[Harp-L] Re: Jayphat like Ron Holmes' box? (long w/ synopsis up front)

[HTownFess <Spschndr@xxxxxxx>]

Anyone familiar with Ron Holmes' generosity to harp players will not
be surprised that he took the time to look at the JAYPHAT schematic
and confirm that it does not derive from his mic buffer design ~at
all~, and furthermore was kind enough to allow me to post his design,
which he had put in the public domain on his website for a while.  The
Holmes_Loop_mod schematic at http://groups.google.com/group/harp-l/files/
is for Ron's entire effects loop design; the relevant portion for
JAYPHAT comparison is at the lower left, on the audio input, to the
left of the jacks.  Ron explained the difference in the email I am
appending and you can plow thru the tech details if you want, but the
short version is that his is a ~truly~ different circuit running a
different JFET a different way; it could possibly be done with the MPF
102 JFET used in the JAYPHAT, but component values would have to be
adjusted and there are reasons why Ron uses a different JFET.  He
mentions that the "common drain amplifier" circuit of the JAYPHAT has
been around since 1963 in its generic form.

You'll also see that Ron is skeptical about the JAYPHAT's performance
except as a rough & ready device that newbies can build.  My friend
Greg Schlacter found exactly what Ron predicts: the JAYPHAT is not a
studio-quality buffer to run high-impedance things like crystal mics
or humbuckers straight into the board with.  That's precisely why Greg
gave me the prototype, because any shortcomings the JAYPHAT has in
that sense are not apparent so far when using it as a mic buffer for
amplified harp.  You can listen to the SuproSupremeFinal clip and
other JAYPHAT demos and decide for yourself; I'll also mention that I
used the JAYPHAT on a Grammy-winning record and on another record that
was #1 for at least half of 2004 on CrossCut's European blues chart,
and nobody noticed any shortcomings in the rig on those occasions.
JAYPHATs have already been built by others over the past few weeks and
found quite transparent for our purposes; our list's own Jim Rossen is
one satisfied builder/user.  I posted about the JAYPHAT due to popular
demand from people who heard its performance.  It works fine.

I did revise the JAYPHAT2B schematic in light of what Ron said,
restored the 5 mfd polarized electrolytic output capacitor in case
that has any bearing on performance.  People have already done
different things there and the box still worked fine; try what you
want there if you are electronically literate.  I will do my best to
get some layout photos up; it's a remarkably simple device to build as
these things go, good thing to tackle for a first scratch build if you
know how to solder.

I will append the text of Ron's reply below.  I will put in another
non-affiliated plug for Ron's products: he is truly harp-savvy, knows
what we need and has been extremely generous in providing technical
insight to the harmonica community, as well as selling us stuff.  As I
mentioned, I found that I can use my early Harp Commander as a mic
preamp with Audacity on my laptop, and get quite decent stuff with a
mediocre mic I had laying around, wound up with a good rough & ready
mono recording rig with one very high-quality section, the Commander,
really warming up the soundscape.  Works for me, though it may be an
older version of Audacity than the one that is apparently getting
squirrelly on people of late.

Meanwhile, Jeff G, I do expect a public apology.  When you question
the provenance of the JAYPHAT onlist, rather than inquiring offlist
first, it is tantamount to a public accusation of intellectual
property theft, and I have dealt with it publicly partly because my
friend Greg Schlacter is not onlist to defend himself.  I have shown
as conclusively as possible that the JAYPHAT does not derive from
Ron's design.

Though this discussion has been useful in eliciting Ron Holmes's
assessment of the JAYPHAT, I hope it has been still more useful in
demonstrating how one should use Harp-L's offlist network to avoid
embarrassing oneself or someone else publicly when delicate questions
arise.  Inquiring offlist first can save bandwidth and increase
civility; I hope listmembers will realize that and remember to take
advantage of offlist options in the future.

Ron's reply follows--Stephen Schneider


Hi Stephen,
The Jayphat2 is a different design. That one is a common source
amplifier
with some gain. Am attaching the schematic for my loop mod circuit.

 The front end (or first 1/2 before the jacks) is similar to the
buffer amp
(no gain, low impedance out) I use on all my Commanders. Mine is a
common
drain, source follower configuration.

The performance is substantially different between the two. The
Jayphat2 is
a nice simple design where some gain is needed and you don't mind lots
of
harmonic distortion.

It is not very transparent and can take only a low level of signal
before
waveform compresses. N-channel Jets suffer from an anomaly related to
miller
effect (but with channel leakage) when high values of base resistor
and
higher voltage are used to operate circuit.

Leakage current rises very fast and causes input impedance (at signal
frequencies) to drop like a rock. So what looks like a 5.1 Meg input
resistance turns into an input impedance much lower-maybe 500K-??
depending
on drain current.

I tested those and they sound harsh I think due to distortion. But,
for a
quick amp using very low signal level they can be usable. That circuit
has
been around since 1963 or so.

The only FET buffer that is transparent and clean is the one I ended
up
with. That's the same one in the loop mod. It has no gain but has
other
virtues like a wide signal swing and complete transparency .

That is the source follower and has a slight loss of gain thru it.
But, it
is absolutely linear and squeaky clean. I use screened 2N5484 FETs but
the
MPF102 is ok. Radio Shack has those.

You could try building my circuit  using MPF102 run at higher current
than I
do. The source current effects base leakage so I keep mine low as
possible-maybe 500-600 uAmp.

R1 is 1 megohm, R2 bias resistor maybe 1K, and R3 2.7K. Similar
values-different places. The 1 megohm base bias is magnified by
current gain
(not voltage gain) of source follower at AC (signal) frequencies.

But, at DC the base is 1 meg above ground which keeps noise down. 5.1
meg is
too noisy. So, a designer looks at everything the circuit does and
optimizes
it across the board.

My source follower and my specially screened JFETs and current I use
are all
carefully optimized for very clean, silky smooth performance. The
circuit
has a "liquid" feel.

 I do my own FET screening and have to buy many FETs to get just a few
that
work properly. The manufacturing tolerances are huge and most parts
are
unusable. I have a big box of new FETS with unusable Idss values.

Now, you can buy a MPF102 off the shelf  and tweak it to work ok but
you
have to dial in both bias resistor and load resistor (R2 & R3) to
optimize
circuit for max clean swing. You can't do that in production.

As you can see this wasn't guess work. A lot of careful effort went
into
making a really good, consistent circuit. This is also the same
circuit I
did in the pedal mods. I added the no-gain buffer amp right on the
pedal
board.

That's the story. The values will change with different FETS,
different
currents, and different voltages. The values I use in my source
follower are
all way different. Plus the operation is different. Plus different
FET. They
all sound different!

 In any event you have to use premium parts including tantalum
coupling caps
to get the performance I get. I know this was a long answer but is
quite
technically precise.

Players deserve accurate information and not the usual bag of baloney
posted
everywhere. Plus-it helps a lot to know what you are doing. So absent
these
days.

Am always here to help. Am always happy to hear my old guys are still
chugging along. You are welcome to publish my circuit as it was on my
website as well. Let me know if you have questions.

Best-
Ron
Holmes Engineering