there are my three patent applications surrounding the 
novelties of the Hagerman 3HMS Spectrometer.  One, so far has been approved.  

01) Hyperbolic Horn Helical Mass Spectrometer.
02) Horn Mass Spectrometer having Fan Deflectors.
03) Horn Mass Spectrometer using Digital Deflection Drive.

So to recap 2000 to 2008:

01) 26 original new products (13 still available to DIYer).
02) 11 redesigns.
03) 5 cancelled projects.
04) 6 SBIR proposals.
05) 10 major consulting projects.
06) 3 patents.


Just spoke to Thom of Galibier.  He wants me to design for him a new turntable
motor control unit.  I had done one previously, but it was for a dc motor.  Now
he is looking at ac motors.  Three-phase to be exact.  These are very well
balanced and perhaps the least noisy of all the types.  No brushes, not so much
torque cogging.  The controller / supply would create the three ac phases 120
degrees apart.  And I was thinking, heck one of the phases alone could power a
Rega or similar table.  So this product could be an upgrade that applies across
the board.  Plus, folks have been asking me for this for year, albiet in kit
form.  And so I will start to mull it over, think of how to do such a thing,
what the controls might be.  My first thoughts are to make a HAL box just like
the LINEFILTER.  Have a 120V/240V input switch, ac power cord.  Then a 50/60Hz
select switch (although I might have this done automatically), and then an 11
position rotary switch for speed trim.  Say have 60.0Hz nominal, and then
increment by 0.1Hz steps, plus or minus 5 steps.  Simple, easy.  Any thoughts?
Email me.


Just noticed that 6moons finally got my press release out regarding the new



And then, for days I've been trying to figure out why sales started slumping a
little after mid April.  Duh!  Taxes.  Sheesh.  I think a lot of folks are 
waiting for a refund.  Things should pick up soon.

May 04

Got the PA-10 wired up.  Took some photos.  Ready to fire this thing up!  The 
mate is also almost finished, so I'll have a working pair this week.  

PA-10 front view
PA-10 back view
PA-10 amplifier side
PA-10 power suplpy side

I must say, this amplifier came out beautifully.  Super compact, everything 
tight with short signal paths.


Ok, trying to do what I can to drop prices again.  Putting the TRUMPET down to
$2699.  The PA-10 and DA-10 down to $2299.  That's about the best I can do, 
considering the amount of labor involved.  Just doing my part to try and
increase overall sales during the slow summer months.

May 02

I turned 48 today.  Been doing some financial analysis, trying to get a better
indicator of performance.  Looking at various things.  Hard to spot trends, as
so much stuff is variable.  My weekly income is up and down, very spikey.  It
just happens that way.  But I did come up with something that might be useful.
I plotted the average number of sale by day of the week.  Seems that on
Tuesdays, Wednesdays, and Fridays I get my most orders.  I wonder why that is.

Sales per day

Actually, I didn't do any real work today.  Just reconciled the checkbook, did
the financial analysis.  My kids had their big Lei Day things at school.  

This weekend I plan on finishing up the PA-10s.  Get some photos, some specs.
And by golly, let's give them a listen in stereo!  Oh wait.  Not enough 2A3
tubes in the shop.  Have some on order.  Dang, I am always ordering stuff.

May 01

Finished the FRYDADDY layout.  Finally.  Got it done this morning.  Came out
pretty good.  Luckily there was a lot of room on the board to get things to
fit, easy to route the long traces.  Hope I didn't make any mistakes, would 
like to start shipping this product soon.  Now have a minimum of ordering 4
proto boards.  Get one free.  So I have 5.  Would save me some bux if I didn't
have to make any revisions.  So boards are ordered, so are is the front panel.
I already have most of the other parts (transformers, etc.).

FRYDADDY schematic


Had a thought.  In order to speed up manufacturing of HAGLABS machines, I think
I'm going to try and always keep one set of each panels in stock.  Not the
entire machine, but at least the panels.  Normally they have a 3 week lead time
and that causes too much delay.  I can get all the other parts in a few days.
So the idea is to build up a stock of panels.  Then, when I get a new order,
I get a replacement set, not the original.  This will reduce my backlog from
3-4 weeks down to 1-2.  Even for the big expensive machines.  Duh, why didn't
I think of this before?  Yeah, I end up with $1000 worth of panels sitting on
the shelf, but I think overall the efficiency improvement will be worth it.


Wondering what I should work on next.  Should it be the RA-10 or the CA-10?  
the former would be much faster.  The latter far more profitable.  But waaay
more development work.  On the other hand, the circuit I came up with is
absolutely fabulous.  At least the analog tube section.  Maybe I spend my 
summer doing that, since I doubt I will be going on vacation.

Apr 30

Hey, this month wasn't so bad after all.  Sales ended up quite good overall.  I
guess my worries about the economy might be a little overblown?  Anyway, let's
keep it up!

I am working again on advertising.  The simplest and easiest way is to auction 
a unit on audiogon.  Cheap.  Except there is the cost associated with not
selling something at full price.  So I take the hit there.  I always like to 
start the auctions out at $1.  Gets lot more people interested.  Trying to get
as much eyeball attention as possible.  Need to get word out about the new
HAGLABS stuff.  I put out some press releases the other day.  We'll see what
effect this all has.  I now have 4 auctions going at audiogon.


Oh, I ran into Stu today (unclestu52).  Sure was nice to visit for awhile.


Almost finished a PA-10 today.  Just need to wire up the power tranny.  The 
upper (amplifier) side of the board looks like this.  Very clean and orderly.
The bottom looks way better.

PA-10 amplifier guts

Apr 29

Ok, I'm having trouble buying some of the FRYBABY parts, the binding post
adapters.  They aren't selling much more, and they never did make me much money
so I am going to just not build anymore.  The DIY kit/2 is still available.
The new FRYDADDY will be the king of burn-in machines.  Somebody tried to order
one today.  I best get my butt in gear.

Made a new HAGDAC for the CHIME GOLD, trying to shotgun that random "tic".
Nope, still there.  So I've replaced both HAGUSB and HAGDAC cards.  Same issue.
I'll have to rip the machine apart.  Just to make sure, I then plugged the USB
output into my DA-10.  Whoah!  Yeah, no tics, but more importantly, what a
nicer sound!  The ambience and clarity just punches through the mix.  I wish I
knew exactly what did it, but the DA-10 is a good step up from the CHIME.  
Seems to me the DA-10 and HA-10 combination is pure magic.  Both best on planet

Apr 28

Thought I found a bad solder joint on the VCXO of that CHIME GOLD.  Fixed it,
gave it a try.  Nope.  Still gets a few tics in while the machine warms up.
Seemed to stop after awhile, so maybe it is something thermal.  I'll check
wiring and other parts tomorrow.  Other than that, the machine sounds great.
Will put it back on audiogon soon.

Dave Clark of PFO wants some HAGUSBs for review.  Oh yeah!  I still get a lot
of business from the 6moons review.  Turns out the HAGUSB is at the perfect
price point, delivers way better performance tham all the other low cost
solutions, and is a lot better than a $2k slimserver.  Hey, you order them,
I'll keep pumping them out.  Dave also wants to try out a DA-10.  

Apr 27

Finally got the PA-10 chassis painted up.  I like to let them dry for 2 days,
so that the paint isn't soft and picks up damage.  I glob it on thick to get
that nice glossy finish.  Must wait!  Do not touch yet.  Can't wait to get the
build complete.  These are going to look superb.  I already know they sound
good.  Looking forward to firing them up with a real system.

That CHIME GOLD I auctioned off a few months ago?  It came back.  Something
wrong with it whereas it glitches once every few songs.  I thought it was a 
bad HAGUSB in it, but replacing didn't quite solve it.  There is something
intermittent in there.  Anyway, I bought it back from the guy.  If it don't
work as advertised, I will stand behind my machines!  So tomorrow I will see
what else might be the trouble.  I'm thinking HAGDAC.  Will run a battery of
tests and then try to confirm I solve the problem.  Once it is fixed for sure,
I will auction it off again.  Yeah, same machine.  This time used for two

Also starting to box up and sort the magazine collections.  Maybe I will 
auction them off later.  

Then there are the menial tasks.  Take up time.  Like the front panel express
folks modified my HAL logo such that it is cheaper to produce.  So now I have
to copy it onto all of my panel designs.  That's like 20 or 30 panels!

Apr 25

Finally caught up with orders, putting the finishing touches on a VACUTRACE.
Thinking about the lab re-organization.  Going to sell off a bunch of books and
magazine (glass audio) I never use anymore.  Lighten up the load, make more 

Orders picking up again.  After reading Forbes magazine last night, they 
actually seemed a bit upbeat that things will be picking up, economically
speaking.  Ok, food and gas prices will remain high, but I think folks will
start buying some non-essentials again.  Like DIY audio.


Did some more layout work on the FRYDADDY.  Very busy here with a lot of
things.  Hard to get the time to get this proto up and running.

Ordered a ton more HAGUSB parts.  Also re-stocking on FRYBABY connectors and


Mostly, what I need to do this weekend is reconcile my bank account.  About one
year behind.

Apr 23

Hot and muggy here today.  Today I cut out the chassis for my pair of PA-10s.
Finally.  Was going to do the HA-10 and DA-10 but discovered I am short on
chassis.  Always ordering more parts.

Meanwhile, I see my page hits are averaging about 1700 per day.  So activity
is still pretty good.  I need to think about advertising and promotion again.
Been slacking a lot lately.  Must do them darn product manuals too.  But first,
I'm going to upload schematics to the website.  I wasn't going to do that, but
what will it hurt?  These are fabulous designs.  You should know whatyou are
buying.  Huge bang for the buck.  Sorry they take so long to build, though.
Try to have that done tomorrow.

Oh, did I mention I got a LINEFILTER wired up yesterday?  Works just fine.  I
will be removing the power switch.  No point for it.  Must run a listening
test, make sure all is well.  Right now I have the secondaries totally
floating.  I think I will reference them in balanced common mode to ground 
with a pair of 100k resistors.  Or maybe 10k.

Apr 22

Since I am custom building all of the HAGLABS stuff, I decided to add an option
for faceplate color.  Not everyone is enamored with the red/yellow scheme.  So
I will offer most products in silver/black, too.  But in return, I'll make the
return policy 10%, instead of 5%.


Apr 16

A dealer on the phone today gave me some good tips on how to improve your web
search engine ranking.  Basically, put big long detailed descriptions in the
page title.  Put all keywords there, that somebody might be searching for.  So
I edited all my pages today.

Meanwhile, it was a good production day, as I finally got two VACUTRACEs out 
the door.  Will try to make more progress on the FRYDADDY tonight.

Apr 15

I've gotten the FRYDADDY layout half done already.  Just inching my way along
when I get some free time.  Should be done later this week, I expect?  Will
then get boards and panel, build proto.  I think this is the burn-in generator
audiophiles are looking for.  Big, robust, powerful, and all in a nice box.  It
is what you guys wanted in the first place.  

And I am back on the PA-20.  Only I am going to call it the RA-10 (Reference
Amplifier).  Keep it in the TEN series properly.  I figure, the paper design is
done, it is mostly a modification to the PA-10.  Heck, it should not take long
at all to implement.  The CA-10 will have to wait.  I'll just pump out as much
stuff as I can right now, get the brand going.  Must try and fabricate sales 
out of this slumping economy.  Something to pay for the CA-10 development.  
Once it is done, I should be home-free.  It will be the killer product I always
needed.  So tonight I will upload some more website info.  CAD drawings and an
RA-10 page.

Apr 11

Ok, it is now official.  New price of FRYBABY is $299.  I was on the hairy edge
of breaking even (they're a lot of work to assemble), and used it mainly as an
advertising tool.  Small uptic in price to reflect my costs.  I have notified 
both Audio Advisor and Music Direct, and expect they'll update their websites
and catalogs asap.  The other change is that HAGTECH will no longer sell them
assembled.  I moved it over to HAGLABS, which will now produce only RED ones.



Ok, added FRYDADDY and CA-10 info to website.  The photos are just panel design
sketches, but give you a fair idea of what to expect.  

Apr 10

Crud!  Finishing up two VACUTRACEs, and discover I don't have any of the power 
trannys in stock.  Ugh.  Also missing the filament trannys for my DA-10.  Ok,
more parts on order.  UPS should get the stuff here Monday.  

Apr 09

Cranking on websites.  I added short descriptions to the product pages, added 
some photos.  Built a StepUp (see guts).  Then I re-did the HAGTECH pages,
added a big HAL banner to link up.  Took out obsolete references, etc.  Sort
of cleaned everything up, try to make it looks sensible.


I think I need to raise the price on FRYBABY to $299.  Starting to lose money
on it.  At the same time, maybe I can add a RED version to the HAL site.  It
will then set the stage for the FRYDADDY, to come out later (already half


Ha!  How about that?  I just renamed the PL-10 as a CA-10.  Why?  Because the
letters looked better on the web page.  Something about kerning.  It will be
known as the CONTROL AMPLIFIER.  Same machine, new name.

Apr 08

Building up new production HA-10, DA-10, and my pair of PA-10s.  Got halfway,
finding part shortages.  Glad I had already bought all of the iron.  Sheesh,
just building machines for myself costs a lot of money.  Mostly these are 
needed for a photo shoot.  Get the pics posted to the website.  Then I can
start to think about product manuals.  I have obsolete ones posted right now.
Also thinking of writing a description for each product.  Just having a photo
and specs seems somehow insufficient.  I'll get these parts on order.  Then
spend the rest of the week finishing a pair of VACUTRACEs.

Apr 07

Ok, finally getting some forward motion again here.  Completed that TRUMPET
over the weekend and shipped it out today.  Fits perfectly in that Box-7 from
the post office.  And I am using shredder remains for packaging.  It's a little
messy, but cheap and it works fantastic.  Fills every crevice and provides good
protection.  And I've been shredding a lot lately.  Grinding up all those old
records I don't really need any more.  Like HAGTECH orders from 2002.  I'll 
keep hardcopy backups for only a year now.  Oh, and that TRUMPET?  It really
did sound better than my tweaked CORNET.  More open, is the word I would use.

Meanwhile, going to get back to the website, fix the photos and links.  Adding
a features section for each product.  Must add the JUMPER card to VACUTRACE
page.  I'm also going to change the billing policy.  Forever I did not charge
credit cards until after product was built and ready to ship.  I'll keep doing
that for HAGTECH, but HAGLABS will be money up front.  As if you paid paypal.  
Each product is custom built to order.  Delivery can be anywhere up to 4 weeks,
typically 3 weeks.  Unless I happen to have certain critical parts in stock,
which I sometimes do.  So HAGLABS will charge your card up front now, then I
can more easily buy parts and proceed with custom contruction.  There will be a
5% cancellation fee.  Same deal as the 5% restocking fee.  

Running out of LSK389 again.  Must order more to keep those PICCOLOs shipping.
Oh yeah, with the differenc chassis connectors in the RED PICCOLO, I can't 
include the internal battery pack, so it will be AC powered only.  In return,
I'll add the 0dB gain setting.  


Ok, got some photos of the LINEFILTER.  Wondering why I put a switch on it.  
Not needed.


Apr 06

Ok, so I had a bad day yesterday.  Forgot to cut an etch, blew up and expensive
273DZ tranny.  Replaced it, found a better way to wire it in (forget the inline
connectors).  New TRUMPET with lots of modifications is now up and running.  
And then I discover I put in the wrong MUTE switches.  Shoot.  What a pain they
are to remove.  Anyway, once I got the right ones in, everything was perfect.
It all looks good, sounds fantastic.  I'll have photos posted soon.  Ok, so 
maybe I was a little harsh, thinking I could not handle this product.  Today I
feel better.  But I'll have to raise the price, as the amount of labor involved
cannot justify such a low price.  I don't have much choice.  The TRUMPET lives!
But at $2999.


Apr 04

Wow, listening to the new PICCOLO was an experience.  I can't believe how much
better these caps are.  Damn!  The red HAGLABS version sounds absolutely
stunning.  This is a hit, worth every extra penny.  I think I'll send one to
Thom to use at RMAF.  It is now clearly better than the tranny.  Not quite as
quiet, though.

Got the parts in today to finish the TRUMPET.  Maybe I can get it shipped out
tomorrow.  Meanwhile, the two VACUTRACE sit there unfinished due to lack of LED
bezels.  Man, I just can't keep up with inventory control.  

Apr 03

Excellent, my Electrocubes came in.  Just in time.  I can now build up the
DA-10, PICCOLOs, and other items that were on hold.  Grinding away here, 
filling backorders.  My apologies for all of the delays.

Apr 02

Shoot, I can't believe how long it takes me to build a new TRUMPET.  The
original was so well designed.  Everything went together a snap.  Hardly any
hand wiring.  Only trouble was all of the woodworking.  The new design requires
a lot of chassis work, machining and painting.  And wiring the boards into the
boxes is a pain!  New inline connectors, wiring the up the RCAs and XLRs.  The
original had everything solder righted to the board.  Great design.  Oh well, 
the new one looks and sounds better.  I wonder if I have to raise the price,
though, as so much extra labor is involved.  Pondering.  For now it stays the

Apr 01

I should have the first production TRUMPET done any day now.  All finished 
except for an inline connector I added between transformer and board, well, I
bought the wrong part.  Once I get the design tweaked, they will be easier and
faster to build.  I still expect a 3-4 week lead time.

Bought that vmware workstation, absolutely fabulous software!  Highly
recommended.  Yeah, it cost me $189, but it lets me run a fake Win NT operating
system inside a window within XP.  It is incredible.  My schematic program runs
great in this mode.  It acts like it is a completely separate computer.  This,
plus the dosbox software, I now have everything running on one machine.  And 
just in time.

Mar 19

The PA-10 and DA-10 boards came in.  Finally, I'll be able to have some
amplifiers to listen to.  And then I realized why I haven't finished the new
TRUMPET.  I scavenged the electrolytics for a CORNET and CHIME.  Damn.  That's
ok, I place one or two part orders per week.

Taxes are done.  Shipped off to accountant.  Computers up and running.  Now I
can focus on assembly and the new website.  I still think I'll have most of it
going by April Fool's.


Ok, got a PICCOLO partially built.  Waiting for the new expensive capacitors to
show up.  Good enough for a photo shoot, anyway.




Mar 18

And now I wonder about credit card debt.  Although it may be counter-intuitive,
paying off my credit cards may not make sense.  At the moment.  If the value of
the dollar continues to plummet, and I'm hedging it will, then the value of my
debt plummets with it.  Let's look at this from scratch.  If I buy $10k in gold
tomorrow with a credit card at 10% interest, what happens in a year?  Usually 
you lose.  But my guess is that next year gold will be double.  So a year from 
now I'll owe $11k in dollars, but own $20k worth of gold.  Basically, the value
of gold stayed constant, but my debt was roughly cut in half.  Am I nuts?  I'm
thinking there is no need right now to pay down my cards too fast.  Instead, I
should concentrate on owning assets.  In that respect, I bought sheetloads of
expensive audiophile caps, a bunch of PCM1704K converters, lots of leaded
solder, and as many circuit boards as I can store.  My next move should be to 
buy a pile of tubes.  Assets keep their value, the dollar won't.  The idea is 
to not own anything in dollars.  Unless they are negative!  Or perhaps I am a
fool, driving myself to ruin.  On the other hand, by next year people will be
desperate to sell off whatever assets they have left, just to eat.  So maybe
2009 will be a good year to find vacuum tube bargains.  

Could be I am spending too much time reading stuff on the web.  Like this guy
Weiss, who must be a genius.

Martin Weiss


Spent the morning doing taxes.  Every month I have to pay Hawaii excise tax on
all items sold (they even tax me on shipping charges!).  Fortunately, if item 
shipped out of state I can declare an exemption.  And that is 99% of my
business.  It just takes a lot of paperwork.  I'm still using my old QuickBooks
99 in double entry mode.  I keep customer invoices in a separate file from 
checking account balance.  Everything is entered twice, makes it real easy to 
go back and find glitches.  Ok, so I am about a year behind on reconcilliation.
Anyway, for federal taxes I made me a spreadsheet that adds up all my deduction
in various categories.  Part, postage, office expenses, advertising, travel,
etc.  Fortunately, it doesn't take me 8 hours anymore, as being organized can
help.  I just go through all checks written and credit card statements to find
my expenses.  I spent $4906 on shipping last year.  

Mar 17

Seems a few of the photos didn't quite make it.  I lost about 5% of the
collection.  Not too bad.

Working on this LINE-FILTER project, and I realized I messed up with the choice
of isolation transformer used in the LC-10.  Had a hard time finding the right
size to fit.  Well, I thought I had one good for 120W per outlet.  Turns out it
was half of that.  Crap.  Not enough juice.  I need to make sure each plug can
power a PA-10.  Or less.  Ok, I find a tranny to do the job, but it is 5"
diameter instead of 4".  Should be able to squeeze in the box.  Going to
increase price a bit.  That's four big trannys at $81 a pop.  Meanwhile, I have
four of the smaller ones sitting on a shelf.  Well, it turns out that it is 
just about right for the new LINE-FILTER.  I'll wire the secondaries in
parallel for 1 amp output current.  Both plugs are together and pump out 120W
total.  That's enough for a TRUMPET and something smaller.  Or a DA-10 and
HA-10.  Or two CORNETs.  It solves my problem for a low cost answer to 240V
operation.  And it is not just a voltage converter, it offers BALANCED ac line
outputs plus line filtering.  I should be able to offer it for $499 direct. 
Box is 7x5.  I've already bought the parts to build a prototype.  It is a
perfect match for a STEP-UP and looks something like this:


And while at the post office today I notice a sign that says fixed rate large
box.  Hmmm, that's new.  They've added a 12x12x8 box any weight for $12.95. 
Unbelievable.  That is the EXACT size I chose for the TEN series components.
Somebody is looking out for me.  The timing of this is perfect.  Not only does
it cut my shipping costs in half, I no longer have to buy boxes!  What a stroke
of luck.


Finished that other ARCHIVER, getting it ready for audiogon auction.  Burned
it in.  Did cap testing.  I realized that I was working with three different
types.  Blue, red, and pink.  The warm, earthy sounding ones were pink.  Not
my taste, as I thought they were too much, tending toward muddy.  The red and
blue have a different tone, more clear, more sparkle.  I think less smear.  The
blue ones are still my fave (but not by much), as they retain the microambience
and microdynamics the best.  So the ARCHIVER will retain the red, the FLUGEL
will use blue, and I'll stop using the pink.  The pink, by the way, are the
panasonic PPS dielectric.  

ARCHIVER in rack chassis


So what happens if the dollar's value is cut in half again?  Let's say oil is
$200, gold $2000.  It could happen in 2008, given the pace of where things are
headed at the moment.  It appears the Fed it entirely willing to sacrifice the
dollar in order to save banks.  Or something.  Did I bet on the wrong horse?
Let's take a look.  Three years ago I came up with the concept of HAGERMAN
AUDIO LABS.  The idea was to take my hard-earned reputation and move upscale to
a new line of products that would be sold dealer direct.  Why would I do this?
Because history was showing me that the best way to move product was to have
somebody else act as saleman.  Companies selling boxes through dealers were
making far more money than the one-website type of company.  Ok, I figured that
was the way to expand.  So I came up with the FRYBABY.  The whole idea behind
that product was to get it into the mail order catalogs and boost the HAGERMAN
name, get it recognized as a dealer product.  Well, my first choice was
Acoustic Sounds.  Chad wasn't interested.  Ok, Wayne at Audio Advisor saw
differently and took the chance.  That was followed by Music Direct.  Ok great!
Two giant audio catalogs.  The FRYBABY was off and running.  And I'm thinking
I have a foot in the door.  

The problem was money.  My idea for a giant preamp, fully tubed, balanced in to
out, was going to be a costly proposition.  How to fund development?  That's
where I got stuck.  Meanwhile, the economy was chugging along, ripe for such a
product.  It was a cool machine, a TRUMPET phonostage with multiple inputs,
step-up trannys, integrated with a balanced tube linestage.  All under remote
control using my new one-button concept.  Inside it had a built-in tape loop.
Except it was digital!  Yes, I had an ADC to USB and USB to DAC as a tape loop.
It would record any of the inputs (level controls built in), play them back, or
just be a line/phonostage.  It was a RIPPER deluxe.  XLRs everywhere.  It also 
had a FRYKLEANER built-in, capable not only of self-break-in, but it would also
burn-in any cables attached to it!  Yeah, it was a dream machine.  Then I added
it up.  Retail price would have been about $20,000.  Holy crap!  That is ten
times what a TRUMPET was going for.  How was I going to pull this off?  I had
the entire design down on paper, but couldn't afford to move forward.  The cost
of R&D was more debt than I could handle.  I had the endpoint, but no way to 
get there.

Then last year, after cleaning up my product line (ARCHIVER replaced BUGLE PRO,
etc.), somebody suggested again to make a tube headphone amp.  Ok, been on the
back burner for 5 years.  So I thought about it for awhile.  It had to be a
kit.  And things started falling into place.  For me, design is full of kharma,
metaphysical inspiration, feng shui, whatever.  Things happen.  Ideas come and
go for no reason.  I started moving in the right direction with choke input
power supplies, tons of iron throughout the design, all tubes, etc.  And then 
it hit me.  The chassis I ended up with would provide a perfect platform for a
HAGLABS series, but only if I changed to single-ended circuits.  And so I 
thought some more.  Sure enough, I could morph this new chassis into the "TEN"
series, leaving me the opening for a future "TWENTY" series (the XLR balanced
of original concept).  It was a solution that met me halfway.  All of a sudden
I could afford the development costs and build me a series of products still
destined for dealers.  And so I aimed for that.  Expensive front and rear
panels along with a classy paint job made the chassis beautiful enough to sit
on a dealer shelf next to just about anything.  My wife chose a unique and
identifiable color scheme.  Yes, yellow LEDs.  And realizing my production
would be limited, I could aim this line solely for US dealers, expanding into
Europe later.  Hence, I was able to compromise and use 120Vac only power
transformers.  Doing such allowed me to keep the trannys inside the super
compact chassis.  The designs were falling into place.  Solutions were

And now the bottom is falling out.  Will there be any dealers left in the US?
Sure, but what are the chances of any of them taking the risk of a new line?
That's my present conundrum.  Do I have a workable business plan?  Or is it the
right plan at the wrong time?  As a small company it is super easy for me to
adapt.  And fast.  I can change policy the next day.  Like recently, when I
changed to free shipping worldwide.  Bam!  Next day it was implemented.  So 
what does this mean for me now?  The question is, what should my new strategy 
be?  With a falling dollar, methinks I need to focus on exports for the near
future.  That's why I invented the new LINE-FILTER today.  It allows a TRUMPET
or CORNET to run on 240V.  I also recently switched the HERITAGE line of
products to factory-direct sales only.  That cut the price to super competitive
levels.  Moreso overseas.

So what am I worried about?  I'm worried about the TEN series.  Am I about to
waste a lot of time and money courting dealers when they're on the verge of
bankruptcy?  The dealers in good shape are the ones selling $100k speakers and
$20k amps.  They never had any interest in my stuff anyway.  So whom am I
chasing?  Did I totally mess up?  I could use some advice here.  Do I
completely change course here and switch the TEN series back to factory-direct?
If so, I shoot myself in the foot.  I screw up years of planning.  I set myself
back a long way.  My chances of getting the big magazine reviews are shot.  My
exposure in the marketplace is shot.  I have no sales channel (other than a
website).  Cutting out the dealers is like cutting off your hand.  Or am I 
saving myself with a brilliant move?  I have no idea.  I'm nothing more than a
circuit designer.  The bonus here is to the customer, who can then buy the
world-class DA-10 for a mere $2299 instead of $3760.  That's the difference.  
For me, the profit margin is the same.  

I consider this, as I find lately I've been really struggling to get boxes
built.  I'm realizing I don't have the production capacity I thought I might.
A hundred boxes a year is about it (a box equals a CORNET or DA-10).  If dealer
demand takes off, I am unable to produce without expanding and hiring a bunch
of assembly techs.  That will turn me into a CEO.  I will spend all my time
managing and no time designing.  Is that where I want to be?  No, in fact, I'd
like to someday do nothing more than design.  Outsource production.  Sort of
like what I think Tim de Paravicini or John Curl do.  No, I don't want to be
CEO or sales manager.  Right now, I can handle a hefty load of production and
assembly work.  That's fine.  But if I am limited to 50 or 100 boxes a year,  
then maybe I messed up thinking I should chase dealers?  Maybe I change HAGLABS
into a halfway measure.  Maybe it is nothing more than a vehicle to get me to
where I really want to be.  Build my reputation.  Build a customer base.  Pay
the bills.  Get me to a point where I can partner with the right people and
finally go big time.

Anyway, that's sort of what I am thinking today, as the price of bread climbs

Mar 16

Good, the file recovery program did it's job.  It's got a wierd interface, and
is a bit of a pain to use, but so what.  It worked!  I haven't checked yet, but
it appears to have recovered all of my lost photos.  I had to click on each one
to recover only them, and not the ten thousand temporary internet pics saved by
explorer.  Anyway, after many hours (and the entire Saturday, all the way up to
11pm), I have 1108 precious photos recovered taking up 749MB.  And I quickly
copied the folder to my new machine for backup.  This time, I think I will make
a few CDs.

Hey, I also have a completely rebuilt XP machine in pristine state, upgraded 
with all of the latest updates (I swear, they must issue one per day!).  The
kids may now play games.  And I can return to audio.

Speaking of which, I came up with a brilliant idea for a product.  All of my
fretting lately over the dollar, well it made me realize that this is a time to
focus on exports.  While my assets are loosing value (as is my debt!) I can
take advantage of the exchange rate, as now my products appear to Europeans and
others not only as bargains, but outright cheap.  The $119 HAGUSB is selling 
all over the world right now.  It is the right product at the right time at the
right price.  Anyone can cheaply turn their computer into a high quality CD or
music transport for less money than any other option.  A very simple upgrade to
your existing hardware.  Anyway, I realize I have to do the same for the
HAGLABS products.  Last year I made a (fatal?) compromise in that I chose to go
with 120Vac power transformers.  It was my only solution to fit inside the
chassis and keep costs down.  So I exchanged the ability to run at 240V for 
tranny inside.  Was that the right move?  In hindsight, maybe not.  However, 
all is not lost!  I realized last night all I have to do is make a smaller and
cheaper version of the LC-10.  I can do the same thing in a STEPUP-like box.
Duh.  Way cheaper.  A single outlet, one big isolation tranny inside (instead 
of four).  Basically, 1/4 LC-10.  Do I call is LC-2.5?  Nah.  Maybe something
like LF-10?  It will offer a solution for dropping from 240V to 120V, while 
adding the benefits of filtering and isolation.  I'll put it into the HERITAGE
line to keep cost under $500 (we'll see).  That way, a bloke in England can buy
a full-blown TRUMPET package less than L2000!  That's a TRUMPET, a STEPUP, and
a LINEBOX, shipped to your door.  Ok?  A stereophile class A tube phonostage 
with MC step up tranny and the ability to run at 240V.  L2000.  Yeah, this is a
backup plan to cover for and expected sales downturn on the domestic side.  

Mar 15

I got another ARCHIVER built.  This is the one without ADC, put into a spare
rack chassis.  Looks beautiful.  This is nicer than most machines you see at
the store.  I'll do some listening tests to compare with my FLUGELHORN to 
determine the best coupling capacitor for the application.  So far, the blue
ones are winning out over the red.  

ARCHIVER in rack chassis

Meanwhile, I think I'll hold off on my ebay purchases.  No point paying $300 
for a 25 year old calculator.  Heck, in a few months there will be a lot more
sellers and a lot fewer buyers, driving prices down.  I think this summer will
be a good time for ebay pickings.


Trying to rebuild my old XP machine.  Ugh.  I've got old install CDs all over 
the place.  Have probably 5 Dell machines, 1 IBM laptop, and I'm on my 3rd
Toshiba laptop.  The oldest Dell I rebuilt years ago into the NT machine.  Its
been by far the most robust.  Odd, when there is a power line glitch, all of
the other machines reboot, the old NT beast keeps right on ticking.  Inside, it
is a dustbowl.  Anyway, I had two hard drives in the XP machine.  One had all
my data files, the other the operating systems (multi-boot).  Decided to
sacrifice the data drive (had copied everything to my new machine already) and
ran the Maxtor reformatter and made a new partition.  Well shoot, my XP disc
that came with the machine would not boot.  The XP disc that came with my
wife's machine would not boot.  I tried a DOS diskette and it would not see C:
drive.  Nor D:.  Ok, plan C.  So I found my old NT install floppys.  After a
couple of hours, I got an NT install working.  Sadly, it could not use the
ethernet card nor the video.  Still, limped along.  When I tried to run the XP
install discs, they said I must have service pack 5 to continue.  Crap!  So
many dead ends.  No net.  So I tried installing drivers for the card.  Nope.
Only works with XP.  Sheet, how do I upgrade to SP5 so I can load XP so I can
use the network?  Ok, tried pulling the net card from the NT machine, as I have
the drivers.  But for some reason it remains undetected.  All is lost.  I'm
thinking I may have to go to Costco and buy a new XP.  One more try, I stick 
the XP install disc I have (SP1) and now it boots!  So running it.  Wiped out
the NT partition I just spent four hours building.  Well, it seems to be

I was thinking, just before I lost all my photos and crashed the system, I
stupidly ran that Registry Mechanic software again.  Thinking it might clear up
the machine a little before I handed it over to the kids.  Well, that was the
last thing I did.  On reboot is when it failed for good.  So whatever you do,
NEVER run that software!  I might have to quit the audio business and travel
the world warning people of the danger.  I still have the hard drive.  Will
stick it into the rebuilt XP machine as a slave.  Hopefully tonight I can run
that pareto recovery software I bought.

Mar 14

I am so impressed with this DOSBOX software, must repeat, it is phenomenal. 


Been trying to catch up on internet news.  Everything is pointing to economic
depression, not recession.  I have to wonder what it bodes for our industry.
Will brick and mortar stores be able to ride it out?  What is the best way to
position oneself to profit or survive such a thing?  And why is it the Fed will
bail out bear stearns but not audio research?

The US dollar

Hey, I'm just a circuit designer.  But one thing I have learned, is that the
price of gold is constant.  It is the dollar that is changing.  

I wonder if I need to turn off my google adsense campaign.  Am I just throwing
money away?


My production DA-10 and PA-10 boards just shipped.  I need to buy me a whole
new couple of matching sets of magnetics for the amps.  I have the panels and
chassis already.  Meanwhile built up a STEPUP.  No guts, but I needed this
photo for the website.

STEPUP photo

Also got the OEM phonostage complete.  This is a pre-production model.  I still
have to adjust panel dimensions, but this is what I expect it to look like
(without the AA badge).  On first listen, I the caps I used are not as good as
the blue ones I put into the FLUGELHORN proto.  These PPS dielectric are a bit
muddy and too warm.  I like the more open clarity and lower smear of the blue
EPCOS caps.  

OEM phono

Mar 13

Ok, some suggestions to try out vmware.  A virtual os machine within a machine.
I also found virtual pc 2007 (yeah, but it is from microsoft) and one called
dosbox, which looks very promising.  Honestly, I would love to be able to run
all of my software on this one computer.  At the moment, the old PC is unable
to boot XP, but it will boot to DOS and run my layout code.  This is ok, but I
have to get files in/out using a floppy disk.  Not so happy.  The NT machine is
continuing flawlessly.  Maybe because it never had to deal with a kid's game.
I'm pretty sure that is what finally broke down the XP machine.  So I am up and
running, albeit not with perfect convenience.  And my palm keeps hitting the
ctrl key!  I have made new backups, just in case.  The USB backup drive is now
running again, but had to be reformatted.  That leaves my old photos missing.
My last shot at them is plugging the disk into my wife's old XP machine.  Will
have to attempt when she is not looking.  

Meanwhile, I will take a break from this computer nonsense and focus on
soldering.  A few more machines need to be built.  Oh, you might also notice 
that the www.haglabs.com website is slowly growing.  I added a couple of 
placeholders for the other pages.  I need to work out how I want the individual
product pages should look.  I might borrow a bit from the HAGTECH stuff.  Seems
to me having the price and specifications conveniently shown is a good thing.
I can also start taking orders.  


This new keyboard is really starting to piss me off.  My left hand naturally
lays in a particular spot for comfortable typing.  Sadly, they put a goddamn
ctrl key there!  So stupid!  I keep screwing up everything I type with ctrl
characters, jumping off to unkown menus, etc.  I just was typing a long asylum
post and on my last sentence, bam!  It all disappeared.  Dell, you owe me a
new keyboard.  This one sucks.


Nevermind.  I ripped the key off.  Everything ok now.

Ok, after looking around, I think I'll use the DOSBOX for my layout software.
That will be my first test.  Actually, that's all I really need to do at this
point.  Eventually, I can then use vmware (looks to be the best supported) to
load up a virtual NT machine in a window.  I like the entire virtual concept.
Very clever, very useful.  I will keep you posted.


Oh my, this was unbelievably easy.  I downloaded the free dosbox program.  Hey,
only 1.5MB.  This is good old fashioned efficient code.  The way it used to be
done.  Loaded nicely.  I didn't even have to load dos.  Just my program.  Easy
to follow a few simple instructions.  And it all worked!  This mischevious and
nasty dos program fired up and ran without a hitch.  The only issue is that it
runs in 640x480 mode.  So it is small in a window.  When I run fullscreen (what
I had been doing for years) it gets lopsided, as I have one of those newfangled
widescreen LCD displays.  Not sure yet how to get it to maintain aspect ratio.
But hey, I am duly impressed.  This is awesome code.  Highly recommended.

Mar 12

I was doing pretty good with the data file transfers, but did one thing out of
order.  I erased a faulty account.  Last year when I ran that registry mechanic
it wiped out a whole bunch of working stuff.  My emails, etc.  I found the data
bases and copied them, sort of had things running.  But I forgot all of my
photos downloaded over the past 3 or 4 years was in a Kodak directory buried in
there.  Crap, I didn't get the photos out in time.  About a thousand pics gone
forever!  All of my yellowstone, england, alaska trip shots all erased.
Nothing went to the recycle bin.  Ok, I had a backup on a 160GB USB drive.  
Except it ain't workin now.  It loads, but doesn't mount up with a drive number
like it used to.  Won't load on any machine.  I think it lost it's partition.  
I hadn't used it in awhile, maybe this was why?  So my backup is dead too.  
Jeepers.  This is like the house burnt down.  Shoot, and it is just because I 
did this out of order.  Had my chance to copy the photos.  Missed it.


Ok, don't panic.  The data is sill probably there.  Just need a way to find it.
Remembered he recent issue of PC Mag has a list of the best 150 free programs. 
I seem to remember something about JPG recovery.  So did some searching.  Well,
there must be a hundred different companies making software to do this.  Most
involve flash cards that fit into cameras.  I just need to scan my hard drive
for data that looks like JPG file.  So I went with one of the paid for google
ads, got me that pareto logic data recovery pro.  Well, sure enough, it was
able to locate 22,800 residual images on the drive.  Took me two hours to go
through the list, one by one, finding the original file names.  Also to see if
there was recoverable data.  The good news is that almost all of them are still
there.  I bought the $49 key to enable recovery.  We shall see.  It appears to
be hung right now.  I have 529MB of photos supposedly being recovered.  I have
a feeling it ain't gonna work.


Yup, totally hung.  Now it won't reboot.  After much futzing, it's the registry
that has been corrupted.  Will not start XP.  Can't put drive into new machine
because cables are different.  Shux.  So now I have a completely dead computer.
A hard drive full of erased photos.  Hmmm.  Maybe I try sticking it into my
wife's XP machine?  I better get a new HD for this computer, as it is still a
good box.  My kids can use it for gaming.

And now I realize I have another issue.  I've been running two boxes with a KVM
swich (shares keyboard and mouse).  One box was NT to run the schematic
software.  I bounced back and forh.  The XP machine was dual boot, with a DOS
partition to run the layout code.  What a mess.  I need XP, NT, and DOS.  I'm a
little reluctant to try and make this new machine triple boot.  I need to find 
a place for the DOS code.  And mainly, my new keyboard and mouse are USB and
won't work on the KVM I had.  Hmmm.  Do I now run two sets of keyboards?  This
computer stuff is losing it's fun quotient.  Shoot.  I could run this new
machine as triple boot.  I already have he system commander software, just 
seemed to have lost the serial number.  Gads.


Ok, does anybody know how to run an NT-only program under XP?  How about DOS?

Mar 11

Well, I am typing this on the new computer.  Slowly transfering files and all
data and crap over.  Tons of programs to install.  Have you ever tried to find
the address book data?  Or your old emails?  Sheesh.  Anyway, I did.  Now I
need to move printer.  Theoretically, I got the hard stuff done already.  The
new computer is a Dell optiplex.  Big.  Looks really rugged.  I like the case.
The problem so far is the keyboard.  Where I normally rest my left palm is now
on top of the control key, so it's booching left and right.  I can't type
normally.  I may have to upgrade.  Or maybe I can just disable the key.


Ok, got image processing running again.  Here's some shots of the new CORNET.
I think it came out beautifully.  This thing is a stunner.  And quiter than the
original.  Check out the guts.  Expensive resistors and caps everywhere.  Very
compact little unit.


Mar 10

Interesting destinations for HAGUSB orders.  Today I sent one to a radio or TV
station in NYC.  Another to team CSC (the bike guys) in Austria.  You'd be
suprised at some of the folks who own one.  

Almost ready with some photos.  Got the FLUGELHORN done, almost there with the
CORNET and TRUMPET.  As soon as I can find some free time I'll get the SU and
PI done too.  That should give me enough photos to launch the website.  
Meanwhile, I keep playing with link colors and other such miniscule things.


Computer finally got here.  Missed the first two attempts at delivery.  And
speaking of which, during lab testing last week my laptop went wierd.  Maybe it
was a low battery warning, but it glitched and locked up.  Started up again ok,
but now the usb-audio driver doesn't talk to foobar.  This is a classic issue
I've been having with that machine.  A few times I did something to get it 
working again, and recently it was flawless.  Now it is totally messed up
again.  Just as bad, now the wireless software doesn't go looking for wifi.
Jeez.  When it rains it pours.  And this is the laptop I was depending upon for


Ok, got new CORNET done.  Took good photos of it and FLUGELHORN.  Just can't
get them uploaded without bluescreening this machine.  The new computer is 
sitting here, ready to get fired up.  Probably take me a week to get everything
transferred over.  You know, emails, favorites, quickbooks, all my data files,
songs.  I agonize over having to reload foobar and the dozen other programs I
have (many not free).  Damn, where did I put all those registration keys?

Mar 08

Finally got around to doing some chassis work.  Been putting it off.  But it
wasn't so bad.  Sure, it takes a lot of work to prep a chassis.  Mark holes
(precisely), drill them, countersink, debur, punch holes, clean and paint.  My
arm gets sore filing the big transformer hole (rectangle) for the TRUMPET, or
punching them damn tube holes in the top.  The punch gets amazingly hot after
grinding out eight holes.  The results, however, were awesome.  I really got
this painting thing down.  The coatings are perfect, and the glossy finish
gives it a mirror-like quality, even though there is the textured powdercoat
underneath.  It looks really good and I hope it shows in the photos.
Unfortunately though, like other shiny hi-end machines, they tend to pick up
fingerprints quite easily.  

Mar 07

All this economic news.  It's a recession folks, pay no attention to those 
pundits behind the tv screen saying otherwise.  I'm not in a panic yet, merely
to the point of apprehension.  So many things in conjunction, it's like an
intereference pattern at a horn mouth where everything coincides, a perfect
storm.  A recession I can ride.  A full blown depression is what I'm worried
about.  What does it do to hi-end audio?  Is there history we can examine?  Was
there a hi-fi cult back in '29?  Who bought hi-end equipment in the '30's?  Is
my product line properly placed and marketed for such an economy?

I look at the price of food, energy, metals.  That tells the story.  When a
donut costs more than $1, we're in trouble (the classic bellweather).  What has
happened over the past 5 years?  In 2003 gas (for me) was maybe $2.00, now it
is $3.50.  Gold was (I have no idea) $500?  Now $1000.  Copper thefts were a
big story last year.  Ok, so if we assume the price of gold, fuel, and bread
are constant, and that the value of the dollar is what is changing, then we can
see a rough devaluation of maybe 40% in the last 5 years.  Not quite half.  Or
looking at it the other way, prices are 75% higher.  

So let's look at HAGTECH.  A VACUTRACE used to be $1.25k, now it is $1.55k.  Or
a 25% increase.  It will soon be $1.7k, a 36% jump.  Only 1/2 inflated compared
to other items.  In reality, the price has gone down!  Think about it.  Do not
measure in dollars, that is the mistake.  Measure in gold, gas, or groceries.
A TRUMPET used to be $1.9K, now $2.5k.  That's a mere 32%.  And a HAGUSB is
still $119.  That's cheaper (in dollars) than it was in 2005.

Ok, so I'm not ripping anyone off.  My improvements in efficiency and volumes
have helped.  The question is, has it been enough?


I completely changed around the coloring scheme of the website.  This was done 
to match the colors of the new chassis.  No, I didn't pick them.  My wife did.
She liked the gold/yellow text on red anodize.  And then black covers.  The
silver is to represent the color of exposed aluminum extrusions and stainless
steel hardware.  Got rid of the green, which was an inversion of my red
(#801010).  The good times font will be used for banners and headers.  I also
changed to a fixed width format, something a lot of newspaper sites do.  This 
is to make it more readable.  My old site stretches with your browser, so you
can force it into any shape.  I picked 768 pixels for a width.  It will always
look the same.  Simplified menus even more.  Left out "dealers" for the moment.
About is now the same as home.  I'm not really going to put up much in the way
of content, the HAGTECH site is for that.  This is nothing more than product
photos, specs, and the ability to order them.  No fluff.  Pure, simple, clean,
efficient.  Nothing more.  I'd edit the photos to look better, but with my
computer crashing around me, I can no longer save JPGs.

Mar 06

Working again on the HAGLABS website.  Trying out various design formats and
layouts.  This time menu across the top.  A lot of people suggest I make a 
fancy website.  But hey, I am inherently rebellious in nature, and I just have
to be me.  I am making this as simple as possible.  It will be easy to look at,
use, and maintain.  No flash, no animations, no shopping cart, no "enter site"
gimmicks.  Just straightforward purity and simplicity.  Same as my amplifiers.
Why would I do anything else?  The site must reflect my philosophy and design
style.  This ain't amazon.com or facebook.  Anyway, I hope to have a new index
page uploaded later today.  I realize I have only 3 of 14 products photographed
for this.  

HAGLABS website

Speaking of style, I have to say the Veena is right up my alley.  I really like
what they did technically and visually.  It might be what I would do if I were 
a speaker designer.  I worry, however, that my high impedance low watt
amplifiers aren't going to be up to the task.  Afterall, I do NOT use feedback,
like most other tube amps (Lamm, BAT, Quicksilver, etc.).  Will I have any
dynamics at all with my 6 clean watts?  What about the DeVore Super 8?  From
what I read online it might be more "musical".  Maybe I prefer the DeVore sonic
presentation better?  I'll keep looking, but pretty soon I have to try and make
a deal.


Reading the horn article in AXP.  Wow!  This has to be one of the best articles
they've published in the last ten years.  Very well done.  Good description of
horn designs and some of their compromises.  One line in there really did it
for me, that the speed of sound is taken to be constant, and that's how you
solve for wavefronts.  Bingo!  I get it.  So simple.  You set the boundary
conditions and solve using certain assumptions, in this case speed.  It's like
solving electric field isopotential lines, just set the laplacian (del-squared
V) equal to zero.  Iterate about that condition and the answer pops out.  Well,
it is the same with doing wavefronts in horns.  Just use a piece of string,
mark off inches or whatever increments.  Pull it tight straight out.  That's
the front of the sound wave.  Now force it along the edge of the horn, that's
the same wavefront (in time) position at the horn surface.  Ok, it doesn't
solve in between (except for a conical), but it's a good start.  Intuitive.  
And now I understand where the various shapes came from.  Designers solved for
different parameters.  A cylinder gives a plane wave.  A conical a spherical
wavefront.  The tractrix came from solving for perpendicularity at the horn
surface.  Very clever.  I get it.  You do this such that there are no
reflections along it!  The wave flows smoothly across the horn.  Of course, if
you build a horn without a smooth surface you just destroyed this very unique
property.  Anyway, I can't wait for part 2.  My guess is that the hyperbolic
is the optimal design.  I just have a feeling.  Probably because the front
surface reaches a limit, whereas all other shapes continue on forever.

I find horn design interesting, and maybe someday I will get me a full system.
My patent #4 did the same thing (which is why I find the analysis fascinating).
The mass spectrometer field solution had a 1/r term in it, and I realized the
way to cancel this was to put a z term to the field strength.  Thus, if z=r,
then the solution has a term z/r that cancels!  Problem solved.  The horn is
a hyperbola of the form z = 1/x, not the generalized formula for a hyperbola,
but the simplest.  It has a special name, I forget.  Rectangular, or something
like that.  Anyway, I'm still guessing the hyperbolic cross section delivers
the smoothest impedance transfer for a loudspeaker.  We shall see.


Hey, I just thought of something while doing dishes.  I bet you anything that
the "honkiness" of a horn comes from these surface reflections.  Hence, the
tractrix has least honk, if done correctly.  The non-perpendicular fronts
develop interference patterns of cancellation and reinforcement.  Probably
very much frequency dependent.  Any of you horn builders attempting rough or
irregular surfaces to combat this are heading down the wrong trail.  It's a
nice bandaid, but not a solution.  Try, try, try, you'll never get there.  A  
good horn has to be smooth.

Mar 05

I think I'm running on overload.

Mar 04

Uh oh, prices finally moving.  I noticed big time trying to order a new batch
of USB cables.  Electrolytics, too.  I'll keep a watch on them, but it could 
force me to nudge up prices myself.  We'll see.  

Mar 03

Dang, the HP15C are going for well over $300 in eBay auctions.  I found the 11C
and 10C are pretty close, with fewer features, mainly without the complex math,
which I don't use much anymore.  Actually, if I can just figure out how to
scroll the digits to the right one place, I can live with my broken one.  It's
just the leftmost digit is in a broken LCD area.  


Looking into ways to run an automatic picture slide show for my new website.
Sort of doing a nice fade in and out of various glamour shots of the new HAL
products.  Found this blendTrans() filter function that is somewhat helpful.
Looks like a bit of work, though.  Maybe I need to just do the flash thing.  Or
insert an applet.


Running short on parts again.  I swear, I'm placing new orders while the
previous orders are still in transit.  Out of screws.  Out of HAGUBS boxes.  It
never stops.  Those things are selling like hotcakes.  And most going overseas.
With the dollar dropping like a rock, I am now shipping most stuff to foreign
addresses.  Hmmm, maybe I shouldn't have made the HAGLABS stuff 120Vac only.


Ok, I sort of figured out how to do this.  I used the VACUTRACE page as an
example.  It used to have a bunch of example curve images that if you clicked
on it they would change.  Ok, so I automated it and put in the fade feature. 
It seems to work.  Maybe just on IE, not opera or safari, I really don't know.
Hopefully they will at least display the first image as static.  Anyway, try it
out, see if the image fade feature works for you.

VACUTRACE page with image fading

Mar 02

Someone emailed me saying they read an article in the Chicago Tribune about the
HAGUSB.  So I dug it out.

HAGUSB article

Mar 01

Ok, I just bought a new computer.  Pretty lazy about this, I always just buy a
Dell.  Yeah, I'd rather have an iMac, but am so entrenched with PC software, I
got no way around it.  Layout, schematics, credit card processing, etc.  This
was not a purchase I really wanted to make, but having been given the two
minute warning, I'd be a fool not to.  Why Dell?  Because I like to really
personally configure the machine.  No box store is going to package all of the
options I want.  For $800 I get quite a machine.  Delivered.  The Dell prices
are tricky, always start with one of their "deals" or special promotions, then
tweak from there.  What else do I recommend?  Optical mouse.  Floppy drive.  I
even splurged for the widescreen 2" displays.  Always wanted to be able to put 
two word docs side by side (I write a lot of product manuals).  Windows XP.

And then I read the stereophile review of the cayin integrated.  Jeez, $1300
retail for this tube amp.  If I built that thing, it  would cost me $1000 in 
raw parts, unassembled.  And that's their street price!  They actually build it
for probably half that!  That's why I don't compete on price anymore.  I sell


Eventually, you are what you leave behind.  


Hey, good news.  Just got notified the first of my three pending patents was
approved.  I don't know if it's been assigned a number or not.  Have to wait
until the lawyer sends me the forms to sign.  Anyway, the basis for the 
invention is described in this white paper.

Hyperbolic Horn Helical Mass Spectrometer

Feb 29

Just so you know, the other speaker I am looking into for RMAF08 is the Veena
from 3A.  I heard the de capo years ago and was mesmerized.  The idea of 
direct coupling really appeals to me.  So far, it's either these or the de Vore
super 8s.  Remember, I gots me only 6 watts.  Any other recommendations?

Speaking of which, the Melody i2A3 reviewed by Tellig claims 18W out of the 
same tubes and configuration I get 6W.  Damn!  What am I doing wrong?  Well, it
is because I operate in full class A.  They must be pushing it very close to 
class B, with a small AB crossover region.  Put the bias at 350V and 15mA,
you're still on the recommended operating load line (which I don't use because
there is a sonically BETTER load line).  This shift in bias gets them from a
limitation of 100Vp signal swing to 170Vp swing.  Basically, only one tube is
operating at a time, the other is in cutoff.  I guess if all you care about is
power, then this is how you do it.  


Ok, good news is that Saelig says they won't charge me the 15% restocking fee
for the Owon scope, since it didn't operate as advertised (no persistence in
XY).  They seem to be good folks.  I've bought stuff from them years past, the
USB scope and an I2C unit.


However, trying to start the FLUGELHORN layout modifications to the ARCHIVER, I
found the file would not read in.  That's ok, I got plenty backups.  And I've
been having boot issues and a lot of blue screens lately.  Gee, I wonder if it
is time for a new computer?  So I figured I would do a chkdsk.  Sure enough, it
fixed the FAT partition where I run layout software.  Then I ran it on my C:
windows drive, see if that helped too.  Well, it started, lasted for awhile and
then blue screened.  Did it again.  Got farther, almost done!  BS.  Sheet.  Am
I ever going to get through this?  Keep in mind, each cycle here takes about an
hour!  I'm wasting the entire morning, wondering if I will ever see my email
again.  Third time it made it.  And the blue screened.  I tried going around it
with other boot disks, etc.  No luck.  The problem is that now every time I 
boot it starts that 1 hour chkdsk routinge.  I finally figured out that
unplugging the machine right as it starts the chkdsk, it then realizes it
actually crashed and lets me start up with "last known working" whatever.  
Ding!  That did it.  Got me booted up and launched into XP again.  Everything
still here.  Email, quickbooks, credit card processing, etc.  Suppose the HD is
on its way out.  At least I have time left to do something about it.  I'm not
worried about losing data (everything of value backed up), merely the week long
hassle of configuring a new machine.  

Anyway, I did get the FLUGELHORN relayout done.  It makes for much easier
wiring and fit into the new chassis.  Will probably order production boards on
Monday.  I always like to give it a day or two (if possible) just to see if I
think of some feature to add (like a ground hole or test point label).  


The first perfect HAL VACUTRACE is done.  I'll take a bunch of photos for use
in manual and on website.  I'm thinking on my homepage a spot of cool product
photos being shown as a slideshow would be neat.  Gives the page some "life".

Feb 28

CORNET2 boards came in.  I will fill all backorders today.  Trying to get as
much done this week as possible, as my wife will be out of town all next week.
Going to be difficult, with all of the trips to school and all.


Finished a CHIME and VACUTRACE today.  Well, except for one is missing the
handle ferrules and the other a plastic LED dome.  Jeez.  Ordered the parts
last night.  So difficult to keep perfect JIT inventory.  

Thinking I'll do some minor makeup changes to the FRYBABY-HAL version.  It will
be red case, of course.  I'll put the gold plated HAL RCA jacks in it, and also
the more expensive solid aluminum knob.  Common pieces with the rest of the 
HERITAGE series.  Price will be a tad higher, maybe $269 or so.  The important
thing is that I don't screw my dealers MD and AA by undercutting them with the
free shipping.  They still get the original FRYBABY at $249.  


Tried out that Owon digital scope.  Whoah, really cheap.  They did a great job
making this low cost.  But hey, if it works, great!  The display is huge, maybe
too much.  At the low resolution it is fuzzy.  Also really low contrast.  Kind
of hard to look at.  Traces are not clean.  Anyway, finally hooked up the
VACUTRACE to it today to test it out for curve tracing.  Fell down flat!  Ok, 
it runs in XY mode, but their display is not so good.  The USB scope I have is
way better.  And the damn variable persistence mode does not work.  As I change
it from none to 1s, to 2s, 5s, and infinity, there is no change in the display
whatsoever.  The function does not work.  I am probably the only one who ever
tried it.  As a result I could not get a reasonable display of curves, as it
would only capture parts of them.  I am very disappointed, I thought this was
going to be really trick.  Sending it back.

Feb 27

Updating schematics today (SU-10, et. al.), sketching up layouts for HAGLABS
website, and I realize I need a better way to connect the output of a PL-10 to
both an HA-10 and PA-10.  Room for only one set of RCA output jacks on the
PL-10 (a compromise, you wouldn't believe the ripple effect from such a small
thing).  So I figured, hey, just put a loop-through on the HA-10!  One set for
input, the other output.  Change the colume control to 100k and it is quite
transparent.  Turn the volume to zero and it looks like a pure 100k resistor in
the audio band.  I'm also adding an extra RCA to the PA-10.  Same thing.  This
way you can loop out to a subwoofer with same signal.  The TEN series is now a
small line, but nicely integrated.  The LC provides power, PL the control, DA
the digital, and PA the output.  You get a complete system by providing the six
source components, such as PC (for DA), FM, SACD, TAPE, and 3 TT.  Yes, you can
connect three turntables (or arms) at same time.  

Decided for sure to do my "what-if" for the HERITAGE series.  And it got big.
It has 4 categories.  1) BURN-IN: with FRYBABY (in red), FRYKLEANER, and
FRYDADDY (which I am again working on the proto).  2) TEST: with VACUTRACE,
JUMPER and BLANK adapter cards.  3) ACCESSORY: with UFO (black).  4) PHONO:
nice and full, all re-packaging of previous machines to go with the new HAGLABS
appearance.  And lucky for you, HERITAGE is only sold factory direct!  

I can't tell you how much stress that decision took off of my shoulders.  Panic
with this possibly faltering economy, how to justify the doubling in price of a
TRUMPET, etc.  It also solves various little issues, like knobs, umbilical, 
etc.  Just plain makes more sense, and most likely is a more profitable
strategy.  It buys me a LOT of time before I have to make any dealer or
distributor decisions.  


Thinking the panel for the FRYDADDY will look something like this.  It fits
into that rugged serpac box, same as I use for VACUTRACE adapters.  Compact,
rugged, portable with handle.  Very professional it will look, just like a mini
VACUTRACE.  Anyway, it's basically a re-shaping of the FRYKLEANER (GOLD) with
more output power and a connector for burning in power cords.  It has the usual
binding posts, XLR, RCA, but the BNC is a raw output, not current limited.  You
can then use this to either monitor the signal, or have it drive something
else, like a speaker or custom rack of connectors.  It makes the machine more
versatile.  Lots more output swing and power is due thanks to an LM1875 power
amplifier and balanced output transformer.  The AMPLITUDE control is like that
on the FRYBABY, but with the added CABLE setting, which is 20dB above line
level.  I'll have to run calculations, but I think I can get it close to
8Vrms (+20dBu).  Price will be somewhere $599 to $699.  



Ok, the text description of what I did for the SU-10 (now STEP-UP) in weeks
past may be confusing.  So I finally got the schematic done.  This is how I get
the CineMag CMQEE-3440A tranny to work best for me.  The outputs float, getting
their ground reference from the phonostage inputs.  The case gets a copy of the
separate TT ground wire (no signal on it).  I then tie this to the tranny
inputs, as they need some sort of common mode reference.  I don't like floating
them in space.  This gives the cart a reference ground, yet it has no signal
currents on it.  It is the exact same ground as what the tonearm gets.  This 
works best for coaxial cart cables.  If they are twisted pair, it is also fine,
although you might be able to do better by balancing it.  This is done by
splitting the primary side loading into two resistors, each of R/2 value.  The
center tap to earth ground.  This forces the cartridge to operate in true
balanced mode, by providing the necessary 3rd terminal.  Anyway, my machine
will be with single-ended RCA connectors.  S1 selects primary side loading.  S2
changes gain.  Gain is set by selecting a primary tap, and then the same switch
kicks in the proper secondary tuning for optimum transient response.  You can
have your cake and eat it too.  Not a huge range of loading (LO, MED, HI, and
extra HI), but it covers the full range.  You'll also find that primary side
loading is NOT as sensitive as secondary side loading is.  Hence, relatively
big steps are ok.  

CineMag wiring

The result is gains of 16dB, 22dB, and 27dB.  Lowest possible loading on cart 
is roughly 21 ohms.  


Reading Entrepreneur magazine tonight.  Honestly, it's not my thing, I actually
read very little of it.  But it does make me think.  Ok, ok, that Kawasaki guy
writes a good column, which I read every issue.  Mostly, I skim until I find a
good tidbit or sidebar.  Anyway, has me wondering if I should put up an eBay
storefront.  Or add automatic paypal invoicing to my website.  Mostly, I
realized I need to operate the new company the same as the old one.  Both 
should have identical policies.  This will also help prevent confusion.  So
what does this mean?  Well, for starters, I had envisioned HAGLABS products to
not have free shipping.  Hmmm.  You know what, that is exactly backwards!  With
the declining value of the dollar, I need to focus more on foreign sales.  That
means I need to offer free shipping, always.  Not just USA, but worldwide.  
Heck, for a long time now I've neglected to charge shipping on HAGUSB orders.
Mostly, it's just a pain in the butt to spend the time putting on an extra
$6.30 charge.  Yes, this is the way to go.  Focus on the customer needs, make 
it as easy to buy as possible.  So there it is.  Or rather, here they are, the
NEW rules for both HAGTECH and HAGLABS:

1)  Free shipping.  Anywhere.  My choice of carrier, though.  That is UPS
ground for large boxes to USA, first class mail everywhere else.  If you want 
it faster, you pay.

2)  30 day trial period.  If you are not satisfied, for any reason, you can
send item back for refund.  See #4.  

3)  10 year warranty.  Transferable.  That's right, I don't care how many times
it changes hands, the warranty is still valid.  90 days for tubes.

4)  5% restocking fee.  

What?  Where did that last one come from?  Well, I have to cover the free
shipping costs somehow.  I think this is reasonable.  It also prevents abuse.
Dealers subject to same rules (yes, they get the free shipping!).  

Feb 26

Lately I've been playing a lot of FLAC files through my proto DA-10.  And I 
just love it!  I am having so much fun with this digital stuff.  It can really
sound good.  I love digital.  And then I find myself playing LPs.  God, I love
this sound!  I love analog.  Hmmm, revelation.  If I had a tape machine or a
decent FM radio station to tune into, I bet I would love them too.  And there 
it is.  The answer to that long running debate on analog versus digital - they
are both awesome!  Merely different means to the same end.

And I am happy to report than some of the latest songs to hit the charts are
recorded really well.  Songs such as "Bubbly", "Teardrops on my Guitar", and
"Delilah" are products of fine engineering and mastering.


Thought I would see what I could do to improve on the HAGDAC, other than 
removing the unnecessary parts from the optional gain section.  Nothing obvious
as it was pretty well tweaked a few years ago.  So I ran another set of jitter
tests, thinking maybe I could lower the output jitter from the CS8415A SPDIF RX
chip.  Triggering off of the SPDIF signal (from my HAGCLOCK modified CD player)
I can easily see edge jitter on a scope of the receiver's output.  On an analog
scope I see roughly 2ns peak to peak.  Now don't think that's a crazy number,
because it likely means something like 300ps RMS.  The problem is that there is
no easy way to correlate PP to RMS (except using the erfc function and BER
rates).  Oh wait I could have photographed it and then looked at the histogram
of the luminance values.  Duh.  Then if it approximates a gaussian curve, the 
RMS is equal to 1 sigma.  Maybe next time.  Anyway, I tried various shifting of
the PLL filter values to see what helped.  Heck, the data sheet recommendations
produced some horrible jitter.  Fast lock times, though.  Finally, after a lot
of tweaks, I realized I was right back where I started.  The good news is that
the clock output from my VCXO PLL reclocker (after this) totally cleans it up.
Absolutely no visible widening of the trace.  Not just a 10x improvement,
either.  Way more.  So I have to believe the 5ps RMS specification for the VCXO
is realistic.  The PLL will add a little random noise to this.  Regardless, it
is rock solid performance.  I wish I could get stereophile to put it on their
analyzer.  Hmmm, maybe if I can get them to review a DA-10?

Feb 25

Finally got that FLUGELHORN prototype completed.  Took more work than I 
expected.  Bought me this very nice twisted pair cable for the I/O signals, it
is silver plated copper, stranded, in teflon with outer teflon jacket.  Nice to
work with.  I wired up every thing direct, sort of point to point.  Missing one
LED at the moment, but from this photo you can see the boards mounted to the
panels, nothing but wires in the middle.  Chassis is lightweight, also consumes
only a few watts of power.


Measurements look ok.  So I hooked it into my system.  Hey, noise floor is
really good.  I was worried about the twisted pairs too close to the ac power
supply.  No problems!  In fact, this machine sounded great.  I was stunned at
how clean and neutral and detailed it was.  With zero break-in!  Hmmm.  I'm
working from long term memory, but this sounds better than the original
ARCHIVER.  Is it the better power supply?  Could be.  I have a few extra volts
to work with, so a tad more headroom.  Wow, this sounds incredibly good.  It
doesn't have that "slowness" that I found with the original.  Remember, I 
always thought the musical pacing was slow and relaxed.  Not with this version.
Everything moves along with great detail and dynamics.  Missing that extra
weight and fullness of a CORNET.  Or maybe it is just the difference in
presentation between the PICCOLO and STEP-UP.  Afterall, the FLUGELHORN has a
PICCOLO-like front end.  Very natural, but not with the usual solid state
staleness (BUGLE).  Oh wait!  I just remembered this sample uses different
caps.  I wonder if that is it?  I'll FRYKLEAN it and listen again later.  If
all is good, I'll paint the covers.  Then I can finally get a real photo of its
final appearance.


Hey, my new digital scope came in already.  I'll have to go check it out.  Have
a new VACUTRACE half built, will make for a good XY test.  


Shoot, I just can't get over this.  Playing the new FLUGELHORN and it continues
to amaze me how damn good it sounds.  Spot on.  So I ran it up against my
reference CORNET.  Whew!  Thankfully the CORNET edged it out, slightly more
real on cymbals and more chesty and full.  That part I don't understand, as
they measure identically on equalization.  It ain't a dB thing.  The CORNET
with PICCOLO front end was awfully close in sonics.  Noise floor just a tad
higher than the FLUGELHORN.  CORNET and STEP-UP is definately quieter again.
More of that organic sound.  My preference.  Back to the FLUGELHORN, very much
quiet enough for use with my LOMC, slightly leaner and more neutral.  Cranked
up some Jacintha (not exactly a rocker) and let it fly.  No fatigue, easy to
hear down into the deep depths of ambience.  Even the bleedthru from tracks was
clearly intelligable.  For the same price as a CORNET, it ain't giving up much,
while offering way more flexibility, features, and gain.  Honestly, I really
didn't expect this much from the machine.  Now I just have to figure out if it
the power supply or the caps that did it.  Have to build another one, I guess.


Sold the last FRYGOLD.  Now it is red.  Same machine, new color, new name.  I
am giving it the original FRYKLEANER moniker.  Hey, this is the culmination of
everything I learned on generations one (FRYPRO) and two (FRYBABY).  FRYDADDY
will be gen 3.  Anyway, same price.  Same everything except color.  Once I get
this unit done I can take photos.


Feb 24

Still wondering about my retail/direct mix.  After crunching a lot of numbers
for business plan, running various scenarios, I find some interesting things.
My complicated pricing formula (different multipliers based on amount of
inventory carried, plus 10% for shipping), which comes up with the retail price
ends up being almost an exact X multiplier.  Industry standard is closer to 
2X.  Shux, in about ten different part lists, they all end up at X.  And that
means once you take out the dealer's 40% cut, you get a Y markup for oem
pricing.  Y = 3/5*X.  Damn, I had been using Y for years!  It has turned out
to be marginal.  It works fine with robust sales, not so with minimal revenue.

So I am simulating another what-if.  What if I sold the new TRUMPET and CORNET
direct, instead of retail?  I would make them roughly $2499 and $1499.  And
chances are, I would end up selling more of them.  They are both older designs,
I'm not so confident dealers will want to to sign up for them.  Plus they have
a history of their mark I versions costing $2k and $1k.  To me, this makes more
sense.  The problem is that I suck at marketing.  I have often before undersold
myself!  Still, seems to me from a profit viewpoint, selling these models
direct could be better.  Besides, if we enter a recession right now, this
strategy makes me way more competitive.  So maybe I do a "classic" series under
the HAL banner?  Hey, Klipsch has their "heritage" series.  Then I separate the
TEN series as  the newer stuff.  Do I screw my dealers this way?  

I like that, "heritage".  Perhaps I organize in this category the CORNET,
move the PI-10 and SU-10 with them, as they sell in combinations.  And
honestly, over a grand for the SU-10 seemed high to me.  They would be maybe
$449 and $649.  In the end, the profit margin is exactly the same for me.  The
premier "10" series is thusly shrunk to LC-10, PA-10, DA-10, and the upcoming
PL-10.  The glitch in this logic is the HA-10.  It is an all new TEN machine.
Do I sell it as a TEN or as the CASTANET?

Damn, I hate wavering on this stuff.


Ok, I looked again at my formulas, and I realize the X multiplier method is
just too simplified, not taking certain things into account.  The pricing is
better with the advanced method.  Plus, using the Y multiplier, the VACUTRACE
goes to over $2000!  I realize this is a bit messy, but it is a transition
period, so I have a good exuse.  The TRUMPET and CORNET HAL models will not
last forever.  They are an interim measure until I can get PL-10 production up
to speed.  So now I get the HAGLABS HERITAGE series sold factory-direct:

PICCOLO (red)    $449
STEP-UP (red)    $649
FLUGELHORN      $1499
TRUMPET (red)   $2499
CORNET (red)    $1499
VACUTRACE (red) $1699
FRYKLEANER (red) $449

And then the HAGLABS TEN series, retail:

HA-10 $2410
LC-10 $2080
PA-10 $3240 (each)
DA-10 $3760
PL-10 $????

Eventually I will add the FRYDADDY as a heritage product to replace the
FRYKLEANER (which is a red version of the FRYGOLD!).  The CASTANET will be sold
as half-kit only.  


I just realized the UFO (assembled) can be a HERITAGE product.  Only have 3 or
4 more left, so it's not a big deal.  On my HAGTECH home page, I moved all of 
the HERITAGE-to-be stuff down to the bottom.

Feb 23

Aw crap, I dropped my trusty HP 15C calculator.  The LCD is broken over the 
left most digit.  Shoot.  25 years I been using this thing.  I'll have to find
another on eBay.  Can't work with anything else.  My brain works on RPN.  Dang!
A good condition 15C on eBay is up to $222 already with 2 days to go.

Sometimes I am nuts.  As if I didn't have enough work to do, I discovered a new
project.  Was sort of simultaneously working on a VACUTRACE, CHIME, and the 
FLUGELHORN builds when I came across the panels for a custom ARCHIVER.  This 
was for the project I cancelled, the panels didn't fit right into the chassis.
But hey, here they are, beautiful panels in every other way, and worth about
two hundred clams.  Last thing I want to do is throw them away.  And then I 
realize I have an old 19" rack mount chassis on the shelf, from about 5 years
ago.  Dang, these ARCHIVER panels fit right onto the existing ones!  Shoot, it
looks pretty damn nice, too.  So here's about $300 in parts, just rotting away.
Fine, let's build it!  I can turn this into a revenue project.  A one-of-a-kind
custom ARCHIVER in a big metal box.  It is actually sort of the fore-runner to
the FLUGELHORN.  So here you see the two partial assemblies side by side.


I will sell this on Audiogon.  Luckily I have maybe 90% of the parts already in
house.  Yeah, this one is all about making a few bucks.  I admit it.  Then
again, anyone who ends up with this for under a grand will score a fabulous
phono machine.  Seriously, compare this jewel to a ph3d.  


I remember all the shit I got for making the CHIME the way I did.  Like in John
Rocker's professional magazine review of it.  He just could not figure out why 
I did what I did.  It made no sense to him (then why the frik didn't you ask?).  
So he bashed me in writing (ultrafitimes).  And many other folks did it in
various online forums.  They all thought I made a huge mistake.  Well, time has
proven me right.  I foresaw all of this.  The main question centered around why
I chose to go from USB to SPDIF and then to I2S, instead of straigh to I2S,
like everyone else does.  Simple Sherlock.  It's to avoid the ground
contamination.  So many audiophiles have problems and resort to nasty solutions
like in the post below, or they end up buying the super expensive Opticis
cable.  Read what this guy has to try and do.  It's insane.

Asylum post

All HAGUSB and CHIME owners have avoided this!  I see similar posts all the
time and just want to scream.  Hello?  Now do they get it?  Now do they
understand the reason for the isolation tranny?  Ok, sorry.  I had to get that
off my chest.

Feb 22

Lots of good forward progress this week.  I solved a ton of technical issues,
bought lots of production pieces.  I'm still wondering about how to split up 
the marketing of various machines.  The CASTANET is the main problem.  The 
question is how to turn it into an advantage.  You see, the retail price with
dealer markup is $2k+.  That's ok, but not when you can DIY the machine for
under $600.  I've wavered on this tpoic before, but I think it is reasonable to
offer some machines stricly direct, not through retailers.  It has to be a
balanced portfolio that makes sense, and does not alienate any customers.  I've
already put the CASTANET (black panels) up for sale.  The thing is, I want to
migrate some of the HAGTECH machines to red.  I re-did the FRYGOLD into a FRYK
red.  VACUTRACE is now red.  The new FRYDADDY to replace FRYGOLD will look
like a VACUTRACE in red (but small).  Yeah, the FRYGOLD missed the mark.  It
does not sell anywhere near as well as the FRYPRO did.  Better machine, wrong
form factor.  So I just take the loss, discontinue it, replace with the
FRYDADDY.  Well, the reasons are many, but to make a long story short, I'll
continue with the HAGTECH professional series, which mainly consists of the
VACUTRACE.  I will make the CASTANET red and move it into this category, and
then add the FRYDADDY.  Meanwhile, I'm going to pull the FLUGELHORN from the
HAGLABS lineup and move it here.  Again, a SS phonostage for $3k just doesn't
seem right.  The damn thing costs me a lot to build, though.  I will only
offer it in 120Vac, so it doesn't compete with the similar AA oem unit, which
may retail for about $2k (import costs & markup).  Same sonics, different 
chassis.  The red HAGTECH units will have the HAL logo.  Given recent inflation
and other considerations, I think these will end up at:

VT-10 (VACUTRACE)  $1699
HA-10 (CASTANET)   $1299
FL-10 (FLUGELHORN) $1299
FD-10 (FRYDADDY)   $599

All great machines.  These are factory-direct prices!  After you buy one,
you'll wonder how I can offer the VT-10 so cheap.  The HAGLABS retail lineup
then is something like this:  Remember, prices are NOT finalized, and are
preliminary.  Each of these is a really, really fine piece.

LC-10 $2080 (Line Conditioner)
PA-10 $6480 (Power Amp) per pair!
DA-10 $3760 (Digital to Analog)
PI-10 $730  (Piccolo)
SU-10 $1080 (Step Up tranny)
CO-10 $2380 (Cornet)
TR-10 $3790 (Trumpet)

So yeah, I continue to organize, upgrade parts lists, set up business model.
Launch date for HAGLABS is still 4/1, although that only gives me a month!
Keeping up with current production makes it difficult.  Right now I am humming
along with my time split roughly into thirds.  One part production, one part
R&D, one part overhead (customer service, order processing, purchasing, etc.).
That seems a pretty reasonable mix to me.  I get all emails answered in a
timely fashion, most orders ship right away, and I still get enough design time
in for new machines.  The R&D is something you have to continually invest in.

Whoah, that should keep me busy for this year.  I hope it leaves time enough to
get a proto up and running of the PL-10.  Afterall, that will be the flagship
component.  Must be ready by RMAF08!  Meanwhile, I'll see if I can slip the
FRYDADDY in on weekends.  Mostly it is already designed.  I should also be able
to do the CLIPPER detector, as a small half-kit.  It will be about the size of
a 9V battery.  Then in 2009 I can do the PA-20.  Yes, I am already thinking 
what I need for RMAF09.  Mostly, a more powerful amp.  Maybe LS-10 too?

Feb 21

I scanned in the final PA-10 schematics last night.  Proprietary stuff, but 
shoot, nobody is gonna copy me.  Nice stuff to look at, though.

PA-10 power supply
PA-10 amplifier

Did some more testing.  At full power the waveforms begin to distort at the
frequency extremes.  So I swept a -3dB level to see where distortion begins.  I
get clean response from 25Hz to 19kHz.  Above and below that, the waveforms 
deviates from a perfect sinewave.  That would be full power bandwidth?  At 1W 
the response is much wider.  I also did a distortion check at 1kHz at a 10W 
power output level.  It was 3.5%.  Hmmm, a lot of other manufacturers state
their power level at 3% to 5% distortion, so I guess competitively I could call
this a 10W amplifier.  Not sure I will do that.  Here's the bench testing at

PA-10 testing


Searching for new knobs.  I need a round big knob for the PL-10.  Can also use 
on the HA-10.  But it cannot have a pointer!  That's what make it hard to
locate.  The best so far is the alcoswitch kn1751bs1/4.  At newark they want
$33!  Mouser is only $21.  Sheesh.  The good news is that it has little brother
sizes for the SU-10.  This will set the machines apart from the HAGTECH stuff.
Damn cost of aluminum!  


Back into the lab for DA-10 testing.  I tried out some silicon-carbide diodes 
to see if they gave any improvement in the inductive switching crossover glitch
(see Feb 07).  No change.  I think this is all a secondary resistance issue.
Current waveforms are beautiful, the glitch only shows up on voltage, and isn't
reflected too bad onto the primary.  Whew!  Ok, so I tried the common mode 
choke (balun) trick installed in front of the diodes.  Again, no matter how I
wired it up, no visible change whatsoever.  Maybe I had too low inductance on
the balun?  I dunno.  But zero change.  Fine, enough research down this drain.

Then I checked out a bunch of surface mount shielded filter inductors for the
HAGDAC I-V circuit.  The pure RC variation produced very little noise.  I got
about 0.2mV broadband on the outputs.  That's 80dB SNR unweighted.  It's good
enough with critical listening tests.  But I really wanted to restore the 
original 3rd order Bessel filter, for improved noise rejection and transient
response.  The stock coils gave me 0.26mV and 0.14mV.  I forget which channel
was favored.  Turned the noiser side coil around (different field orientation)
and got it down to 0.2mV.  This seems better than before.  Tried the 308-1513-1
which definately added some buzz, up to 0.24 and 0.36.  The lower inductance
513-1392-1 gave me .14 and .14 after I figured out orientation (they are 
opposite each other).  The PCD2127CT was also very good at .16 and .14.  Using
high gain and headphones, both of these solutions we pretty much buzz-less and
pleasing.  The 553-1100-1 gave .16 and .38, but I didn't bother to turn it
around.  I like the 513 and the PCD better than stock.  They both have improved
construction, but I think my choice comes down to the 513, as it is larger with
more ferrite.  At 0.14mV noise it sounds the best.  It is rated 330uH at full
current, so at almost zero current, inductance should be much higher, hopefully
close to the 470uH nominal of the filter.  It is not critical.  And so that
problem is solved.  Now I can order production DA-10 boards.  With A-weighting,
this is probably 85dB to 90dB.  


Alrighty then.  Just ordered tons of production pieces.  Not ordering huge
quantities of circuit boards, as it may not be wise to sit on so much inventory
at this time.  And I hate to dig into debt.  The price is that PA-10 boards end
up costing me $49 a pop, and the DA-10 is $52.  Hey, a C2 half-kit at $99 is a
steal!  Where is my margin?  I also forked over for production panels.  Bought
me DA-10, HA-10, PI-10, SU-10, and PA-10 sets.  Also bought a set for the
Artisan Audio project, an upgraded and modified oem version of the ARCHIVER.  
So now I should have the major pieces to get production started.  Almost done
with units #1 of CO-10, TR-10, and FL-10.  The design for LC-10 is lagging, but
the big isolation trannys are on their way.  I have to have at least one
perfect example of each new machine for website and brochure photographs.  That
reminds me, I'll need some for reviewers.


Some customers reporting issues with tics on their HAGUSB.  Anyone else out
there getting random glitches?  I'm thinking it could be an issue with xtal
loading.  Changing C9 to 47pF to compensate.  Might have had a batch of
marginal 33pF caps.  I haven't confirmed this yet.  If you are having trouble,
let me know.  So far, this is may be 1% to 5% of units.  Presently running
some tests.  

Feb 20

Really worried about this inflation thing going on.  The devalue of the dollar.
Ok, the good news is that I get a ton of sales from EU.  Everybody elses prices
are jumping 20%, or so it seems.  So I have to keep track of my costs, reflect
the final retail pricing using my formula.  It's very hard to keep exact
numbers on every part used in 20 or so different products.  No, I'm not totally
computerized with automation (such would cost me more than what I might save?).
So, I double checked my parts for HAGUSB.  It's a perfect example.  I figured 
my suppliers track their prices really well, so any change by them is a real 
one, and I should adjust.  Well, it hasn't hit me yet.  The price for HAGUSB
parts hasn't really changed over the last two years.  Why is that?  I think the
suppliers are getting more efficient.  They have kept pace.  That means I do
NOT have to raise my prices.  I was thinking HAGUSB might bump up another $10,
but it does not have to happen yet.  Honestly, I am suprised at this outcome.

Maybe my calculations are off, but if that satellite is to be shot down at 3:30
GMT Thursday, then that would be 5:30pm tomorrow night here.  The boats are 
straight west of here.  Shoot, I can stand on my roof and see this.  Too bad it
won't be dark.  Let's see, 10 degrees longitude west of Maui, we got 25k mile
circumference, that's 700 miles.  Aw crap, I can't see it.  Nobody can.  Too 
much curvature.

Satellite 193 kill zone

What?  And now I read they may delay the shot because of weather.  Hello?  The
weather out here is absolutely friggin perfect.  The seas are dead quiet.  

Meanwhile, I just sold the last assembled CHIME.  All out of chassis parts now.
I can continue to sell the kit/2 style for a long time.  Anyone wanting an
assembled machine can look to the newer DA-10.

Relayed out the PA-10 to put in all the recent changes.  Little stuff, mostly.
Add a 33 ohm resistor to ground for the speaker outputs (they were floating). 
Add the 1.5k in series with interstage primary, change filament dropping 
resistors to 3 1W in parallel rather than a single 2W (which don't come in the
0.33 ohm value required), and add the traces to back panel for RCA input.  This
design came out better than I expected.  Superb topology, and thanks to that
6E5P driver tube it delivers ultra-low distortion.  Shux, 0.12% at 1W, with no
feedback anywhere!  And less than 0.5% at full power?  That's amazing when you
consider the huge voltage swings involved.  


Reading EDN, I came across an ad for a new cheap digital scope.  Hmmm.  It is a
dead ringer for the Tek TDS320 from 15 years ago.  Made in China.  But hey, it
might not be so bad.  There are times when I wouldn't mind having a digital
scope.  They can come in handy.  Shoot I think I might buy one of these things.
It looks to be a perfect mate for a VACUTRACE.  Does XY mode and has a hardcopy
button which send the screen to PC (or printer?).  How handy is that?  For this
price, it is hard to pass up.  

Cheap digital scope

Ok, I just bought it.  Will let you know how it turns out.

Feb 19

Hard to get back to work.  We are having some phenomenally idyllic weather
here.  Got my Lundahl trannys in already.  Fast!  They must want to be my
supplier.  They look so nice.  I'll have to fire them up and see what they do.
Wish I had my HP4261A.  The handheld inductance meter I have gets easily

Whipped up a quick circuit yesterday to do the clipping detector.  I had the
parts already on hand, so made one up today.  Decided to use a BUGLE as the
circuit board.  Easier than a full custom breadboard.  Came up with the circuit
below.  Then I realized, duh, I use a single +9V battery supply, but forgot
about the negative signal swings.  They will reverse bias the substrate diodes
on the inputs to the comparators.  So I reduced signal level by nine.  Using a
2.5V reference, I can therefore measure peaks up to about that level.  The
extra clamp diodes on the input are not necessary.  I combine both channels to
trigger a single one-shot, which fires an LED for 1ms minimum.  You could use
an LM556 or two of them if you want an LED per channel.  For me, as a
diagnostic tool, I don't really care.  

Clipping detector schematic

For the PA-10, it reaches full power at 0.95Vrms.  That's 1.35Vp.  I take that
and divide using 100k/10k pad and get 0.122Vp.  That's the level I adjust the 
pot to.  All I have to do is insert this device inline with the PA-10 inputs 
and I'll know when I am driving it too hard.

Clipping detector board

Feb 17

Trying to catch up on a VACUTRACE order.  Them things take forever to build.  
Also a backlog of FRYBABYs.  Built me some more interconnects, and experimental
speaker cables.  I figure, hey, this amp has a damping factor of 2.5, so why
would I need 12 awg cables?  So I built some using twisted pair stranded silver
plated copper in teflon.  Only 24 awg I think.  We'll see if they work as good
as the thick solid silver I have right now.  Interconnects are military surplus
silver plated copper, stranded coaxial in teflon.  Nice small diameter.  I
loathe the garden hoses most of hi-end uses.  Burning in right now.  And then I
get this email regarding the FRYBABY, MM asks "are you a sane person, or just a
con artist?"  He he he.  I told him "neither."

Meanwhile, I added a set of RCAs on the back of my CASTANET that are connected 
to the HI outputs.  Thus, it makes me a simple low gain linestage.  I used it 
today to play the PA-10 into a frankenvandy.  DA-10 as a source.  Nice to not
have it cranking or pushed into overload (DA-10 pumps out 1.9V, double what the
PA-10 can handle).  Anyway, one-channel sound is still thin.  Bass is there,
deep, but shelved down.  Everything is too midrangy.  I'll not know until I can
get the second channel running and the speakers placed, or buy an SS amp.  Was
thinking of something like a Bryston 2B.  Pretty cheap on eBay.  Would make a
good reference (high damping factor).  I should also try the 4 ohm tap.  The 
other thing I discovered is that the microphony was not the 2A3, but the frame
grid 6E5P.  It has the same "ping" sound as the 7788.  Some o-rings helped.
Still, you can hear it with your ear up against the speaker.  I'm wondering if
I wait until production boards come in or just build me another proto.

I might not buy me that new Mac for a music server.  Burning through the cash
way too fast lately.  I don't mind spending it on parts that I will end up
selling in a product, but I do mind spending on myself.  Seems I could use my
old IBM laptop (the same one from RMAF06 and RMAF04).  For some reason, the 
usb-audio driver is working perfectly now.  I has been having trouble swapping
between the EMU for spectrum analyzer measurements and the USB audio out.  Now
it never glitches.  So I must have done something right in one of the myriad
dialog boxes.  No idea what it was.  Oh hey, that reminds me.  I am also able
to reproduce the random dropouts some people are experiencing.  This happens
to me regularly (up to 10 times per song) when I stream through WiFi.  Move the
FLAC file to my local disk and all is well.  So, now that it is working well, I
can just make some playlists that are appropriate.  Been buying a couple tracks
every week now.  Yesterday the acoustic version of Plush.  Anyway, the old IBM
can play CDs at the show.  So bring 'em!  I'll see if it can also do double
duty as a scope for the VACUTRACE.

Started looking into the LC-10 again.  I have one half built.  Need to buy 3
more isolation trannys.  Looks great, except the fuses are under the line
filter and you can't reach them.  Need to do a slight redesign.  Putting
together a ton of notes on stuff to do.  Tweaks here and there.  Like that I
have to move an electrolytic on the PA-10 board in order to shift the binding
posts to the middle.  Also using just 3.  One ground, a 4 ohm tap and a 9 ohm
tap.  Interestingly, between the 4 and 9 ohm taps gives you a 1 ohm tap (riding
on top of the 4 ohm signal.  That should translate to an output impedance of
0.4 ohms.  DF is constant at 2.5.  Might be an interesting test.  More notes?
I find the 6688 is so short it doesn't stick out of the chassis.  How to
change?  The PA-10 monoblock weighs in at 15 pounds.  I need some sort of
inline connector for power transformers, so the top cover of PA-10 and TR-10
can be removed.  

Feb 15

Ha!  I told you I would prove it.  My tricks worked.  Splitting the loading up
(separation of variables) allowed me to re-tune the interstage tranny in a much
better fashion.  I realized this morning that I can use series damping as well
as parallel.  So I did that for the 74kHz resonance.  I used the input
capacitance of the 2A3 itself as the snubber capacitor, then upped the grid
resistor to 10k.  Seems high, but not really.  That's very low for dc.  Anyway,
that ran great, and I could then damp the 26kHz resonance with parallel 
resistors.  This time, I loosened up a little ad used 47k.  I also added a 1.5k
in series with the primary.  This doubled the output impedance of the driver
stage, from 1.5k to 3k.  But hey, it balanced out the edge injection from the
interwinding capacitance.  Gone are all the ripples and squiggles.  I did not
have to damp so hard, actually resulting in higher overall gain!  The amplifier
output is now completely tame and handles overloads nicely.  I also got the 
power bandwidth out to 25kHz.  So that is reasonable.  I'd like to get more, so
will later do some experiments with the other tranny. (hard to tell out of
circuit).  Anyway, here's the new response.

PA-10 5kHz square wave
PA-10 full power sine (6W)

A few more tests showed the onset of clipping is 6W, but in overload can reach
13W.  Input sensitivity is 0.93Vrms.  Bandwidth from 6Hz to 25.5kHz.  Output
impedance still 3.5 ohms.  Distortion (measured the old analog way) came out to
0.12% at 1W and 0.43% at 5W.  That's incredible!  Full power tube amp with no
feedback and it's still at a half-percent.  Fabulous.  That easily surpasses 
the CYMBAL.  Input ac power is 95W for roughly 6% power efficiency.  Please 
don't compare that to your average class-D amp.  Then again, here's the
clipping with a 6dB overload condition (1.8V input).  Can your amp do this?
Beautiful symmetry into grid current class A2.

PA-10 clipping

Ok, now the interesting stuff.  The "harmonic balancer" operation of the old
western electric 93A.  That's the value of resistor that sets the ratio between
parallel single-ended and pure differential operation.  The CYMBAL is all 
differential.  Well, I figured the resistance value would be about 250 ohms,
based on my unsolved theory and what WE did.  Well, first proto showed a low
value of about 10 ohms to be best.  Ok, so I built the machine at 50 ohms.  
Here's what the spectrum looks like (I used the unbroken channel on my EMU):

PA-10 distortion at 50 ohms

Not bad.  So let's see what happens with I raise the value, shifting towards
differential mode.

PA-10 distortion at 1000 ohms

Whoah, the 3rd comes up and 2nd goes down.  Ok, that's the wrong direction!  
How about 10 ohms?

PA-10 distortion at 10 ohms

Very good.  Let's lower it to 5 ohms.

PA-10 distortion at 5 ohms

Aha!  The power supply noise just jumped 10dB or more!  Ok, so I decided to
compromise and run as high as I could and still get low 3rd and low 60Hz.  That
was 15 ohms.  I'm pretty happy with this result.  Not sure why noise floor came
up, maybe I moved some wiring.  This is my hand-built wein-bridge generator
into a pot, the amp, a 10x scope probe, into my hand-built probe buffer, into 
the EMU.  Wires everywhere scattered across two benches.  Anyway, this is not a
bad result.  Distortion digitally measured is 0.11% at 1W.

PA-10 distortion at 15 ohms

Ok, so why does that 3rd stay up there so high?  If I could only drop it 10dB
this would be the best amplifier on the planet.  Hmmm.  So I connect the raw
generator input and measure that.

Generator distortion

Maybe?  It is good for 0.002%, but I see some 3rd order in there.  I wonder if
this is too low to matter or if it contributes.  Mostly, I think the problem is
a matching of 2A3 tubes.  The circuit runs absolutely fine with mismatched
tubes, however I may find that matching them on an ac basis will lower the 3rd.
That's a more likely answer.  All I need now is to buy some matched 2A3.  These
tests were all done with the cheap current production Sovtek 2A3.  Imagine how
good some RCAs would be!


Closed it up with top and bottom covers to see what effect putting the power
tranny is it's proper location (closer) would have.  Well, very low noise, buzz
and hum.  More than acceptable.  In fact, I'm thinking I have enough headroom 
to try and move the RCA input to the back panel again.  I'll throw in some
tracks on the board that I can jumper in, thus giving me the choice of front or
back.  No need to do another proto.  I am ready for production.

One thing I noticed, is that these Sovtek tubes are microphonic.  And I am not
using chassis mounted sockets, they are on the board, which flexes and gives a
nice suspension.  In fact, the music itself stimulates the microphonic.  Turn
off the music and the "shimmer" hangs in the background, decaying in about 5 or
10 seconds.  Ok, this is way down in the noise floor, with my ear to the
speaker.  But still, I expected better.  The driver tube is fine.  I really 
ought to get me some RCA black plates.  Better give Stu a call.  Also can hear
microphonics on my RB300 tonearm.  At high gains, just touching the arm (while
not in groove) will get into the cart.  I'm seeing how everything is
microphonic.  I've had cables where you can clearly hear them being handled.
Must pay attention to these sources of smear.  

I can't wait to build the stereo pair.  Having one channel sucks.  You can't
tell anything.  It sounds lifeless and thin.  Although, I can hear an
incredible openness and clarity on vocals, even if tonally off.  Shoot, gotta
also get me a SS amp to for A/B comparisons.  My Hafler ain't working.  Either
I fix it or buy something else.  A good reference is nice to have.  

Feb 14

Still working business plan stuff.  Yeah, I read all the mags.  Trying to write
down my unique selling point, customer description, marketing strategy, all in
simple short sentences.  You know, they say you need an 11 second "elevator
pitch."  What do you do?  Bam!  You blurt out a concise 11 second description 
of your business that makes perfect sense to anyone, grabs their attention, 
and makes them want to throw money at you.  Well, so far I got this tag line:
"Hi-fi for those who want the absolute best but can't afford it."  Bam!

Ok, maybe that doesn't make sense.  But you know what I'm trying to say.  Now
start throwing money...


People ask me what do I do.  I say "I design and manufacture hi-end audio
electronics".  Usually I get a blank stare.  Then I add "My company builds
stereos like Ferrari build cars".  Oh.  So here's what I came up with for the
seven sentence business plan.

1. (action) "Visit website and sign up for a test drive".  (the best way to
sell these things is to see and hear them in person)

2. (advantage) "The best audio money can't buy."  (this one needs work)

3. (customer) "The demanding audiophile, especially analog lovers."

4. (marketing) "Internet forums, google, reviews, shows."  (I need to add in

5. (niche) "Performance, masterly engineered electronics with cool and useful

6. (image) "Hagerman is the premier circuit designer in all of audio."
7. (actual) "Down-to-earth engineer."

Does this make any sense to you?  I have a hard time with it.  Trying to learn
marketing is like trying to understand women.  This magazine article was useful
in that it really tried to narrow your focus.  Go very specific.  No
generalizations.  Zoom in on what you do, your market advantage, your customer.
Mostly, I think the best tidbit I got was at the end, where it said the image
you present is huge; make nice color brochures.  


Not having much luck with package design.  I'd like to not have a huge box.
The chassis come in two sizes, 10x7 and 10x10.  Height is normally 5" with
clearance, but the TR-10 comes in at 6" and PA-10 at 7".  So perhaps a 12x12x8
box?  The PI-10 and SU-10 go in the small 6x6x6 boxes I already have.  Ok, so I
check out the prices of foam, the kind you can cut up and glue into various
cushion shapes.  Well, it works out to about $5 per square foot of 1" thick
stuffing.  Double, by the time I get the material shipped out here to Hawaii.
Hmmm.  So what about those nice expandable foam in place bags?  They're called
instapak and go for about $5 in a 12x12.  Would need two of them.  I think they
take up too much space, though.  Too costly.  Do I just stick to chewed up 
paper from shredder?  Virtually free.  


DA-10 back on the bench.  Seems I can induce a little buzz here in the shop,
where machine is dead quiet in listening room.  Could be nearby equipment, or
the variac I am powering it through.  The key, however, is that the philips pq
tubes are noisier than the RCA.  So this is somewhat tube dependent.  I wonder
if they are affected themselves by nearby magnetic fields.  Steel pins or
plates?  My new HAGDAC inductors should be here soon, I can then test those.
I don't think the diode change did anything.  The good news is that I did a
new distortion test.  Using the same 1kHz flac file, I put the output on my
trusty HP333A.  Well shux, distortion at full scale was a mere 0.08%.  Now that
is much more like it.  Not the 0.5% the damn EMU told me!  Shoot, I wonder if
I blew out my EMU.  Anyway, ignore ALL of the previous distortion plots in this
blog.  They are suspect.  It might also explain why I could not find the null
in the PA-10 harmonic balancer?  Also full scale output was 1.8Vrms.  Hey, not
bad, considering the CD spec is 2V.  So the 7788 was a good choice.  That's
easily more than 20dB headroom.  BW of 8Hz to 410kHz.  Speaking of transient
performance, here's a shot of a digital 1kHz -3dBFS square wave.

DA-10 1kHz response


So I finally fired up the PA-10.  Shazam!  It worked.  I started with no tubes,
then just the 6E5P, the the 2A3, then the 5AR4.  Each time, check ac and dc
voltages.  Only one minor mistake, swapped positions of a 22k and 100k
resistors.  Easy to do when there is no silkscreen.  Anyway, it runs like a 
charm.  All of those changes came out right.  Perfect voltages, bias, etc.
Well almost.  The output current fell to 46mA.  I want 50mA.  So I'll just 
change the 5.6V zener to 6.2V.  I forgot to measure distortion and sensitivity,
but other numbers came out to 85dBA SNR (it was really quiet) ref 1W output,
6W output power prior to clipping (it does about 12W peak), 3.5 ohm output
impedance, 9Hz to 22kHz bandwidth.  I am still limited by that interstage
tranny.  But the important thing is that I got the new self-balancing and
harmonic cancellation architecture to work.  So here's some plots.

PA-10 5kHz square wave
PA-10 5kHz full power

The square wave is a mess, very much bandlimited with some squiggles.  The
sinewave at full power is nice.  I get 20Vpp into 8 ohms.  Yeah, only 6W.  I
think I can edge this up a little by fixing the bias.

So what to do about that bandwidth?  I looked online at the Lundahl stuff.  The
LL1635 looks pretty nice, but is a large bulky package that probably won't fit
into my tight space.  Perhaps it I make a bracket for it?  Better yet, let's
just fix what we have.  Got out the spare tranny (low nickel version) and took
some measurements.  Hmmm.  700 ohms on one secondary and 800 on the other.  The
leakage inductances weren't matched either.  I hokked up to a pulse generator
and fed primary through a 1k ohm resistor, approximating the driver tube.  I
got the same messed up response, but not as rolled off.  A sinewave sweep
revealed resonances at 26kHz and 74kHz.  Huh?  There should be just one.  Then
I remembered earlier I had found different damping resistors per secondary.  I
did that and things were getting better.  Applying finger pressure on the leads
simulating tube capacitance added the rolloff effect seen in the scope photo
above.  Hmmm.  Swapped primary leads.  Hey, the glitch is now reversed.  Doh!
My eureka moment.  I was driving the primary unbalanced.  One side had the
signal the other was grounded.  Well, the interwinding capacitance to the
secondaries was off to one side, and hurt one more than the other.  Added an
extra 1k on the grounded side of coil.  Bingo!  Now things are balancing out.
Losing some gain, but this is a good area to play in.  I finally realized that
the tranny is constructed with primary wound first, then one secondary, then
the other on top of it.  Hence, the outer secondary (even though it has the
same number of turns) is longer, resulting in the higher resistance.  Balancing
the interwinding capacitances gives better symmetry with the transients.  Also,
the outer coil is less coupled to the core, giving it the higher leakage
inductance.  That's why they have two different values for damping.  One likes
to see 47k, the other 22k.  Interestingly, using the values for inductance and
frequency, I ran my snubber calculator (from the famous paper) and came up with
values of 18k and 21k.  Very close to the experimental 22k.  Finally, the last
trick is to greatly increase the grid resistor.  This will decouple the load
capacitance from the secondaries, hopefully allowing them to pass much higher
bandwidth (the driver stage is good to at least 500kHz).  This separates the
two rolloffs, making it easier to tune.  Hey, nothing so bad about a 10k or so
grid resistor.  Two 40kHz rolloffs in series is way better than one at 20kHz.
Note, this is the exact same trick I pull on the BUGLE, with the 316 ohm
output resistor (keeps interconnect off the feedback).  Really, two rolloffs
are better than one.  I will prove it tomorrow.

Oh yeah, then I hooked the PA-10 up to a frankenvandy.  Almost zero audible
noise.  I had to get my ear about 3" away to hear it.  One problem solved.
Played some music, not bad!  I hate listening to one channel, especially when
signal is not mono.  Usually sounds very odd.  But this was ok.  A little thin,
but plenty bass, very clear vocals.  So it is going to work.  


While doing a training run, came up with this idea of selling practice boards.
Such would be very useful for the DIY novice or anyone to practice their
soldering skills.  Doh!  I've been cleaning out old stuff, and realized that I
have bunches of bad boards.  I can't use them or sell them.  For a working
machine, anyway.  I will add this to the order menu.  Practice boards for $9.
Free shipping.  Your choice of large (like a CORNET2) or small (like BUGLE).
Practice for big or small parts.

Feb 13

Redesigned the switching topology of the PL-10 today.  I'm using hermetically
sealed telecom relays (silver contacts) for input and gain switching, thus
controllable via CPU.  Being gas tight makes them last over time without any
corrosion.  Anyway, the new style has maximum flexibility for choice/gain etc.
It uses a LOT of relays, as I wire them up in parallel, not series, for minimal
impact on signal.  Worst case is 4 contacts (mc phono), best is 2.  There is
input select, gain, mono, and phase.  I have as many set to compensate on/off
for virtually constant current load on power supply (one goes off, another 
goes on).  Minimum is 4 relays, maximum is 5.  The oddball is mono.  It is a
single switch to connect channels.  Although, if I were nuts, I could add a 
dummy load to compensate.  Loading, volume, and mute are by vactrols, and I 
think EQ will be too.  The phase relay coils are wired opposite, so it doesn't 
matter which polarity is selected.  Yeah, these little things matter.  You
focus on all the tiny details and they begin to add up.  Each one alone may
have no sonic impact, but together they do.  

And now I turn my thoughts to packaging.  I need something a little more
professional than a used DigiKey box.  Problem is that I have two chassis sizes
and some have trannys sticking out top or back.  Sitting in traffice I got an
idea.  Just copy what they do for a DVD player!  I can use one generic box size
and use foam inserts to hold box in place.  Ok, so they won't be custom molded
styro inserts, but I think I can cut shapes and glue together to approximate
the same thing.  Thus, all chassis will fit into same box.  Boxes get lots of
cushion on all sides.  I do not have to double-box.  The inserts are skeletal,
leaving room for tubes (I ship uninstalled), a power cord, manual, etc.  Hmmm,
that means I should also put a 1-page installation sheet (insert tube A into
socket B), with the standard product warnings (do not use in bathtub).  Yes, I
can make it look very professional, exactly what you folks should expect.  Me
also have that stencil so I can paint the HAL logo on the side of box.  

Feb 12

Finishing up TAS last night, I read the factory tour of Vienna Acoutsics, and
it says they sell 20,000 pairs of speakers per year.  And that makes them a
"medium-sized" audiophile company.  Shoot, what does that make me?  I sold less
than 200 boxes last year (not counting HAGUSB) of assembled items, not half-
kits.  So they outsell me 100:1.  Only 50:1 if you include the HAGUSB.  I'm ok
with that.  I don't need to be a large corporation.  It is much better to be a
micro company that delivers solid value and craftsmanship.  

Buying more parts today.  Digikey, Mouser, Lundahl, Cinemag, Sears.  Yeah, have
to get a new timer for the dryer.  Production is caught up and still humming.  
Just need to get back to assembly of the HAGLABS machines.  One more CHIME to
build too.  Should be able to get that PA-10 fired up any day now.  How long
can it sit there waiting?  Honestly, I spend most of my day with current
production.  Not that I am complaining.  Just very hard to move forward fast
with new developments.

Meanwhile, a friend suggested maybe tweekgeek as a manufacturer representative.
Hey, I'm open to it.  Do I really want to sign up dealers myself?  Or get 
someone like Jeffrey at HighWaterSound to do it?  I've been keeping my options
open.  So what do you think, is this a good place for me?  Would you buy from
this guy?

Tweek Geek

Feb 11

I woke up in the middle of the night with a new solition for the PL-10.  At
least in terms of phono input architecture and front panel control layout.  I
moved the INPUT from way off to the left to the middle, so it is next to VOLUME
as the main controls.  This puts the most used functions right in the middle.
Easier I think to bounce back and forth.  Then on left side are the input
specific controls that tune each source (GAIN, LOADING, EQ, etc.).  Off to the 
right are the output related controls (MONO, PHASE, etc.).  I'm not sure where
to put BALANCE.  Seems to me it might be nice to make it source specific, but I
got no more room on that side.  I suppose I can wire it up to operate either
way, as it is just a software change.  So I updated the panel designs this 
morning and it all works out nicely.  Better fit, more sense.  One compromise I
made was to have only one set of output jacks.  I removed the MONITOR jacks
because I thought it would lead to sonic degradation (they were unbuffered).  
In return, I switched to 3 phono inputs.  I am also making them all generic.
Instead of one MM and one MC, I am making them all selectable.  Each phono
input can be routed as MM, MC low gain, and MC high gain.  Loading for each is
independent.  The other compromise was that this leaves room for only 3 line
level inputs.  Ok, so I can't be everything to everyone.  This is an ANALOG
lover's machine.  I can't explain why yet, but it will become obvious once you
start to use it.  An analog guy can run an MC cart, a MM cart, a mono cart, a
DA-10 (digital input), an FM tuner, and a tape machine.  What else do you need?
Oh yeah, SACD.  

Feb 08

Oh yeah, baby.  After I finished messing around with the DA-10 rectifiers I put
it back together, slapped on the new back panel (re-wiring the RCAs), and put 
in some new 7788 tubes that came in from ebay.  Buttoned it up and put back in
listening room.  Well heck, for some reason the noise floor dropped out.  It's
gone.  No faint buzz, nothing.  Super quiet.  This must be about -100dBA.  The
inductors are presently not in the HAGDAC.  Just the RC filter seems to sound
the same.  So was it the new tubes?  New panel?  Something about the way the 
input jacks got wired?  Could be the new diodes I put in (from TRUMPET).  Will
have to experiment to find out.  Good to know I don't have to mess with layout.
Also, futzing with chassis mounting, board size, etc., I got the suspension
bridge anti-vibration principle operating again (originally in TRUMPET), and 
the tubes stay quiet.  I have now surpassed the CHIME in every possible way.


Or did I?  Finished building the CHIME, so I plug it into my system.  Do a
quick A/B with the DA-10.  Damn, the CHIME sounds just like the DA-10.  I mean,
this is pretty close.  The difference is that the bass is more pronounced in 
the CHIME (I have no idea why), and the DA-10 has better retrieval of that
microambience stuff.  Not bad, considering the CHIME had no break in.  So yeah,
the DA-10 has the edge in neutrality and small signal recovery, slightly lower
noise.  I guess that makes a CHIME half-kit the best deal in digital audio?  I
would put the DA-10 up against anything out there.  Very musical and
non-fatiguing with superb, realistic detail, without that hyper-etch common in
most digital.  The only advantage the big $20k machines have are custom digital
filter algorithms.  


Tonight I hooked up the spectrum analyzer and ran a full scale 1kHz 16 bit tone
into the DA-10.  Here's what it looks like.  It's about 0.45% distortion, all
2nd order.  There are remnants of 3rd and others, but they are -90dB or under.
Keep in mind, the smallest signal you can get off a CD is -90dB.  And this is
for a peak signal, normal music will have much lower 2nd harmonic.  When I get
if figured out, maybe I can see what a -10dBFS signal looks like.  

DA-10 distortion

Actually, I'm suprised the number is so high.  The CHIME is at 0.03%.  But that
was just measuring the output stage.  I'll re-measure the DA-10 output stage
alone to see what it is actually doing.  Maybe I have a problem with my test
setup, the PA-10 showed the same exact 2nd harmonic when I was testing it.  I
trust my analog test equipment a lot more than this fancy digital FFT stuff.


I just discovered a CORNET MC on the shelf.  I thought it was a box of packing
peanuts!  Will auction it off on audiogon.

Feb 07

The DA-10 caps broke in really nicely.  Everything smoothing out, details
cleaned up.  I hear a change in bass, though.  Not as emphasised and more
tuneful.  Maybe that wasn't the caps.  Anyway, the ambience and low level
signals really take on an aura now, at least with headphones.  I wonder if that
will translate into soundscape with speakers.  Should be able to finish the 
CHIME tomorrow and get a comparison.  Now I am finding some not-so-good tracks
are becoming listenable.  Like hip-hop.  Seems to me that genre is to poorestly
recorded in terms on audiophilia.  Dry, stale, compressed, way too loud.  I
bought a couple just to try.  The Fort Minor tune is ok on a car radio, but
nowhere else.  Until now.  It is actually listenable on the DA-10 (if you don't
crank it).  It's not that the track sounds good now, nor is it enjoyable, 
musical, or have any emotion.  I'm just saying that I can sit an play through
the entire piece without my ears bleeding.  Ok, so crud does not turn to magic.
It ruthlessly exposes the track for what it is.  I can hear exactly what the
artists are trying to do.  It makes sense.  That's all.  I'll probably never
play it again.  

I'm also thinking that maybe I can try different diodes in the heater circuit.
They could be the source of some of that noise.  If I can clean up the diodes,
I clean up the magnetic flux, and hence much less spray onto the other
circuits.  Am using the MBR1100 schottky diodes at the moment.  But still, even
these things are not turning off.  It's that choke!  You end up with a period
where both diodes conduct around the zero crossover.  This causes a
discontinuity in the voltage waveform of the secondary.  The flux in the choke 
is fine, it's the tranny that ends up with the hiccup.


Ok, so I did a simulation of the rectifier circuit.  In SPICE, I was unable to
accurately recreate the waveforms, but it did let me play around with various
parameters to see what might make things worse.  Doing so identified the 
transformer secondary coil ESR as a culprit.  Basically, in a choke-rectified
dc supply we have diodes feeding a choke, not a capacitor.  It looks basically
like this:

Input circuit

In real life using a DA-10, I measure the B+ circuit.  This has ESR of 240 ohms
in each tap, a pair of UF4007, L of 9H with 300 ohms ESR, 200uF capacitance, 
and roughly a 30mA load.  Tranny puts out 250Vct, B+ is a little over 100Vdc.
Here's the voltage on one of the diode anodes:


And then the diodes cathode (inductor input).


You can see the glitch of concern.  The question is how to get rid of it.  Can
be a source of buzz.  This is a solid-state diode phenomena, as they have
finite storage times, where even with reverse current the don't turn off right 
away.  Schottkys are generally better.  In fact, I think I might some of the
Cree silicon carbide types.  Anyway, what we see here is about a 20V
discontinutity.  This happens when the off diode starts to turn on.  Now that
inductor is pulling a roughly constant 30mA, no matter what the tranny voltage
is.  So as the "on" diode swings negative in voltage, it still delivers the
30mA to load.  In fact, the load is split between diodes.  The "off" diode
turns on right away.  For roughly 0.6ms we have both diodes conducting, and the
inductor voltage stuck at about ground.  What we need is a way to get the one
diode to turn off faster.  The high tranny ESR makes this worse, as it is
unable to switch current fast enough (and still supply the 30mA load).  One
solution is to use a heavier duty tranny.  I tried inserting capacitors in
various location with no luck.  

One possible solution I found at the izzy wizzy site.  He uses a balun in
front of the diodes.  This might be really clever.  It could help snap the
diode off by preventing simultaneous conduction of both diodes.  

Izzy Wizzy

I tried making one from a toroid I had lying around, but wasn't able to get
enough turns on it to make a difference.  So I will buy one of the commercial
line filter baluns and give it a go.  This could be a very elegant solution.  
Should work with heaters too.  


So maybe you think I'm a little off my rocker, going overboard chasing down
ghosts and phantom demons.  Hey, it's just a little squiggle on some signal in
the power supply, far away from any audio circuits.  And you may be right.  But
this may end up as progress.  Who knows.  Diligent pursuit of excellence,
nothing more.  The envelope gets pushed in all directions, not just forward.
Look at the PA-10, it borrows ideas from the 1930s.  My new remote control aims
forward going it in a way nobody else ever has.  

Anyway, speaking of the PA-10, I got some more photos.  Pretty much got the
board completed, just need to wire up the power tranny.  Hey, this thing is 7.5
pounds without chassis or the power tranny!  The DA-10 came out to 9.0 pounds 
in full dress.  

PA-10 top
PA-10 bottom

Feb 06

Reading stereophile again last night (yeah, always doing market research) and
checked out the review of the BAT Rex preamp.  Wow, that is a LOT of tubes for
just a preamp.  Then at 6moons this morning I see the link to the new Thorens
pre, 12 tubes.  Yikes!  Maybe that's the other part I am missing?  It's not
just about how much metal one can put into a chassis, but it seems the total
number of parts is important.  Products are getting more and more complicated,
perhaps to justify the price?  Shux, and I'm trying to make everything simpler.
Doh!  Rex, which means king in latin, is "...an eighteen tube powerhouse that 
combines simplicity...", and I was thinking my LS-10 design with one tube was
simplicity.  The boat has left without me.  I think I'll stick with finesse.
And that other word, eloquence.


Found the source of hum/buzz in the DA-10.  It's the inductors in the I-V
filter.  They are picking up a magnetic spray from the power supply.  Not
exactly sure what part, but must be near the diodes.  I was able to measure
current through the heater chokes, and it is a really nice smooth continuous
waveform (almost a 120Hz sine).  So theory is holding up.  Ripple voltage is
about 10mVrms perfect sine.  Same for B+, a 1mVrms perfect sine with ac line
wander (rejected by CCS).  Outputs are solid without the wander (under 1Hz
stuff).  Since one channel was much quieter than the other, I unsoldered the
inductor in the I-V filter and turned it around, hoping the new orientation
might do something.  Well, yup.  It did.  Now that channel is the quiet one.  I
get about 85dBA SNR.  Listening, I can only hear a faint buzz in the one
channel if I crank it way up.  So I am close.  The good news is that the
layout is fixed!  The only noise issue now are the filter inductors.  So, I'll
look around for better shielded ones, or perhaps try out a simple RC filter
instead.  Cut a cover, and now unit is on repeat burning in.  Microphonics no
longer an issue.  I checked SPDIF connection, it was perfect.  Everything else
is right on the money, except that the LO and HI LEDS are a touch brighter than
the CLK and DAT ones.  Easy to fix.  

Ok, I found 4 alternative 470uH coils to try.  All shielded.  This cool filter
has a bessel response and starts to rolloff at 35kHz.  Alternatively, I could
remove the 15nF, replace the 470uH with a 1 ohm resistor, and then I get an RC
rolloff at 32kHz.  It will be lowest noise.  Anyway, all I have to do is fix 
the HAGDAC.  Another possibility would be to move part of this filter onto the
main board, use a toroidal inductor style.  Nah, I don't really want to do that
as it is better to keep the DAC HF loop currents local.  So I guess I am ready
to order production DA-10 boards.  Cool.  That is good progress.  Designs done
for HA-10, DA-10, SU-10, PI-10, CO-10, and TR-10.  As soon as I get this PA-10
licked (hopefully next week), I'll be able to kick off a launch.  LC-10 can
wait.  I think I'll kill the FL-10.  Nobody is asking for it.  The again, I
bought a sheetload of 2SK389 JFETs at top dollar.

Oh bother, the things I have to do to launch.  Need to work out packaging,
labelling, boxes, shipping cushions, product manuals, photos, website, ads,
ordering, measurements, build production units (not protos), determine pricing,
parts lists (mostly done), buy long-lead inventory, make production tooling,
assembly process, part bins, painting, cnc programs, send out review samples,
sales literature, and sign up some dealers.  So, I guess I build one perfect
example of each, take photos.  Then I code website and write manuals.  Take it
from there.  Still aiming for 4/1/8.  Then I have 5 months to make a PL-10 and
LC-10.  The big show is waiting, and the date is fixed.


I realized last time I tuned the SU-10 I had a 100pF cable loading the output.
Well, now that we have proven this is a really bad thing sonically, I re-did 
the measurements with no capacitive loading, other than a scope probe.  The
transient peaking went down as a result.  The tranny was happier.  Therefore,
I did not have to load it as heavily on the secondary.  So this time, I loaded
it as light as possible.  The MED and HI gain settings are always similar.  
This time I got 37k, 110k, and 80k.  The MED setting actually needed less
damping than HI.  I think this is because I don't parallel the primaries for
MED.  So I rebuilt my personal unit with 34k, 100k, 100k.  I'll try it out
tonight, see if lightening up a little opens up the dynamics.  

And so I start to crunch the numbers, but think I won't use exact values.  I'll
run 39k on LO, double it to 82k on MED, and double again for 160k on HI.  Why?
I dunno, it just seems right.  There is so little peaking on the HI setting
that even at infinity it is ok.  And I really don't want to clobber the
dynamics, so let it sing.  Don't hold it back.  Unload it.  Anyway, I now get
cartridge equivalent loadings of 363, 132, and 44.  Hmmm, those are almost in a
3:1 ratio.  Gains ended up at 16dB, 22dB, and 27dB.  For primary loading I'll
use inf, 392, 133, and 44.2.  I think this will give me the best compromise and
all around performance.  Makes me wonder, do I need to sell short cables with
it?  Or build them right into the unit?

Feb 05

Working on building up the PA-10.  Short a number of parts, so placing more 
orders today.  Besides, I need more HAGUSB boxes.  Everything going together
nicely.  Good fit.  I rotated the board mounted output tranny for better wiring
and to minimize coupling with the power trannys (273DZ not shown).  It's a good
start.  I can't wait to get this design nailed down so I can fire up my audio
system again.  All analysis lately is using CASTANET.

PA-10 top of board
PA-10 bottom of board

I wonder how many of these I will eventually sell.  Bought way too many boards
for the CYMBAL.  Hardly sold any.  Very few people want to fork over this sort
of cash for so few watts.  I'd have to fork over two grand for a minimal sized
production order.  Or, I could buy them two at a time, in the style you see
above.  No soldermask or silkscreen, but otherwise the same quality.  At $180 a
pair it ain't cheap, but at least I only have to order what I need.  Seriously,
I just threw out about two grand worth of blank CYMBAL boards because I knew I
would never sell them (it made more room for storage of newer items).  The
PA-10 are pretty much for me.  It's a breakthrough product from a technological
standpoint, but I wonder if it can compete in the marketplace.  It's not really
what people are asking for.  Anyway, not really a big deal from a business
perspective, as I am not depending upon it for revenue.  I realize most of my
sales will be phonostages, linestages, and dacs.  I don't need to fool myself.


I'm wondering what I can do with those hammond chokes and trannys.  I noticed
that after many years of storage in Hawaiian salty air weather, the brackets
rust (see DA-10 photo below).  Maybe I can dip them in paint?  Or another coat
of varnish?  If I dip them, how will they dry?  With a point coming off the
top?  Or maybe just a good spray will do.  I guess I can experiment with spray
paint prior to assembly.  Just need to worry about losing track of what the
choke values are.  


Ok, got the new DA-10 fired up.  Everything came up perfectly.  All supply
voltages were spot on.  Pulled out the HAGDAC board so I could inject a signal
generator.  Bandwidth of tube section was 8Hz to 410kHz.  Super fast, clean, no
ripples.  Heck, this thing could put out 26Vrms, huge signal swings.  That's
20dB more than it will ever bea asked to do.  Noise was way down, almost
acceptable.  I'm still getting a little bit of buzz in the left channel.  Works
out to 0.5mVrms unweighted, or 72dB SNR.  In normal use, I couldn't hear it
until very high listening level.  I'll see if I can knock it down some more.
Will also do a battery of complete lab tests.  I haven't even looked at USB or
SPDIF or any digital stuff.   Listening, whoah.  This machine is ruthless.  Ok,
that was my first impression.  Zero burn-in.  I'm trying out these new caps per
recommendation from Romy.  After some time it got better.  Or maybe I just got
used to hearing more.  I'm not sure the bass is quite there yet, but mostly I
hear more ambience.  That's all I can think of to explain it.  A larger sound
image (I'm using headphones).  More background stuff.  Same tone and character
as CHIME.  In fact, I am building a CHIME now for a customer, so will be able
to A/B the two.  Anyway, here's a shot of the 10kHz step response.

DA-10 step response

I guess I should have posted the 100kHz response, as you can't tell anything 
from this photo.  Except that anyone who says tubes are slow and rolled off are
full of it.

Feb 04

Ok, new batch of C2 boards on order.  And it got me thinking.  What can I do to
jazz up the older kits?  I look at the circuits for the CLARINET and CORNET2
and they really don't need fixing.  Not really possible to convert them to the
more upscale and expensive TEN series.  And we already know that with component
upgrades they can be really killer.  No, what they really need is a facelift.
The boards and circuits themselves are just fine.  They are classically good 
and will not become obsolete.  I'm thinking a chassis upgrade, but not sure
yet which direction to go.  Maybe use the FPE route?  A fancy top plate like
that one guy did with a CHIME?  Something different, though, than the original
Lansing chassis.  This leads to a different paradigm - the really high-end kit
builder - someone who doesn't mind pumping a grand into a phonostage.  Hey,
maybe the design should accommodate conemag trannys.  Most of you C2 DIYers are
doing this already.  Maybe I can just make it easier.  Anyway, I think I can
extend the life of these half-kits for quite awhile.  

Feb 02

Ok, maybe I am a bit sensitive to these scams, as I have been scammed myself
lately.  And people continue to try every week.  I get a lot of foreign
addressed orders with credit card numbers that are either invalid or missing
digits.  Perhaps they assume I will just ship the item without checking?  Half
the time it is an honest mistake, a simple typo on the keyboard.  We all do it.
However, sometimes the customer just disappears into the night, never to be
heard from again.  These incidents are increasing.  Then again, maybe the rate
is constant and I am just getting more orders in total.

I am still waiting to hear back from LJ.  You owe me $900.  Are you really just
checking the machines out, or did you pull a fast one on me?  Give me a call!  
I could write up the full story, but in hindsight I look like a fool.  He did a
great job building up my confidence over time.  He even sent in a downpayment 
of $550 two months in advance.  Go ahead, deposit it he said.  Then another
payment on CC when I shipped.  It was supposed to be a trade deal with cash.  
He had called a number of times.  Well, now I realize he never used an email, 
the phone number he gave is not his.  The address I shipped to doesn't appear
to be a residence on the closeup photo from google earth (hey, it is useful).
And the credit card expired on 12/07, right after the second charge.  Great,
now I have no way to contact LJ, no way to complete the trade.  Searching
through online white pages I find an LJ with 18 addresses in the last 10 years.
Hmmm.  Sure, this could all be completely innocent.  I really don't know.

Anyhow, now ALL customers suffer from the actions of a few.  Nothing will ever
ship again until full payment has cleared.  No more special deals.  No personal
checks.  No wire transfers.  I'm sorry, but my policy has to remain diligent.
I really can't afford to get ripped off too many times.  From now on, I only
take visa, mastercard, money orders, or paypal.  Or cash.  But even that will
make me suspicious.


Now if I can get back to finishing the DA-10.  I'd like to post some photos
and show you how beautiful it is.

DA-10 top of board
DA-10 bottom of board

It is very apparent how simple the vacuum tube amplification section is.  A 
high-gm frame grid pentode strapped as triode, CCS plate loading for super
linear performance.  Tons of gain, low output impedance.  I'm running the tube
at 13mA so it can drive any cable, linestage, transformer, and/or passive
volume control.  Bottom shows the heft of the power supply.  Lots of iron and
capacitance.  Plenty of energy storage.  All passively regulated.  The bottom
is all power supply, plus two 1uF output signal caps.  If you look carefully, 
you can see where I split the ground plane.

Feb 01

As long as I am doing rants.  I finished reading the eTellig article.  And then
the guy says the amplifier can deliver high current output for 10 microseconds.
Whoah, shiver me timbers.  I am impressed.  Not really, as a 10us burst is one
cycle of 100kHz.  Yeah, five times greater than you can get out of a CD, this
amplifier will deliver a peak current above normal.  Again, this is stated with
a straight face.  

And those computer virus folks aren't the only suspicious ones.  I really have
to wonder about the ID theft hoax.  Who is actually doing such?  Seriously, if
I wanted to be an ID thief I would make a website called
www.freecreditreport.com or www.checkyourssn.com.  Can you imagine, people
don't know boo about the site yet they fill in online forms with their name,
address, phone number, and ssn.  Hello?  While your at it give me your bank
account number too.  Anyone can make such a website.  Just what do you think
they are doing with those thousands of VALID ssns you just gave them for free?

Jan 31

I dunno, maybe I have the wrong focus here.  I look at these CES photos and 
wonder what is going on.  It seems to me everyone is working on visuals.  The
most important aspect of audio is how cool the box looks.  Everybody trying to
outdo the other on industrial design.  Appearances.  So many new companies and
they all offer the same stuff but with a new chassis style or look.  Hey,
triangles!  Every possible new rounded shape for a loudspeaker.  What ever 
happened to sonics?  Is anybody working on technical breakthroughs or true
innovation?  Or is it just that shiny coat of paint next to the mirror that
reflects off of the tubes that are illuminated by blue lights in spires?  Again
I am incredulous.  

CES photos

This business is about sales.  Make nice box, sell it.  That is all.  So why am
I wasting my time trying to innovate?  Maybe I am waaaay off the mark here.  I
spend so much time developing new circuits, new topologies, and new features.
Maybe some day I will learn.  I realized, duh, it HAS to look real nice.  That 
is what you are paying for.  A luxury item.  But is it now the only criteria?

Eloquence.  Yeah, that should be my new catch word.  I liked elegance, but this
it better.  Audio components that "speak" with emotion, clarity and persuasion.
Simple circuits with short signal paths, backed by huge power supplies.

I got the new DA-10 mostly built.  Looks stunning.  I love this job.  Should be
able to get it fired up this weekend.  Have to take Friday off.  

Gah!  Just discovered I ran out of C2 boards again.  Dang, will get new batch
ordered up tonight.  Got some new packaging in for these large boards.  Was
using large padded envelopes, now I have large hard cardbord flats in white.
Seems to be working really well.  A bit more sturdy and will bend less in


Huh, just found another virus.  This time it was dssagent.exe from one of those
kid games.  I have this theory that the people who want your keystrokes so they
can futz with your life are the same ones who own the companies that sell kid
games.  It all appears normal as these are actual CDs you have to buy at the
store.  Yeah, it's called "brodcast".  Hide behind some product registration
excuse.  Anyway, the viruses really slow the machine down.  You know where to
look when things hang by using the windows task manager.  Just look at what
process is consuming all the cpu cycles.


Oh gads.  I was just reading the feb stereophile, eTellig.  Some amplifier guy
states: "when you are dealing with low-level signals, you are talking about the
behavior of individual electrons" (regards to MC phono inputs).  What a crock
of shit.  I can't believe they publish this stuff with a straight face.  To me,
this brings a serious reduction in credibility.  A 0.1mV cart into 100 ohm load
is roughly 6,000,000,000,000 electrons per second (if dc).  That is quite a few
individuals, don't you think?  This is an insult to those of us who have
actually dealt with low electron count systems.  For me, that includes CCD
camera design, where low light and odd scientific applications (astronomy)
require imaging where conversion efficiency is paramount (electrons per photon)
and both dark noise and shot noise are big contributors to the signal.  That's
why we had to cool the damn things way below freezing.  Also mass spectrometer
design, where photomultiplier and electron multiplier detector tubes generate
output pulses for each individual electron.  And I'm not talking 6 trillion.  I
mean one.

Jan 30

PA-10 board on order.  Also bought some more CineMag trannys.  Will also place
production order for foil caps (8 week lead time).  Good thing I am not
designing to predetermined price points, I would have to make compromises.  The
trannys for the PL-10 are $350 my cost.  But hey, I want to make a statement
(sonic) here, and I feel the output trannys are the way to go for this tube
linestage.  So I look at the Lundahl alternatives.  I like the 1681 ($95) for 
MC input, and the 1674 ($110) for output.  If the CineMag stuff doesn't work 
for me, I think this pair will.  Easier to mount to PCB and the windings are
much lower resistance.  Bandwidth looks to be better too.

Jan 29

I still worry about the economy.  What a time for it to hiccup.  Here's an
article from a guy who kinda sounds like he knows what he is talking about:

Money and markets

Ok, so the value of the dollar will plummet much farther.  Will that make my
equipment cheap to EU?  Or is it part prices will rise accordingly thereby
pushing up my prices too (I got no room to wiggle)?  How will hi-end audio deal
with it?  Does everyone switch over to kits because that is all they can
afford?  Do they stop going to expensive salons?  Am I nuts to try and
introduce a dealer line at this time?  I could go factory direct at lower
pricing, but without the exposure.  Or is perhaps this the future?

Buying spree


Hmmm, looking at the graphs for the output tranny, that 8k might be on the high
side.  Sure, it works, but 3k is a lot better.  I should look at power supply
changes to get the 5687 back in there.


Ok, I re-did the power supply for 2.1 amps heater and 65mA B+.  The 5687 is 
perfect at 15mA, 175V, -7V, with an rp of 2.5k.  Exactly what that tranny is
designed for!  Overall linestage gain is then 11dB or 12dB.  So I use 60mA in 
the tubes, that leaves 5mA for CCS bias.  I can always power the LEDs straight
off the tranny ac.  But wait!  I realize in my zest I forgot the damn CPU and
relays.  Sheet, they get powered by the heater string too.  Then I find I
already did the supply design weeks ago.  And it is almost identical to what I
just did yesterday, but with a heater split to +/-3.1V, and an extra 1 amp of
current.  Ok, should be fine then.  Requires 3 more chokes, but that's not a 
big deal.  Going in circles I ended up back where I started.  Good, so the
5687 is back.  Phono is 6688 and 6CG7.  Now I need to get that PA-10 layout 


Spent most of the day on PA-10.  This layout is a pita.  Very crowded with a
lot of big parts.  Big items on both top and bottom of board.  In order to get
maximal performance, I had to move the connectors to the front panel.  Then 
rotate the output tranny (which takes up the entire center of the board).
Power tranny ended up staying where it was, on top of chassis.  Moving the RCA
inputs to the front made room for the ac wiring.  I also had an issue in that
the 5AR4 tranny wires were right next to the input.  Doh!  Anyway, this 
should be much cleaner.  Wish those 3.3uF caps were smaller.  Lot of little
circuit tweaks.  Improved on the CCS design.  Still routing, maybe I can finish


Never mind.  I worked into the evening.  Layout done.  I'll get proto ordered
tomorrow.  Meanwhile, the DA-10 board came in.

PA-10 layout

New PICCOLO review just came online.  Here it is:


Jan 28

I realized I must generate maybe 30 hits a day on my own website.  It's my home
page in explorer.  So the typical 1700 hits a day is really a bit less, thanks
to me.  Every once in awhile I google myself.  See what's up.  This time I did
Hagerman Technolgy and followed it down about 15 pages deep.  Found I've been
mentioned in EDN magazine (the cult rag for EEs).  This is in Steve Leibson's
blog.  He links back to my blog.  We've exchanged several emails in the past,
and it is nice to receive advice from the former editor.  He's the reason I
ended up giving products names like "DA-10".  

Leibson's law

Ok, so I do some calculations on various triodes & such to compare for use as a
phono front end.  I already know that a 12AX7 works pretty good.  That's my 
reference benchmark.  My earlier thought was to try and improve on headroom and
dynamics by going to the 12AY7 (6072), more current, lower rp.  It was a
sacrifice in gain that might not be a good idea.  Anyway, at a reasonable
operating point of 4mA, -2V, it gives roughly a mu of 44 and rp of 25k.  Giving
myself a fixed 100V drop on the plate load resistor (merely for comparison), 
the load is then 25k (compared to 150k in CORNET).  Gain is thus about 27dB and
equivalent noise resistance of 1.5k.  Which ain't bad.  That's about what I get
for the 12AX7.  Ok, now try a 5879 pentode.  At 100V g2, -1V bias, it runs at
4.5mA with 1mA g2.  A 22k load gives 31dB gain, but at a noise resistance of
8.4k!  Aha, that's the reason pentodes get blamed for noise!  I'll probably get
similar numbers for the EF86 and 6SJ7.

So what about this 6688?  Does the frame grid high transconductance make a
difference?  Yup.  Playing around awhile, I come to a sweet spot of g2 of 80V, 
g1 at -0.5V.  Plate current is thus 5mA and Ig2 is 1mA.  Plate load is 20k.
However, the noise resistance is a mere 310 ohms!  Shux, this is 6dB better
than my triode reference.  That's nothing.  Get this, the gain works out to
48dB!  Holy cow.  This is unbelievable.  I get better noise, higher gain.  This
is like miracle performance and I am only consuming 6mA of B+.  And the darn
tubes sell for $6 or less.  I don't get it.  Why is nobody else using this
tube?  Romy uses the 7788.  I think the 6688 in pentode for a first stage is
just as good.  Superb performance at low cost.  Maybe this is the breakthrough
I was waiting for.  I can change the plate load to 22k, then the RIAA caps work
out to a perfect 0.1uF and a hair over 33nF.  That's EQ and 28dB gain with one
tube stage.  I can follow this with a 6CG7 normal triode gain stage at 8mA for
22dB gain, 5k output impedance.  Will this be low enough to drive the 50k to
100k attenuator of a PL-10?  I hope so.  No cables to drive.  I end up with a 
50dB phonostage in 2 tubes.  Maybe the gain is a bit high.  I'd like to be 
about 6dB lower.  No big deal.  There may be losses I haven't accounted for.  
The linearity of the curves looks really good.  I don't anticipate a distortion

Hmmm, what about the 6688 in triode mode?  I found these curves out there in 
cyberspace.  Also discover the 6J9P is supposed to be a Russian equivalent.

6688 triode curves

An output stage using such would be roughly an Ro of 3k at 31dB gain.  Not bad.
Almost as good as the 7788.  I wonder if I can use the 6688 in the DA-10
instead of the 7788.  Oh wait, looks like B+ is too low for that.  

Now I wonder about dc biasing.  How to set and stabilize the g2 voltage?  I
could borrow from the self-biasing of a triode, use a large R into a C for G2.
If fed from B+ the 220k gives me 80V from a B+ of 300V at 1mA.  Is that stable?
A little negative feedback would make me choose to put that resistor on the
plate instead of B+.  Then maybe the tube will find it's proper setpoint a
little more consitently, at a desired current.  Too much Ib and g2 drops (a
lot), thus forcing Ib down.  So maybe this is a good way to balance.  Another
subject for research.


Aha, biasing a pentode is simple.  Turns out the ratio between Ib and Ig2 is
relatively constant.  Usually about 4.  Looks like for the 6688 it might be 5.
So a 220k from B+ is just right.  All you have to do is keep the ratio of the
two resistors constant.  Then auto-bias is done via cathode resistor just like
a triode.  In my case, 82 ohms (6mA at -0.5V).  

So I think about the linestage again.  Nothing seems to click for me, until I
insert the tranny.  Yeah, a 4:1 output tranny is it.  Forget the differential
stage, etc.  Just a normal gain stage with CCS load, driving a tranny.  Hey,
that's exactly what I have on the PA-10.  HA-10 too.  With a 5687 at 20mA, it
gives an Rp of 2k, really low.  Into the tranny is comes out as 125 ohms.
Probably 150 once you add in ESR.  That is incredibly low for a tube preamp.  
It should drive anything.  Super linear, low distortion.  It also fixes my
gain problem.  For phase, I have to insert relays on the primary to swap them
out.  My choice for tranny looks to be the CM-9589.  Impressively, I am down to
4 tubes!  This is great.  A pair of 6688 for the phono front end.  One 6CG7 for
the second stages.  Then one 5687 for linestage.  CineMag trannys on input and
outputs.  Vactrols for loading and attenuator.  Relays for switching.  This
hugely complicated machine is finding simplicity.  In fact, I could use the
6CG7 as the linestage tube.  Running at 8mA it gives mu of 20 and an rp of 8k.
Sounds high, until you divide it by 16.  Then the output impedance is a mere
500 ohms.  Still incredible.  That's like cathode follower performance.  It 
could be done with just 50mA of 300V B+ and a heater current of 1.8 amps.  I'll
start looking into power supply design (redo what I originally had).  With the
5687 it would be 70mA and 2.1 amps.  Gobs of headroom either way.  With the
6CG7 I have the option of partitioning in two ways.  If I separate L/R, then I
have the phono 2nd stage in same glass as line stage.  Or I mix L/R.

You see, it's not just about the parts you choose.  It is how you put them
together that counts.


Just realized I can use the LS-10 test bed to try out the 4:1 output trannys.
Duh.  That's perfect.  It'll reduce the gain down to something reasonable.  And
I can measure performance on the bench, read distortion, etc.

Ok, did a paper design for a PL-10 power supply.  Not many options here for a
low current high voltage center-tapped B+ transformer.  Hey, what about the
Fender tranny I use in the new CORNET?  I burn off a ton of voltage for that
design.  And maybe I don't really need to operate the tubes at 200V.  With the
pentode I can drop the way down to 100V, so a new bias point at 130V to 150V is
fine.  Redesign the chain for 250V at 6688, 260V at input to amplifier.  With 
my usual LCLC design I find it takes 5H minimum, so I use my trusty 10H at 205
ohm 157J.  That's 10V drop, so first LC needs to put out 270V at 50mA.  The 
22772 tranny is 337V with 175 ohms internal impedance.  A 5AR4 is 75 ohms.
That gives me 455 ohms for input.  Or 23V drop.  Using my online web calculator
I find this string works out perfectly.  Stability tuning with 100uF.  Without
the extra resistance, the second section needs 400uF for criticality.  This is
very much like what I did for the TRUMPET, but with only only one polarity.  I
will add a 1.5k / 100uF RC filter to isolate the 6688s.  That way they end up
with a perfect 250V ultra-clean, zero ripple supply, decoupled from whatever
anything else is doing.  The 5AR4 supplies the delay.  For heaters, a single
155B 6mH choke will do the job.  I set the power supply output to 7V (then I 
get to add a RC filter in the amp of 0.39 ohms / 20,000uF) 1.8 amps.  I have
headroom, so burning off a little power I can detune the LC tank.  That is, it
won't need 100,000uF!  So add another 0.39 ohms to the 6mH choke and we can get
away with just 20,000uF.  I don't even need to resort to schottkys, so good
quiet switching diodes are fine.  This calculates to a 20Vct filament tranny.
Or a Hammond 166J20.  Final heater ripple is 10mV at 120Hz sinewave.  Proven 
good enough.  Wow, both B and H supplies, choke rectified and filtered for only
5 chunks of iron.  This power supply is elegantly small and simple.  Just the
way I like it.  The amp is 4 tubes and 4 signal trannys, 4 vactrols, and a
bunch of relays.  Doesn't look like I can do it on one board, the CPU and IR
remote stuff will require a 2nd board.  Super.  I get all of the performance I
was looking for without too much cost.  Short signal path, not too many caps.
Hey, I might be able to get the retail price of this thing down to $7k.  That's
way better than original assumptions.  

Excellent.  For the price of a steelhead I offer a phonostage that should rival
or better it in performance.  Add to that a linestage with 3 or 4 inputs.  Add
a way cool remote control (volume, balance, select, loading, mono, phase, eq). 
Finally, after two years, I have a design that is unique and innovative.

Jan 26

Took the morning off to study pentodes.  Really, I never did much with them,
always trying to tweak the most out of triodes.  But hey, maybe they have a 
place for me.  After all the 7788 is working wonders in the DA-10, and the 6E5P
is a superb driver for the PA-10 (I might want to also consider the 6CL6).  
Anyway, the buzz has it that pentodes are too noisy for phono.  I wonder.  The
6AU6 is an old standby.  Found the 5879 and 6CF6.  Still liking the 6688 and
frame grid construction.  The RCA receiving tube manual turned out to be the
best source for theoretical and practical information.  Went through six books
I had.  Huh, gain is easy to calculate, at Av = gm * Ra.  As a first stage, it
makes the RIAA equalization easier to have a pentode (or cascode), as the
output is basically a current source.  I just set my R's and C's to do the EQ.
With a 100k load, a gm of 10mA/V gives us a "mu" of 1000.  That is a lot higher
than the 12AX7.  Perhaps a raw gain of 40dB into the EQ, resulting as a +20dB
overall.  Then another +20dB stage behind it.  Actually, 30dB from both stages
(one pentode, the other triode) is another good operating point.  The 7788
gives maybe +30dB with a 1.5k output impedance.  That's a +40dB phonostage in
two tubes.  I like the concept.  Should also have way more headroom than the
12AX7 second stage.  

The noise issue with pentodes is from the "partition".  Breaking the electron
stream flow into two flows, one to G2, the other to plate.  Found a great web
page that summarizes a lot of this.  And gives a formula for noise.

Tube operation info

Great, so now we can look at noise in terms of an equivalent resistance.  What
we have is an 8 * Ig2 / gm added term.  The 12AX7 in the CORNET and TRUMPET
give me decent noise performance.  Especially with a tranny front end.  All we
need to do is operate the pentode such that it doesn't add too much.  If the
noise resistance doubles, that is a 3dB increase in noise.  I can live with
that.  So we set Ig2 = gm / 8.  If gm is 8Ma/V, then we operate with Ig2 at
1mA.  Or for a 6688, typical operating point is 16.5mA/V and Ig2 of 3.3mA.
That gives us an added term of 1.6.  And Rn = 394 ohms.  That's actually a lot
less than a 12AX7.  So maybe pentodes ain't so bad afterall.  Just need to 
implement things properly.  It's all about how you put the pieces together.

Then I got to thinking.  Hmmm, maybe there are other things we can do with a
pentode.  So many grids to play with.  We could make a low gain triode by
strapping G1 and G2 together?  Boy, my work with CRTs many years ago and the
mass spectrometer electric field simulations are starting to pay off.  It
really helps trying to visualize the physics of what is going on inside the
tube.  Anyway, had this wierd thought on a differential stage, where the G2
were crossed over to the G4 of the opposite tube.  I'm guessing this is
negative feedback (degeneration), not positive, which would cause instability.
The resulting gain would be cut down.  A linearized cross-coupled output.  One
thing I really like about the differential stages is that it eliminates the
cathode bypass capacitor.  What a great thing!  Well, I don't know where I'll
end up here.  But I would like to do the phono section in two tubes.  Pentode
looks to be fine for the first stage.  Second gain stage has to have low enough
output impedance such that it can drive the volume control load.  Without any
loss of dynamics.  Maybe a 6688p / 6688t is the right combination?  Huh, this 
is actually very similar in architecture to the ARCHIVER.  

Now I need to study biasing.  Will a pentode bias as easily and consistenyly as
a triode?  My CORNET stages are plug and play with tubes.  Any 12AX7 will fire
up and run ok in it.  I want to achieve the same with pentode.  No pots.  

Jan 25

Listened to the PI-10 with different caps last night.  I gave it 12 hours of
burn-in, but the caps were still smearing.  Not as good as the stock caps.  Ok,
so I have to start using ultimate stuff.  I'm on the verge of pulling the
trigger on production quantities of top notch audiophile caps.  8 week lead
time!  Will start with 1uF and 0.15uF.  Minimum 200 pieces of whatever value I
pick.  Still have enough PPFMX on hand to cover most of TR-10 and CO-10
production, so these new caps are for the rest of the HAGLABS line.  Haven't
decided yet on tubes and values for the PL-10.  So I'll hold out on buying the
rest of what I need.  

Presently the design for PL-10 uses too many caps.  Maybe too many tubes.  I
had made the preliminary choice of the 6072A for phono section, 5687 for line
out.  But still, not enough gain on phono side, too much gain on line.  So 
since I am having so much luck with the CineMags, I looked into maybe putting
some on the line out too!  A 4:1 does some good things.  One, it kills all that
extra gain (-12dB).  Second, it drops the output impedance by 16 (ignoring coil
resistance).  Looking at the 5687 or similar tube with CCS plate load, then 
even at 4x signal, the distortion should be low enough.  I'd like to keep it at
0.1% or lower, full signal.  Basically, this means I need to rethink the entire
PL-10 amplification topology.  Maybe a pentode on phono input?  Maybe not a
differential output stage?  


Ok, so now I have a linestage.  Pulled the LS-10 into two pieces.  Made it a
two-box design.  I cut up a CASTANET board and put it into a BUGLE PRO chassis.

LS-10 power supply

This is much quieter.  Still a huge +20dB gain, though.  But hey, gives me
something I can use for the time being.  This is just a test bed anyway.  Can
modify it later to use the 12B4.  Umbilical is hard-wired.  I built the LS-10
using the original HA-10 chassis parts.  So nothing fits.  But that's ok.  It 
is just a circuit test.  On the bench, the LS-10 can put out huge voltage
swings.  Bandwidth looks perfect, out to over 50kHz.

LS-10 in use

My listening test station changes every day.  Here you can see the SU-10 and
CO-10 proto driving the CASTANET.  The LS-10 is the box on the right.  Volume
on left, select on right.  The white plastic box on the bottom is the new
separate power supply.  I'll give it a listen this weekend.


Thinking more about using a pentode as a first tube in phono.  When used
properly they don't seem to be nonlinear at all.  So I looked around, what
would be good choices?  I came up with 6AU6, EF86, and 6688.  I kinda like the
last one.  Lots of gain.  Cousin to the 7788, which is doing me real well in 
the DA-10.  Dang, and the 6688 are cheap!  Huh, I never would have guessed
that.  The question is how to get the noise down.  I think things might work
best if G2 is kept below anode.  Say put G2 at 100V, anode at 200V, B+ at 300V.
Such would minimize G2 current.  Or maybe tie G2 straight to B+.  Hard.  Don't
let it move.  I worry about noise.  Are pentodes really that noisy?  I need to
study this.  Anyway, just thinking that maybe a 2-stage phonosection would be
cool.  It would go well with my 2-stage PA-10 concept.  Fewer capacitors.  Like
a 6688 (pentode) and 7788 (triode).  No feedback.  Passive EQ.  The output
needs to be low enough impedance to drive the PL-10 volume control.  So 3k or
less is great.  I am really loathe to add another stage or follower just to
buffer things.  That's exactly why I did the DA-10 the way I did.

Jan 24

Bunch of changes to DA-10 layout.  Discovered an error, the ground plane
extended all the way to the sides of the boards, and without the soldermask,
will connect to the side extrusions.  This makes for some unwanted ground
connections to chassis the the wrong places.  Not sure how much hum this added,
but I fixed it anyway.  Now there is a channel on the sides to fit into the
grooves.  Did a lot of ground plane changes, cutting it up into sections,
keeping supply currents isolated from amplifier.  Re-wired the chokes to cancel
flux fields, also tried to cancel the criss-crossing ground currents on the
board.  The filament supplies are spread out across each side of the B+.  It
looks nice, but causes for a lot of current to cross the plane.  Best if I keep
+6V and -6V mirrored and cancelling.  Moved some caps that interfered with the
HAGDAC.  Made some part changes on HAGDAC to balance supplies and improve PLL
operation (CS8415A part).  Added traces for output wiring.  Moved a LOT of
stuff around for better signal currents and minimal crosstalk.  Added FILTER
switch to back panel for selecting rolloff modes.  Fixed other various things.
Made board 0.025" smaller, to keep it from rubbing against front and rear
panels.  I think the tiny 0.012" gap on each side will help reduce chassis
vibration conduction into the board.  Anyway, parts on order so I can build up
proto #2.

DA-10 layout

Finished building up the proto PI-10.  I used some special 0.1uF caps to see if
I could get some improvement there.  Better chassis, jacks, etc.  It has that 
HAGLABS look.  I also put in some pots to null out the dc.  Worked fine.  But I
don't think I'll use the pots for production.  With tube phono stages it really
isn't necessary, and I don't want such in the signal path.  Burning in now,
will give it a listen tonight.  

More parts on order.  Doing my best to keep the economy humming.  Bought some
more silver/teflon wire.  The first batch made for some great interconnects.  

Oooh, just had a thought about LS-10.  I think I will build one, but semi-point
to point.  I have this large rackmount aluminum chassis sitting on the shelf
now for about 5 years.  Semi-point, as I can cut up a CASTANET board and use it
for the power supply, then a VACUTRACE adapter card for the tube sockets.  Or
maybe I just hard wire them to chassis.  Yeah.  I'll build it for the 12B4 and
take my chances.  This would at least prove out the sonics and schematic.  Gets
me partway there.


Wait a minute.  That's too much work.  I think I'll keep the existing LS-10
layout, use the 6H30, and remove the power supply.  All I need to do now is put
the power supply stuff in another box.  Baby steps.  I can modify this result 
to use the 12B4 later (not an easy pinout change).  Then I discover that I can
cut up a CASTANET board small enough to fit into a spare FRYPRO chassis I had
left in stock.  Plastic, but so what?  I cut the supply section out, then 
mount it.  Much easier than drilling steel.  And it is pretty much guaranteed 
to work, as I can make the umbilical long.  I'll even hardwire it.  And so, the
"head" portion of the LS-10 will sort of look how I originally envisioned.

Jan 23

Updated website.  Many changes in anticipation of HAGLABS.  The transition is
ongoing.  The big news, CASTANET is available assembled for $999.  FRYBABY is
now a half-kit.  Build your own.  ARCHIVER is discontinued (to become FL-10
FLUGELHORN).  CORNET discontinued (to become CO-10).  FRYKLEANER discontinued
(replaced by FRYBABY).  BUGLE PRO discontinued.  Eventually VACUTRACE, UFO, and
FRYKLEANER GOLD will migrate to HAGLABS site.  

Jan 22

More research into speakers.  It seems I am on the wrong track with the Maggys.
I love the sound, but I can't drive them.  Not even my megapower PA-20 (the one
with four 6080s running full blast) can power them.  However, the DeVore Super8
is looking like a front-runner.  It should work well with the PA-10 (what I
need for RMAF08).  

So many little things on my plate, can't keep it straight.  FLUGELHORN really
needs a mono switch right?  I wonder if it can go on back.  Adding filter mode
switch to DA-10.  Must relayout.  Aye, the PA-10 is waiting for relayout too.  
Building up a production CO-10 and TR-10.  Gotta make changes to SU-10 design.
Will the short cables help?  FRYBABY kit manual has to be written.  I took
build photos yesterday.  Made short cables.  So many parts to buy.  TR-10 needs
new switches on front.  Gotta try an get an RMAF08 room on 11th floor, did I
ask Al yet?  What about those Super 8's?  I should get the parts and put 
together that LC-10.  Ciufoli is waiting for a gif banner.  Should find better
wire for the umbilical.  Have to write new TR-10 manual.  CO-10 too.  What to
do about LS-10?  What is LJs phone #?  Is he sending back the CORNET MC or not?
Gotta add the HAL logo to FRYKLEANER.  GB is still waiting.  Clive Meakins has
a new review up on ETM.  I can't even remember which product it was.  RB needs
a new font for his logo; I have no solution.  How come musicgiants doesn't have
Ingrid Michaelson?  Should finish my FL-10 proto.  Can't seem to get pricing
from the Electrocube folks.  I need to get started on taxes.  Time to order a
new batch of boxes.  Well, at least I am keeping up with orders!

Ok, for FL-10 the mono switch doesn't fit well on back.  And to make sense from
a customer standpoint, it needs to be on front with all of the other controls.
Then I realized, the LOADING control can do both.  I just reduce the number of
choices to 100, 220, and 47k ohms.  Mono & stereo, for a total of 6 positions.
Easy to do.  Nobody was using the 500 and 1k positions anyway.


Just wrote the FRYBABY kit manual.  Four hours of pure fun.  But it looks ok.
It is uploaded, but I haven't done any website changes yet.

FRYBABY kit manual


Finally, back to listening.  Took my short 8" cables to the outputs of SU-10.
Oh my.  That was something.  Seems to have gotten me some of those dynamics
back.  More punch.  Wow, now it is getting really close to the PI-10.  The MED
gain setting (21dB) should load at 101 ohms.  It sounds closest to the PICCOLO
at 47 ohms loading.  Very similar.  The difference is now one of just weight.
And I can't tell which is correct.  The PI-10 is thinner, the SU-10 thicker.
Or more weighty.  The PI-10 is not thin, but neutral and clear and open.  Only
by direct comparison could it lack the heft of the upper bass region.  I'll 
still give the punch and slam to PI-10, but with the SU-10 really close.  Like
I said, can't tell which one is truer.  It comes down to a personal preference.
The clean solid state presentation, or the earthy tranny presentation.  One
thing I've proven for sure (and I've been promoting this on forums for years)
is that those output cables have to be really low capacitance (ie short).

This is really good sonic performace from such a modest transformer.
Technically, the TX103 blows it away, having much lower coil resistances and 
far greater bandwidth.  Superior measurements all around.  I wish I still had
some to compare sonically.  The CineMags certainly need the right care and
handling.  If not played right they will not sound their best.  Implemented
poorly, they can be really dark, rolled off, and without dynamics.  It took me
days of work to really get them to sing.  And no we know how.

Jan 21

Shux, I forgot to mention DA-10.  Got new tubes in, working just fine now.
Bias is much betterm right where I expected it.  The strong tube has almost
zero microphonics!  So the 7788 is the right choice.  I just need to find some
good ones out there.  The second tube was slightly microphonic, but good enough
for listening.  And so I did.  But not without laptop trouble again.  I always 
keep losing the usb-audio driver, as the EMU-0404 or something crashes on top 
of it.  Futzed around with all the dialog boxes.  Wish I knew what worked, as I
am back in the saddle.  The foobar/usb-audio combo is working again.  And it
sounds great.  The DA-10 has that unmistakable CHIME sound, howeverm the bass
seems cleaner and less pronounced.  Less fat?  Is it the improved USB section?
Or improved power supply?  Not sure why.  But my first impression is that this
does sound better than a CHIME.  Maybe because there is no cathode follower?
I dunno.  Happy for the good news.

And so I listened again to the mods to SU-10.  Hmmmm, very nice.  Slightly
warm, very full tone.  I think it is still missing the extra punch, although 
for some reason cymbals sound more like real pieces of metal.  Like I said 
earlier, organic.  However, it's just a tad on the bright side.  And at each
gain level.  Then I remembered I set the tuning as high as I could.  That
meant some peaking in the square wave response.  Ok, so back to the lab, found
the new parallel resistance values that gave flat tops.  No overshoot.  They 
were 19k, 36k, and 47k.  Yeah, the HI loading came way down.  A gain killer.
Still, the response for each gain is perfect in shape, Bessel-like, which I 
know sounds good.  So all calculations had to be redone.  And of course I once
again soldered in new resistors.  Haven't heard it yet, put it on a FRYKLEANER
instead.  Will get to it later.

I know I sound like a crazy man, going in circles, changing my mind, doing
everything different, not what is on the data sheets.  Fine.  Maybe I can 
explain the madness.  You see, I'm getting about 26kHz through this tranny.
That's a decent audio passband, especially with the smooth and clean response.
The problem is that MC cartridge loading does two things.  First it tames an
ultrasonic resonance (LC).  However, that resonance occurs anywhere from 100kHz
to 1MHz.  At those frequencies, the tranny is reflecting not much back from 
the secondary, so any loading disappears.  The cart has no damping.  That's why
I changed to putting extra damping on the primary.  That, plus I greatly
preserve gain.  Secondly, loading the cart with resistance forces it to deliver
power.  That is, it must source current, not just be a voltage generator.  Such
changes the sonics of the cart.  The passband loading accomplishes the latter.
Also, what I don't want to do is create a tone control.  Too much secondary
loading ended up rolling off the treble.  That's not the idea.  We want to load
the cart for current and/or resonance control (fortunately, the resonance ain't
gettin through this tranny!).  Anyway, secondary tuning turned out to be a 
game with rolloff and peaking.  I want neither.  And so I tune the tranny for
optimal transient response.  I take a hit in gain.  You could put a small cap
in series with the damping resistor, but that introduces the sound of a cap.  

Let's rehash.  Here's an approximation of the system:

Equivalent schematic

Lg and Rg are the generator inductance and resistance (cartridge).  Cp is the
capacitance of tonearm cable.  Rl is our primary side parallel loading.  Rp is
the primary ESR (52, 26, 13 ohms).  Rs is the secondary ESR of 3.9k.  I am
ignoring the primary inductance (parallel with coil) and leakage inductance 
(series with secondary).  Rd is our damping resistance in parallel with Ra and
Ca (phono loading and cable capacitance).  Now in the lab I approximated the
generators as LO = 25 ohms, MED = 10 ohms, and HI = 5 ohms.  That is, a cart of
less than 0.25mV is likely to have an Rg of 5 ohms.  Once I have the Rd values,
I can determine reflected loading and gain.

    L       M       H
Rd  19k     36k     47k
Ry  267     101     34
Av  15.8dB  21.6dB  26.2dB

Wow!  The gains are very close to the PICCOLO.  I'll round off to 16dB, 21dB, 
and 26dB.  Ok, what to do about primary side loading.  I decided I want to be
able to cut each reflected value in half.  At least in the passband.  So the
values are:

     L    M    H
inf  267  101  34
267  135  73   31
101  73   50   25
34   31   25   17

That's four settings.  Wide open is secondary loading only.  I really like how
the all-important 21dB setting gives roughly 100, 75, 50, 25 in steps.  That's
pretty nice.  And we no longer have a tone control, only loading control.  The
bandwidth of each gain remains contant.  LO measures in at 32kHz, both MED and
HI at 26kHz.  Internal wiring isn't so bad.  No need for a PCB.  Serendipity
strikes again, as now the LOADING control is on the bottom, GAIN on top.  
Exactly the same as PI-10.  These step-ups look virtually alike.  Do almost the
exact task with equal flexibility.  To my ears, the PI-10 is a little more
neutral and faster.  Pick your poison.  I like them both.  I'm curious to hear
how the primary side loading pulls the cartridge into alignment.  


Ok, listened.  I like this loading a little better.  Brightness gone.  Very
earthy in tone.  Relaxed.  Maybe laid back too much?  By comparison the PI-10 
seems more neutral tonewise, and faster.  I think the SU-10 is lacking maybe in
transients.  It lacks that hit-you-in-the-chest punch and dynamic.  Not too 
much top end sparkle, yet balanced and pleasing.  Musical with a throaty vocal.
It's very good, but not perfect.  I can't get that last bit of slam to come
through.  Am I being too critical?  It's just a different presentation.  The
microdetails and very soft passages are superb.  Ultra-quiet noise floor.  I
don't think the primary side loading is going to be that useful.  At least not
on this cartridge.  Wait a minute!  Sheet.  I think I know the problem.  Am
using 1 meter output cables.  They need to be much shorter than that.  I'll
make up some new ones tomorrow.

Jan 20

So I got the SU-10 hooked up last night.  Needed to run a ground wire to phono
to kill some buzz.  Wow, was this thing ever quiet!  I can't believe the noise
floor.  A mere trace of a hiss, waaay down.  It must be maybe 10dB quieter than
the PICCOLO, which was already perfectly satisfactory.  I had to play around
with a lot of the gain/loading combinations until I found the best match.
Seems like I stumbled upon the solution, which is LO/LO, MED/MED, and HI/HI.
That makes perfect sense from an electrical standpoint.  The lo gain needed
the 10k on secondary, med needed 36k, and the hi none (or just 47k default from
phono).  Anyway, I ended up at MED/MED to best mimic the frequency response of
the PICCOLO (which I know is dead neutral).  

SU-10 guts

Anyway, the SU-10 was a bit more organic sounding, the PI-10 more open and
clear.  The SU-10 had great tone and very realistic presentation.  The PI-10
definately faster, more dynamic, with a lot more punch.  The SU-10 was much
less in-your-face.  More relaxed.  Again, the SU-10 was way quieter.  There
were no bass problems with either.  No bottom end loss, no nasality, no
thinness.  In fact, the SU-10 had a very full tone.  Nonetheless, I found the
PI-10 to be more exciting.  More edge of your seat thrill.  Does that mean the
SU-10 is lacking?  I dunno.  The tonal character changed a LOT with loading.  I
could easily live with either of these two step-up machines.  But keep in mind
this is only my first impression, with only 1 hour of listening tests.

Theoretically, with an ideal tranny, I get a combination of gains and loadings.
Or loadings of 6, 16, 25, 36, 63, 101, 145, 252, 580.  A good range.  But that
doesn't account for ESR losses.  So I took it to the lab.  I have some photos 
of the square wave responses, but will post later, after optimization.
Bandwidth was not particularly spectacular.  At LO gain I got about 39kHz, MED
and HI delivered 23kHz.  So they roll off early.  Did not matter much between
100pF of cable loading versus a scope probe.  Not too much change in response
with drive impedance.  I soldered up an attenuator so I could properly drive
the inputs with 2.5, 5, 10, 25, or 50 ohm source.  I did notice that gain was
severely affected by loading, especially at 1:36.  Tried to measure inductance,
but my readings were backwards, with higher inductance at higher gain.  So I
discredit whatever I did.  Never trusted that LCR meter anyway.  Measuring the
coil resistances opened my eyes.  Each primary was 13 ohms ESR.  The more coils
I put in series (for reduced gain), the resistance was 13 * number of coils.
Perfect.  However, the secondary came in at 3.9k!  Wow, that is high.  No
wonder I was getting so many losses.  Can't believe nobody on the net ever 
mentioned this before.  This tranny might want to be loaded on primary, not
secondary.  I have to rethink this.  On the other hand, at low gain there is a
definate peaking, and it needs about 20k loading.  The higher gains are fine
with just the 47k of phonostage.  This makes sense, as my listening tests with
LO (resistance) loading made the sound very dark.  Not bad though, at low gain.
Where does that leave me?  I'll have to recalculate proper gains and loading.
Then maybe tweak the original values I chose into something that works better.
This is a very fine tranny, I'm quite happy with it in many ways.  Just need to
tune it in.  Oh, for reference, I looked up their specs for a 1:10 step up, and
it gave 19.75dB gain into a 100k load.  That means it has about 3k resistance 
in the secondary!  So my 3.9k for the CMQEE-3440A is probably correct.  Note,
this translates into a -0.7dB loss right off the bat.  At 10k loading, it's
-2.9dB.  Also, shorting the secondary gives you 100 ohms on the cart at low
gain (3.9k/81 + 52)!  And we can never load the cart at less than 13 ohms.
Hence, my wonderment over primary side loading.

So what is the proper loading for a cartridge?  I've heard plenty of
recommendations, like 10R, where R is the internal resistance.  Something is 
telling me the correct answer is R-squared.  I don't know why yet, just a
feeling.  Maybe I can figure out the math or theoretical background for it.
So my Dyna 20XL has 5 ohms internal ESR.  So according to the new theory, best
loading is 25 ohms.  I've been running 100 ohms in the PICCOLO.  50 ohms was
just fine, too, but I like the added sparkle on the top end.  My MED/MED on
the SU-10 gave exactly 100 ohms.  Hmmm, maybe it should be (2R)^2?  

Whoah, there is something going on here.  Here's the calculated loading on cart
adding in the primary and secondary ESR:

   L     M     H
L  203   352   680
M  65    102   184
H  22    32    53

Vertical axis in gain, horizontal is loading.  Wow, the 203 ohm loading sounded
way darker than the 53 ohm setting.  So I think I was doing mostly transformer
response shaping.  Not cartridge loading.  Ok, let's look now at what happens 
to gain.  For this calculation, we reflect the ESR of primary onto secondary,
add it to the 3.9k, and then attenuate with external loading.  

   L     M     H
L  13.2  16.2  17.7
M  17.1  20.9  23.0
H  20.3  25.0  28.0

Holy moly!  No wonder I ended up at M/M for my listening.  It gives 21dB gain 
at 102 ohm loading, almost exactly the settings I used on the PICCOLO.  The
loading gives HUGE losses in gain.  Ok, this is a good argument for putting
loading on primary.  The question is, does the tranny prefer to operate with
current or not?  

Ok, so let's rethink this.  What is the lowest secondary loading that this 
tranny can handle before it loses gain.  I put a pot for loading, then drove
the primary with the appropriate impedance (5, 10, 25 ohms) for each gain
setting, then tuned for maximum tolerable resistance (before too much resonance
peaking).  We get:

   L     M     H
   28k   43k   60k
With the 47k loading of phono, we never reach this point at high gain.  It is
almost perfect as is.  At medium gain, we're close enough, or should have a
510k resistor in parallel.  Low gains needs a 68k always in parallel.  At 47k
there is way too much ringing.  So I think we need to critically damp the
transformer separately from the cartridge loading.  Hence, automatically switch
in 68k and 510k resistors at the right gain settings.  This will mess up the 
wiring, but can be done.  Then, cart loading is done on primary.  Doing this,
we reflect loading of 446, 172, and 52 ohms respectively.  Our loading is thus
in parallel with these.  Gains are now independent of loading and we get 17dB,
23dB, and 28dB respectively.  Nice, we maintain or ~6dB increments of gain.

The question becomes, how to implement primary side loading, as it no longer
scales with gain.  I'd like to see each of the nominals cut in half, or 223,
86, and 26.  That would require a 6 position switch?  Let's say I used the same
values as PICCOLO, or infinity, 1k, 470, 220, 100, 47.  Then I get for loading:

      L    M    H
      446  172  52
1000  308  147  49
470   229  126  47
220   147  97   42
100   82   63   34
47    43   37   25
Now we've put the tone (loading) control on the cartridge, not the tranny.  It
ain't perfect, nor linear steps, but I get the range I want.  This is better 
viewed as a plot.

SU-10 loading, 6 position

Hmmm, nice drops for the LO gain setting, not the other two.  Ok, how about

     L    M    H
	 446  172  52
442  222  124  47
174  125  87   40
52.3 47   40   26

This kinda makes more sense.  Bandwidth is constant at roughly 25kHz, peaking
is constant (flat response), and each value of loading can be halved.  The 17dB
setting covers carts from 0.5mV to 1.5mV, 23dB for 0.2mV to 0.6mV, and 28dB for
0.1mV to 0.3mV.  And just enough loading adjustments to go from open to dark.
Nothing extra to go too far.  Only the middle ranges we need.  

SU-10 loading, 4 position

Now I have to swap locations of gain and loading controls.  I wonder if this
gets messy.  Do I need to make a layout?

Jan 19

I think the PFO guys do great show coverage.  But only from the standpoint of
showing everything in all of the rooms.  They blanket the place.  What is
missing is any significant analysis of sound.

CES show coverage (go to bottom of page)

Still, you get a really good idea of what was going on.  Indeed, the show was
huge.  All the big players were there.  The industry is not sitting on its
feet.  I see a LOT of new designs.  New players.  The choices in front of
consumers is bewildering.  In fact, I bet the choices are greater now than they
have ever been.  So what do I see?  Lots of really expensive stuff with an
emphasis on visuals.  Not much budget gear.  Who is our customer?  Is it now 
the guy with so much money that price is not a factor?  Only the visual status?
And how does this tie into the fading economy?  Perhaps the hi-end is immune
from future recessions, as most of the products are just big budget sculptures
designed to impress?  I dunno.  I think the customer base will be split.  The
haves and the have nots.  Those who have been investing in the budget (under
$3k) gear will probably be the ones to hold back on new purchases.  The super
rich will have no trouble.  I wonder if I have positioned myself correctly.
The $2k to $8k range might just end up as a dead zone.  My products will have
the sonics, but not the price and glitz of a $10k to $20k machine.  And that
could be a mistake.

The good news is that I'm the only one using red anodized faceplates!  My
chassis are also much smaller than the norm.  Higher WAF.  But maybe that is
not what you need?  I am hoping my features and innovation can carry the day.  
If the answer to success is "fashion", then my focus on sonics and value was
the wrong choice.  Has performance really taken a back seat?

Jan 18

It is a new day.  We move forward.  Anyone heard the DeVore Super 8 speakers?
They were recommended to me as a possible choice for RMAF08.  Not sure I really
want to show up with the frankenvandys.  I'll be having my 8w push/pull PA-10
monoblocks to power them.  At 90dB efficiency, they won't get very loud.  But
then again, a lot of the complaints at RMAF are that the rooms play too loud!
If you've heard these speakers, please let me know what you think. 

And now I have been informed that 12B4 tend to be very microphonic.  Shux.  Is
that project going to die?  I suppose it is time to move it to the back burner,
and use a CLARINET for the time being.  Meanwhile, starting builds of HAL
production CORNET and TRUMPET.  Final design.  Not prototype.  My wife says the
power umbilical should be yellow.  Ok, how am I going to find that?  Parts are
in to finish up the SU-10.  So I will do that and listen this weekend.  Am also
trying out various capacitors for use in the PI-10.  If I ever feel like it, 
the FLUGELHORN can be wired up too.  All parts sitting in a box.  Boards all
built.  Maybe I can get the DA-10 relayout done today.

So is the economy going to tank or what?  How is this going to affect hi-end
audio?  I just worked my ass off for 7 years making minimum wage building up
HAGTECH.  Finally breaking through into the black.  I have momentum.  HAGLABS
is launching right at the start of a recession?  

Spent some more time thinking about how to make this transition - a duality of
both HAGTECH and HAGLABS.  Something that makes sense.  Move products around.
Re-organize.  And I think the answer lies in having a HAGLABS "direct" set of
products.  HAGTECH is going to be the DIY site.  That means machines such as
the VACUTRACE needs a new home.  That's why I put the HAL logo on it!  The
starting lineup of "direct" or "oem" machines will be VACUTRACE, FRYKLEANER 
(turn the gold version into a red one), and the UFO.  I only have 4 UFO left,
so that's just a temporary thing.  Pricing will be $360, $1560, and $380 (yes,
the FRYKLEANER gets cheaper!).  The HAGLABS "retail" products will then be
TR-10, CO-10, PI-10, SU-10, HA-10.  Soon to add the DA-10, PA-10, FL-10.  This
means I need to make three big changes to the HAGTECH lineup.  First, the
FRYBABY needs to be available as a half-kit.  Second, offer the CASTANET in
assembled form ($999, black anodize).  Third, the ARCHIVER disappears (turns
into FL-10).  The only problem with this is that the CASTANET will kill HA-10
sales.  But that's ok, as I doubt I was gonna sell many anyway.  The CORNET
will also be removed from HAGTECH (buy the last ones now!).  The idea is to
have the website to make sense, thus leaving the CORNET2 to reign.  

Jan 17

Checked out that tube in the DA-10.  Yep, the bad tube bias shift way up, and
it lost compliance against the CCS.  Ended up clipping top part of waveform
(bottom ok).  But only on loud signals.  Did a quick check of amplifier
response by injecting a signal through a 1k resistor.  This hits the 50 ohm
filter loading, plus the DAC 1.2k output impedance, plus the filter looking in
reverse, so it had some response shaping to it.  I'm not sure what the filter
does in reverse, but the output signal produced a very nice 10kHz square wave
with a little bit of overshoot.  BW was 60kHz.  So very much like a Butterworth
response.  I can't see any limitations from the tube.  The good channel put out
huge voltage swings.  Maybe 20Vrms.  Or ten times what we actually need.  So I
will go ahead with a relayout, buy more tubes.  

Aha!  My assumed cleverness screws me.  I got this weird 60Hz injection on one
channel that was huge.  Nothing made sense.  Finally tracked it down to maybe
a broken socket.  Nope, it was the damn rubber washers I put inder the socket.
I jammed the ohmeter probes into them and yup, sure enough, they were quite
conductive.  I was shorting one of the heater lines right to a grid.  Replaced
it with a nylon one.  Now the LS-10 is running ok.  Great output swing, moreso
than the DA-10.  A few quick measurements show bandwidth from well below 20Hz
to about 80kHz.  Very clean square wave with zero overshoot.  Gain was high,
at +20dB.  That's 4dB higher than I expected.  And that's with a 6dB pad up
front!  Maybe that is too much.  Have to think of ways to drop it down, like
a resistor in parallel with the plate choke (ruins linearity).  Anyway, I am
still getting some hum injection, more in the left channel.  It is most 
likely from the power tranny into the choke.  I'll remote connect the choke to

So I did a listen.  Hooked up the LS-10 between the CORNET and the HA-10.  Way
too much gain, but at least I can hear what it is doing.  Oddly, there is a hum
pumped into the HA-10 even with the power off.  I wonder what that is from.  
The sonics are fine, but I can't live with any hum, so that needs to be
addressed.  Perhaps I was a little too optimistic regard the placement of
chokes.  In the HA-10 I have the advantage of a 19dB reduction in the output
tranny.  Not here.  I doubt it is the heaters, even though I am running them on
ac.  This is really hard, placing power and gain in one small chassis.  It's
all a mess of big magnetic fields.  

This is getting me depressed.  The 6H30 just isn't the right tube for the job.
I really want a linestage of about 12dB gain.  Any more is a total waste.  And
the last thing I want to do is to mess up a circuit to trash gain.  So I look
around for other possible tube choices.  The 12BH7 and 5687 come to mind, but
they're not that different.  I need a tube with a mu of 4 or 5, rp of 1k.  A
45 or 2A3 works, but we don't need that much power.  What makes sense?  So I
browse through some books.  Maybe a 6F6.  Close encounters are the 6GU7, 6CS7,
and 6CM7.  All too much gain.  Luckily, the RCA receiving tube manual points
out a 12B4A as a vertical deflection ampplifier.  Or something like that.  Ding
ding ding ding ding!  Holy crap, this is it!  This is the linestage tube.  Best
fit for requirements I have seen yet.  Plenty of overhead.  Super low mu.  I
should have started with this one.  A google shows others have already found 
this tube and confirm it as a reasonable choice.  12B4A!  And they are sooooo
cheap.  Maybe we can change that?  But this is not the only problem to be
overcome.  The available heater current is at 1.0 amps, I'll need 1.2.  Might 
be able to pull it out without saturating core since B+ can be kept low.  I'd
consider same circuit, bias at 85V, -10V, 15mA.  The tube is coasting, but 
should still have low enough rp to reach down under 20Hz.  Easy to change
layout to 2 tubes.

The big remaining problem is the induction into chokes.  I'm afraid this could
be insurmountable.  Only solution is dual chassis, which makes the product cost
way higher.  Cannot fit into the smaller chassis either, as the trannys are too
tall.  And so I am depressed.  The LS-10 may be kaput.  I'll know for sure when
I run the inductor test.  


Ok, just ran the inductor test.  Dang.  I put the 30H choke on wires so I could
position it anywhere, relative to the power transformer and power chokes.
Well, it wasn't hard to make that hum go away.  All it required was enough
distance.  I had thought that turning the core on its side would hum buck, but
no, not a whole lotta change in induction.  The pickup is 60Hz from power
tranny and 120Hz from chokes.  Anyway, it takes 7 inches center to center to
reduce hum to a very quiet level.  The chassis is big enough to get close to
that spacing, but only if you don't have an ac input connector or RCA jacks.
Actually, I could probably do it all RCA jacks on front panel.  But that won't
fly.  The circuit modified to 12B4 is just fine (680 ohms for R12), but the
chassis isn't.  Great linestage circuit, but only if used in a giant chassis.
Interestingly, there was no hum caused by the ac heaters.  

Jan 16

Came up with a small improvement in tube socket mounting.  I added some rubber
washers under the socket, then bolted it down to the board using a stainless
steel screw.  Ok, so now it is mechanically rigid, and flat.  The two-piece
ceramic is not loose.  Yet there is still the rubber shock mount.  Now, when I
solder the pins, they are mechanically floating within the carrier, no longer
stressed to one side.  I found me some perfect rubber bushings and ordered them
for production.  I can also retrofit this change into the TRUMPET.  Still
looking into the possibility of a brass counterweight underneath.

Jan 15

Got to listen to the DA-10 last night.  My first impression was good/bad.  Some
things came out really nice, like vocals.  But louder passages got congested 
and mixed up.  The hum isn't really a problem anymore, nor can I hear the 
microphony.  Not a listening levels, anyway.  The thing is, I listen to noise
floor on all of my amplifier with HUGE gain, maybe 20dB above loud.  That way I
can dissect hum and noise issues during prototyping.  I can hear chassis.  I
listen for power up / power down thumps and such.  So maybe the microphony
ain't so bad in real life.

However, this does give me an idea!  On the original TRUMPET (ca 2002) had the
tube sockets bolted to the circuit board, which then hung like a suspension 
bridge offering vibration damping.  The sockets were tightly tied down, not
just ceramic floating amongst soldered pins.  There are two benefits to this, 
but I digress.  Anyway, my HAL stuff also has provisions for bolting the
scokets.  I was smart enough to throw that into the layout.  So how about
adding a counterweight?  What would happen if I bolted the socket to a large
brass counterweight?  A heavy mass to absorb vibration.  A sonic sink.  I
wonder if such would work.  I might need to experiment here.  What a cool and
unique feature!  The extra mass would provide local mechanical stabilization
for each tube.  Even better if I use loosely coupled sockets?  On something
like the DA-10 this could be really useful.

Ok, back to the listening.  Something was wrong.  I was overwhelmed by the
strong bass and fullness of the sound.  Very thick and robust.  Terrific bass.
But during some passages I get this disturbing mess of distortion, like when
you get crud stuck in the stylus.  Shoot, it got worse with time.  I thought
at first is was smear from the capacitors.  Finally, I realized it was only
in one channel, my brain was working overtime trying to compensate!  One
channel was crystal clear.  Maybe I am deaf, but it wasn't really apparent
what the trouble was until I ran one channel at a time.  Gads, it was getting
bad.  Ok, swapped tubes between LR.  Bingo!  That was it.  Bad tube.  Now the
other channel was full of distortion, definately related to loudness in the 
midbass.  Vocals could still come through ok.  Sadly, the tube with the 5kHz
microphony was the good sounding one!  It might be an RF oscillation, not sure
yet.  But one messed up tube.  More on order.  I should run some usual lab
tests.  If ok, then maybe I can stick with this tube.  When they work, it all
sounds really good.  

And so the LS-10 boards came in.  Luckily I already had on hand all of the iron
needed.  Spares from other projects filled in the rest.  I cut out the section
to make room for my ALPS blue velvet.  Dang, I should have tested out the stock
volume pot first, make sure I wired it up correctly.  Then cut it out.  Oops.
Only coupling caps I had on hand were the ones froma TRUMPET.  Hey, this is
going to be my personal unit, so might as well not put in anything junk.  I'm a
little worried that the power tranny is too close to the right channel plate
choke.  Could induce residual hum.  We'll see.  

LS-10 proto top
LS-10 proto bottom

I really love this circuit.  So simple, yet so capable.  Just like the HA-10.

Jan 14

Was wondering if I could fit the cinemag trannys into the new CORNET chassis.
Don't quite fit.  Always thinking of product improvements.  Usually I just get
myself into trouble.

Researched the web to see if there were any other low cost PC-based scopes that
will work with a VACUTRACE.  Can only find this one.  I own the fore-runner to 
it, so I know it works.


Jan 12

This DA-10 has a bit of hum, and a lot of tube microphony.  Something really
odd about it, not like any other tube noise.  Not a ping, but a shimmer, like
a resonant screen is shaking.  One tube is so resonant at 5kHz the signal looks
like radio, music with a 5kHz carrier riding on top of it.  I have to wait at
least 10 to 20 seconds for it to die down to residual levels.  Maybe I need to
mechanically decouple the tube sockets from the board, as the power supply
chokes are also mounted to it.  Maybe all it takes is a tiny bit of 60Hz motor
vibration to get things kicking?  Really noticeable sonically when you touch 
the tube.  I need to buy some more samples.  Maybe I just don't have good ones?
The hum I think I know what I did wrong.  Looking at layout I definately
created a very large loop with the heater currents.  They measure great, with 
only 20mV of 120Hz sinewave ripple on them.  All from just one level of LC
filtering.  But I'm pretty sure I am putting some ground currents where they
should not be, mixing in with the low level DAC output.  Easy to fix, actually.
Just a stupid mistake in the first place.  I also realized I goofed on the hum
bucking.  One side of supply mirrors itself for stray field cancellation, and
also mirrors the other side, for far field cancellation.  Well, forgetting one
supply was -6V while the other was +6V, the flux was already reversed, I didn't
need to reverse it again.  So now they are additive, rather than cancelling.  

Anyway, the main problem is the microphony.  If the 7788 tubes end up to be a
problem here, at such high gain.  Then I'm in quite a design corner.  The B+
supply is really low, can only support CCS or plate choke loading.  Cannot do
a cathode follower.  So no going back to the original CHIME tube circuit,
unless I do a total redesign of the power supply.  Can't use the 6072.  Maybe
a 6H30 or 5687, but then gain is too low and heater circuits need all new 
design.  Basically, either I get the 7788 to work, or it is back to the drawing

Jan 11

Ok!  I got the DA-10 fired up.  Oops, some diodes backwards.  Then the power
supplies ran perfect.  I got exactly 6.2V on each heater.  The HAGDAC regulated
down to +/-5V.  Odd, I didn't need to bleed extra current on the negative side.
Hmmm, I figured the HAGDAC was unbalanced in current draw.  I guess not.  The
gain was just right from the tubes, however, the 7788 are microphonic, in an
odd sort of way.  Not the normal ping.  More of a shimmer that can last over
10 seconds.  I'll see if dampers can quiet them down.  A little bit of hum,
which might be due to my output wiring.  That was a mistake.  I need to put
those tracks down on the board.  The coaxial cable is a bitch to work with.  
Oh, I also changed the CCS, as the 12V zener drop was killing my headroom.  So
made it 5.1V, dropped the tube current from 15mA down to 12mA, and now I have
plenty of headroom.  Noise level is way down, except for that hum.  At normal
listening volume you don't hear it.  But that ain't good enough for me.

And the TR-10 is up and running.  Will listen to it tonight.


It has been listened to.  And is it ever quiet!  Moving that power supply away
dropped the noise floor way down.  Nothing there but a wonderful benign hiss.
Makes a great combo with a PICCOLO.  Been a long time since I heard a TRUMPET.
It just amazes me.  I've been happy as anything with my tweaked out CORNET.
Yet, here it is.  Better sound.  Just when you think it isn't possible.  Bam,
a better phonostage.  More natural, clearer, better articulation.  Shux.  I am
blushing.  I have to rethink what I am doing for the PL-10.  That design had
6072 for gain.  I think it has one too many stages.  Would prefer to simplify
it, as I have done for LS-10 and others.

Yesterday I was thinking about just forgetting the TR-10.  Soooo much work to 
assemble one of these things.  And not easy.  But now I see someone posting a
photo of it on asylum.  And listening to it.  Shoot.  No choice.  I gotta put
these up for sale.  It's just too good of a machine.  Even if I can't mass
produce them.  Maybe I can talk Mackris into being a dealer.

Jan 10

I think maybe I'll just get the new iMac.  It's $1200, though.  Shoot, I could
get a Dell for half of that.  But me wanna run OS-X so I don't have to mess
with audio drivers.  Just plug in my DA-10 and go.  Play CDs and saved files.
I could get the macbook for $100 less, but what's the point?  Mainly I need
this for RMAF.  I suppose I could wait awhile.  My laptop will go to the shop.
The iMac will be my dedicated music server (why would anyone buy an Olive?).

Did some more work on LS-10 layout last night.  Forever moving things around
trying to re-position for better routing of signals, controlling of ground
currents.  I made a cutout feature so now I can have the option of using an
ALPS blue velvet pot for volume.  Also, I added the ability to jumper the 4th
input.  That is, you can use the 4th set of RCAs as an input, or wired as an
output for monitoring or recording.  5th set of RCAs is the tube output.  So if
you need 4 inputs, you got it.  Like me, if you want one unvolumed output to go
to an HA-10, you can choose that option.  Layout only needs another 3 or 4
hours of work.

Ok, finished the layout.  Now I can turn my phone on.  It looks something like

LS-10 layout

Finally got back to the new TRUMPET.  Wired up the power supply.  Aye, I can't
believe how much work this is.  Takes forever to put this thing together.  I'll
have to see what I can do to streamline it.  Very difficult to wire up the
power tranny in the new chassis configuration.  Makes me want to just build
CORNETs.  Anyway, supply up and running.  Can build the umbilical tomorrow.
Amplifier done and ready.  Hopefully, I will have it running and can do the
listening tests, for noise, etc.  I'll be so glad to finally get this machine


Jan 09

Putting together some of the new products.  Ok, after getting the USB port off
by 1/2", I decided it would be best to move the panel holes for RCAs too.  On
everything!  This actually, helps me a little with keeping inputs as far away
from ac and outputs and such.  Lines things up a little better.  So I think I
am ready to order me an LS-10 set of panels.  I'll wait until I get the DA-10
board fully debugged.  Meanwhile, the FLUGELHORN (aka FL-10) boards got mounted
to panels.  This stuff fit perfect.  Need to buy me some more chassis.

FL-10 panels front (missing switch)
FL-10 panels back

And also started putting together the stepups.  The PI-10 (PICCOLO) and SU-10
(transformer STEPUP) use same chassis and I/O configuration.  They look very 
much the same, do basically the same thing.  Your choice of technology.  

SU-10 and PI-10

Now I have to finish the TR-10 (TRUMPET).  It's already sold!  I better get
moving.  Then build up first production versions of CO-10 (CORNET) and HA-10.
One of these days I'll get back to re-doing the layout of the PA-10.  I need me
a pair really badly for my listening room.  So my schedule is doing ok.
Website will not go live in January.  I decided to wait until I have a full set
of production machines ready.  Especially needed for photos!  And then I have 
to write up manuals.  There is really no point launching early.  I'm setting a
new target date of 4/1/8.  Products to be (in no particular order):

TR-10 (trumpet tube phono, re-issue)
CO-10 (cornet tube phono, re-issue)
FL-10 (flugelhorn jfet phono, re-issue of archiver)
SU-10 (step-up tranny)
PI-10 (step-up jfet, re-issue of piccolo)
LS-10 (tube linestage)
HA-10 (tube headphone amp)
DA-10 (tube usb dac)
PA-10 (tube power amp)

Jan 08

I dunno, seemed to not be so worried lately (except that I'm always in a panic
about RMAF08).  Read that interview at ETM on Tenor amps.  Sheesh.  Is that how
most companies are run?  You can have it.  I have no desire to go in and out of
business so often.  Seems to me a LOT of hi-end companies do this.  Hmmm, maybe
this is a strategy I am missing?  Could be I am screwing up a lot less than 
other folk.

Then some audiogon thing on CES says Empirical Audio is leading the charge on
pc audio.  Fine.  Holy cow, then I see the photo of the guts of this ultra
machine.  What a mess!  Sorry Steve, but that's not very pretty.  I'm sure it
sounds great, though, as it seems to have all of the great techno stuff a 
leading edge dac should have.  In fact, same as the CHIME I came out with three
years ago!  My new DA-10 should easily compete sonically, technically, visually,
and will likely cost less.  And when you take the cover off, it will look like
everything belongs, not a patchwork of other people's boards.  Yeah, I'm just
doing a little advertising and self-promotion here.  Started building a
modified HAGDAC today for my DA-10 proto.  The main board is already built (see
photos below).  All I have to do is put things into a chassis and fire it up.
In anticipation I bought telegraph road from musicgiants.


Hmmm, just got notice from Hammond that they don't want to do the modifications
to the stock chassis.  Guess that means I do them myself.  I have to get better
at drilling and tapping holes.  Should make me a better machinist.  Also, no
question I need that compressor / paint sprayer now.  That's ok, it will give 
me a better automotive finish, and I paint AFTER the tube holes have been
punched out.  

Did I already post a schematic of the LS-10?  If not, here it is.  It looks
just like a CASTANET, but without the output tranny.  I think the choke for a
plate load is a great thing.  This should sound like a CLARINET, but with more
output swing and lower output impedance.  Extremely simple circuit, which is
why I am confident it will sound good.

LS-10 schematic (draft)

Arg.  You work so hard.  You crank and jam on the layouts and panel designs.  
They look great on paper.  You run over everything again.  It looks perfect.  A
work of art!  So you order the parts.  Dang!  They come in and things don't
fit.  Arg.  That beautiful DA-10 board.  Well, I tried to plug in a HAGDAC
today and discovered I placed the connectors 0.1" too far apart.  Then the USB
connector is exactly 0.5" in the wrong place.  Ugh.  Ok, that is why we build
protos.  Work the bugs out, however small.  This happens every time, yet it
still bugs me.  It's just part of the process.  Tomorrow I'll see if the
FLUGELHORN panels fit.  

Jan 07

Hey, maybe the reason sales drop in February isn't because of Valentines, but
because of the super bowl.  Everyone buys a TV that month instead?  Meanwhile.
thinking about record cleaning.  Had this thought of spinning a record really
fast.  I mean fast.  At such angular velocity, anything near the surface would
cause a ground-effect mini-vacuum.  I wonder if you could make like an inverse
turbine where the flat disc spins, not the blades.  Could it vacuum one slice
at a time?  Or the entire record.  Probably just bend the vinyl until you get a
head crash.  Hmmm.  Maybe the centrifugal rotational forces can keep the record
flat.  Anyway, the idea is to make the record part of the pump itself.  

As for the LS-10, I decided to continue using the same volume pot and select
switch as the CLARINET, but with a twist.  I'm adding holes in the board such 
that I can run jumper wires to an ALPS blue velvet.  All I need to do then is
carve out a section of the board.  The problem was component shaft mounting
height off of the board.  Some controls are 0.25", these were 0.375".  The ALPS
is about 0.45".  By luck, the 0.375" used in the HA-10 came so darn near the
true centerline of the chassis, I just went with it.  A slight board flex was 
all it took (about 0.01").  So now I can fit the ALPS on the chassis centerline
(good for visuals) by cutting away the circuit board.  I have the option of 
using either pot.  Ready to finish the proto layout now.

Spent the weekend carving up HAGUSB chassis.  Most efficient to do many at the
same time, as it is an agonizing process of mechanical manipulations and CNC
machining.  The XLR versions require two separate CNC runs, and use parts from
the RCA boxes.  It's about a 10-step operation per box.  When cranking through
these, the average time per box is only 10 to 15 minutes.  But when you run 
through 50 boxes, that adds up to 10 hours labor.  Just to cut holes for the
dang USB, RCA, and PHONE connectors!  Ugh, 10 hours of weekend lost.  But hey,
I now have a large stock of ready-to-go boxes in both flavors.  Should last me
at least two months.  Reminds me of a cool video I just saw on making a vacuum
tube by hand.  I thought this was well done:

Triode, by hand

Jan 05

Hmmm, I just realized I also need a computer transport for RMAF.  I don't think
I can depend on my IBM anymore.  Should buy me a Mac.  I love the new iMacs.  
Me just don't want to shell out so many bucks.  I'm burning through enough cash
as it is.  Ugh.  Computer and speakers.  Did some more research into the Maggys
and their $500 MMG.  It seems really good, and I can probably live without the
bass (for the show), but there is no way I have enough power.  Everything out
there says you need 100W and much more.  Crap, my plans for the 1.6s may not 
pan out.  I have to go listen to them at 2W of power.

Finally starting to catch up again on production.  Been hammering away with the
soldering iron.  

Oh, discovered I ran out of chassis for the CHIME box/2 option.  That's it.  No
more.  In fact, I only have two more assembled CHIMEs.  I guess the timing is
ok, as I need to promote the new DA-10 (which ain't quite done).

Jan 04

For some reason I am all worried about RMAF.  At the pace I am working, it's
gonna be a close call.  I don't even know if I can get a room yet.  Folks from
last year get first dibs.  The questions are: 1) will my linestage be done in
time, 2) what speakers to I bring, 3) what turntable do I use.  Maybe I can
arrange a loaner table from Thom again.  Or I can bring my tweaked out Rega; it
ain't no Galibier.  The frankenvandys might cost too much to ship.  Who can I
partner with?  Do I use the Horns again?  Do I go for the Maggies already?  At
least they aren't too bad to ship.  I could try out the MMG at not much cost. 
How loud will they actually get with my 8W amps?  

And so, in a paranoid state, I designed a linestage today.  A backup.  This 
would be an LS-10.  It is just about the same as the HA-10, minus the output
transformer.  This raises gain and lowers distortion.  I use the same 30H plate
loaded choke on a 6H30.  Thought about maybe a 5687.  Otherwise, much of the
electronics is the same.  All I have to do is move the power transformer to 
make room for RCA jacks, then add a selector switch.  It is not a bad idea.  I
could whip through the entire design process in maybe ten hours.  So much is
borrowed from previous work.  I would then have a linestage to use with my
PA-10 amps.  Maybe I only make one unit.  Maybe it is not a bad product to add
to the lineup, as not everyone will want the combo phono/line of the PL-10.  It
would also be pretty low cost.  I should probably use an ALPS blue velvet for a
pot.  Yeah, maybe I better make one of these on the side.

Meanwhile, earlier in the week I designed a new FRYKLEANER.  Was thinking, what
would the next generation look like?  Well, more power.  I figure a stronger
output signal would be good.  Also, put it in a rugged chassis like the
VACUTRACE.  Make it nice for an audiophile with bucks.  Make it a HAGLABS
product.  The problem, I think, is that people don't like the FRYGOLD form
factor.  It is not selling anywhere near as well as the old FRYPRO sold, even
though it has way more outputs.  I think nobody want a rack mount.  Anyway, the
paper design I came up with uses an output transformer to create a balanced 
and higher voltage drive.  Had to make an internal ac power supply, as a wall
wart will not fit within the case with lid closed.  But a power cord will!  No
hurry on this design.  Just a thought exercise at the moment.  

A VACUTRACE buyer thought a version with USB output would be really cool.  I 
had thought about doing this years ago, but since the design is entirely analog
there was no way to do it properly.  But hey, what about a next generation?  If
I had to do it all over again, how would I do it?  Easy, I would use a
microprocessor!  In fact, once I get the PL-10 done, I'll have the processor
stuff down pat.  So maybe 2009 or 2010.  A new VACUTRACE, smaller, USB
controlled.  In the meantime, I will think of what features and improvements
can de added.  There is still zero competition out there.  Over the years, many
have tried, but they all get stuck.  What they didn't realize was that the
software, processor, and ADC stuff is the EASY part.  

Jan 02

Good news, everyone.  The HAGDAC beat out all other contenders in a shootout
done by a very reputable hi-end company.  They are trying out various DAC
boards that could fit with their fancy JFET I-V converter stage.  Apparently
the HAGDAC not only sounds the best, but also has by far the lowest measurable
distortion.  He he he.  I can't wait to hear it in my new DA-10 chassis.

Here's some shots of the DA-10 proto.  I wasn't really trying, but it seems
almost all of the power supply parts ended up on bottom, amplification and
DAC on top (shown without HAGDAC).

DA-10 proto top
DA-10 proto bottom

Seems like a lot of power supply for just two tubes.  Overkill?  I'll find out
by listening.  My major concern is induced hum via all those stray magnetic

Ok, so I did my taxes.  Hawaii excise tax, that is.  I have to pay every month.
Business tax.  Luckily, out of state sales are exempt.  But I still have to
prove such with documentation, so it takes a lot of work and organization to
compile each month.  That is, without wasting much time.  Anyway, it also gives
me up to date analysis of my sales.  I made a spreadsheet where I keep track
over the years.  I find some neat things.  If I average out the monthly take,
the audio sales looks like this:

Sales by month

The start of the year is pretty average and steady, but with a dip in February
thanks to Valentine's Day (no kidding!).  April is taxes.  Then May-June-July
drop way down as folks take summer vacation.  I don't know why August pops back
up.  Then September drops, maybe because of school?  And finally November and
December are the killer months, where holiday sales make the year!  

Now if you go year to year, I break my income down between consulting (yellow)
and audio sales (pink).  Apparently my strategy is working.  I used to generate
roughly half of my income by consulting on the side.  That stayed true for four
years (and probably earlier, too).  But then I focused on HAGTECH.  And it paid
off!  Consulting quickly slid to zero, and audio sales took off.  I can only 
hope that pink line keeps going.  If my HAGLABS gamble works, it will.  The
last two dots are '06 and '07.

Revenue by year

Jan 01

I forgot to do a recap of last year.  Oh well.  It's in the past.  Whatever.  
Let's look forward!  Did I make any resolutions?  Nah.  I'm on the right track
already.  Just have to stay focused.  Let's see, what is in store for 2008?

1) Launch HAGLABS website with real products and an order form.

2) Part of above - finish the design and builds of CO-10, TR-10, PI-10, SU-10,
and HA-10.  Add the DA-10, FL-10, PA-10 and LC-10 as soon as possible.

3) Get that premier PL-10 design going!  It is the key.  The centerpiece.

4) Finish the ARTISAN phono.

5) Get home system rebuilt around PA-10s and the modified Vandys.  What to use
for a volume control?

6) Complete the PL-10 in time for RMAF.  Do a great show! 

7) Start up dealer network.

Meanwhile, the DA-10 board came in, and I built it up.  Everything fitting very
nicely, as if this was a 2nd revision already.  One booch, I put some 1uF caps
where the HAGDAC board is supposed to go.  And I forgot to buy 1 ohm resistors
for the 2512 spots on input filters.  Anyway, found some old 155B chokes, rusty
but working.  Then dug up some 716P orange drop caps.  Figured I'd use them for
output caps.  Hey, it's a proto.  Mostly I need to tweak in topology and values
before I plug in the uber components.  So far it looks beautiful.  Mostly, I am
super busy filling production orders right now.  Hey, lots of sales is a good 
thing.  I don't mind putting off the lauch of HAGLABS line a little, as long as
the money is coming in.