"Kirk Patton" wrote lots of good stuff!

Excellent Kirk, thanks for the most credible advice I've ever had
about anything. Could have done with a lawyer like you. You are
immortalised in my library. Thanks too for the links.

Setting off for Hull, UK, where I believe the boy might be
found...perhaps this time there will be traffic on the bridge, so I
won't be too scared to cross on my own.

While I'm thinking of rack-mount valve amps, it is important to use
good PS transformers that won't vibrate. A rack cabinet can make a
very effective sound-box for a humming transformer.

cheers, Ian

Hi Ian,

This sounds like what I usually call a "pin-1 problem" . . . named
after the
ground pin on an XLR connector . . . it is way too common in
commercially-produced gear. Basically, what happens is that
different
stages inside a given piece of equipment are grounded in different
places,
and when current flows from pin 1 of one XLR to pin 1 of another
XLR, the
accumulated resistance in connectors, wires, and P.C. board traces
is
sufficient to allow a small AC voltage to develop, and if an
amplifying
stage's input is referenced to the wrong part of the ground trace,
then the
AC voltage is amplified as humm.

The reason for the ground current flowing in the first place is
leakage
current onto the chassis from the power transformer's
primary-to-secondary
reactances. I'm assuming two pieces of audio equipment here -- a
synth
sound module and a tube amp, the tube amp has a grounded cord and
the synth
doesn't, and there is an unbalanced audio connection between the
two pieces.
The symptom as described is that there is no hum present in the
system,
until there is a very low-resistance (rack flange) connection made
between
the chassis of the synth and the chassis of the tube amp.

So here are some of my guesses as to the problem:
1 - Internal grounding mistake in the design of the synth sound
module.
Wouldn't surprise me a whole lot . . . PC board designers have off
days too
. . .
2 - Less likely: Your input stage/phase splitter star ground point
has a
lower impedance to the chassis than the one that has the mains
cable. In
this case, the ground current from the synth chassis would flow
through the
rack, to the amp chassis, to the input star ground, through the
wire that
connects the two star points together, and to mains ground. The
resistance
of the wire between the two star points would cause a voltage to
appear on
the grids of the output tubes. Only problem is, that it would be
mainly
common mode (though not perfectly balanced) and would also be
bypassing the
higher (voltage) gain stages in the amp. Hmmmmm.
3 - More likely than #2: Your input jack isn't isolated from the
chassis,
and is actually making a better connection to the synth's chassis
than your
star points, from the point of view of the rack (grounded to front
panel
with rack ears, while other star point is at the back of the
chassis?) In
this case, the ground current from the synth chassis would flow to
the input
jack, to the first star ground, (and then maybe to the second star
ground
via either the wire or the chassis), and to the mains cable
ground. The
resistance of the input jack ground to the first star point in
this case
would be the primary mechanism, resulting in a hum voltage
directly on the
input.
4 - Humm is a side-effect of an oscillation. Maybe not terribly
likely;
adding a grid-stopper resistor on the input tube may be a good
idea if there
isn't already one there.

If the situation is as I described it, then I would lift the
chassis ground
on the second star-point, and ground the mains cable to its own
point, with
a low impedance to the rack mounting ears. I would also isolate
the input
jack, and make sure that there is a separate, low impedance ground
from your
first star point to the rack ears/front faceplate as well.

also, check out:

http://www.jensentransformers.com/an/an004.pdf

http://www.jensentransformers.com/as/as032.pdf

for insights into these types of hum mechanisms.

Hope this at least gets you pointed in the right direction,

Kirk Patton