Ian Iveson wrote:

"tomrp" wrote

It gives me a place to start.

or maybe not...

if it is a dynaco III clone, then it is a combination of fixed and
cathode bias, with the two cathodes sharing the same, unbypassed
bias resistor. Adjustment is through an additional "fixed" bias
supply to the grids.

In that case, if one valve switches off, the voltage across the
cathode resistor drops, so the other valve will pass more current.
In this respect my previous argument still applies.

However, if the adjustment is to the fixed bias part of the circuit
like the dynaco, you may really need to trace the circuit, because
the dynaco lacks a DC balance control as far as I can see
http://www.triodeel.com/dynamk3.gif

It may possibly be a dry joint or poor socket connection or a short
to ground between the bias circuit and one grid. That valve may then
pass enough current for the common cathode resistor to turn the
other valve off.

Conversely, it could be a dry joint or poor socket connection to one
cathode as before, which would tip the balance the opposite way.

Could be lots of other things too.

If you can check whether the pots are associated with the cathodes
or the grids it would be a help. Much better if you could trace it
out, or tell us how it differs from the dynaco. Grids are more
likely I think. Could conceivably be both...

The point I am labouring is that, if the valves are balanced like a
see-saw across a shared cathode resistor, then your symptom can be
caused in four opposite ways: either side can be pushed or pulled,
and the result appears the same. Combine this with an intermittent
fault and things get tricky. Without a circuit you could have an
intractable problem.

cheers, Ian

He has the circuit.

But he has no schematic apparently.

My way if I'd never seen the amp before in my life would be to
catch it in the act of playing up while its upside down on the bench,
and prod around measuring evrything quickly, lest something fry,
before I could catch it in the act of frying.
If the bias current changes for either tube suddenly, and another pair
of tubes
also does the same thing, it ain't the tubes, its most likely the bias
circuit,
or a stuffed coupling cap.


Sometimes you can get an intermittent tube socket.
90% of tube sockets faults are due to the pin grippers working
open, and not gripping the pins tightly.
Usually its the grid pin that goes open, and the tube is unbiased, and
it
then starts to develop a positive charge on its grid, and the tube
current begins to climb,
and in a minute or two you have a red hot anode, and possible tube
meltdown.
spontaneous reconnection of the grid pin won't stop the runaway
because by this time the grid has become hot.
The amp must be turned off, and left a few minutes to cool,
then if the grid id reconnected, the tube may be OK.

Many old amps have had their tubes removed and plugged back in so many
times
that many of the grippers have become dodgy, and the tubes almost would
fall out
if the amp was turned upside down.
Get an awl and prize *all* the contacts of all grippers together a
little.
File the awl to a shape which fits down into the hole for the pin entry.

Don't over do it, and don't break the plastic socket material.
Ceramic sockets are prone to easy breakage.

But then if the fault didn't hit me in the eye, I'd trace the around the
circuit
and draw my own schematic of what is present in the amp in the bias and
output stage.

Armed with information, the fault would reveal itself.

If indeed it is like the schematic at
http://www.triodeel.com/dynamk3.gif
then if one tube stops conducting, the other will conduct more
as there appears to be a 112 ohm common cathode R.

And there is no way of balancing the currents of the two tubes.
Nice one Dynaco! it must have caused lotsa swearing by lotsa owners over
the years,
and maybe even premature tube death.

Quad II also have a common R for the cathodes of KT66,
and sometimes I see amps with 90mA in one tube and
35 mA in the other, and the sound is mud.

A better baising schematic is at
http://www.turneraudio.com.au/htmlwe...0ulabinteg.htm
Since drawing that schematic 4 years ago, I have found it more prudent
to use 100k bias R
instead of the 220k shown, since often 6550 tubes start conducting
some grid current even at idle, and instead of a slight negative grid
with respect to the
bias supply, you can get a couple of volts positive across the bias R,
which turns the tube on more, which causes even more positive charge to
develop at the grid
so tha even with say -40v at the bias pot the can be -38 at the grid.
Maybe KT88 suffere the same problem.
EHKT88, EH6550, and Sovtek KT88 are all identical internally.
I don't know if they start conducting grid current after say 1,000 hrs
of use,
but of twelve JAN GE6550 type A tubes I bought in 1995,
about 4 began to run about up to 2 volts across their 220k bias R,
and then they seemed to stay like that.
Individual bias pots for each output tube is a good idea,
but the balance idea I use in the schematic above works fine,
and I have a two transistor circuit which takes a signal off the 10 ohm
cathode R
and if there is more than 5mA difference in bias current an LED turns
on.
as long as there is enough turn travel to balance the amp
so the led stays turned off, all is well, but if the
pot won't balance the bias currents, one or both tubes has drifted too
far,
and a new set is required.
LEDs weren't around when Dynaco was designed.
Using a twin triode and neon tubes for a balance indicator circuit was
too hard and costly.

Patrick Turner.