[TerryJ]

"Patrick Turner" wrote in message
...


TerryJ wrote:

"Patrick Turner" wrote in message
...


TerryJ wrote:

Opinions?
http://home.comcast.net/~suptjud/6BX7AMP.JPG

It sounds fine with both 6BX7s working or with either removed. I have
no
way
to measure distortion except the scope and my ears.

The two 82k equal value loads to the 6SN7 should not be equal to get
the
same amplitude
of oppositely phased drive signals to the 6BX7.
it looks like the 2 pairs of cathodes of the 2 x 6BX7 should also have
a 1,000 uF cap across each 500 ohms
to lower the Rout of the amp, but it isn't so, since each half of each
twin triode
is hooked up to work the OPT balanced so the circuit will work with
either 1 of 2 6BX7 used,
although there would be a power & distortion & Rout difference.

Its usually hard to make a triode amp that sounds poor, but the circuit
could be optimized
with better ac drive balance.
Try removing the 470k from one 6SN7 anode to 6SL7 anodes and strapping
whatever needed
R across the 82k nearest the 6SL7 so that + / - drive voltages are
equal
at 1/2 full output voltage into load.

Patrick Turner.



OK, I tried a few things.

1- Balanced the splitter by paralleling a 1 meg resistor with the left
82K
and removing the 470K. I could not hear a difference. Proper test
equipment
would probably find it. I thought about changing the 470K to about 680K
and
recheck the balance but did not. The 470K reduced the left triode output
to
about 5-6 % less than the right. With the 470K removed gain of the left
triode was about 5% more than the right.

2- Connected the input directly to the 6SN7 grid. There was not enough
gain
to drive the amp to more than half power but fidelity was very good....
very
listenable. Note there is no feedback with this arrangement.

3- Paralleled a 2200 ohm resistor with the 2700 ohm connected to the 6SN7
cathode and plugged a 6SL7 in the socket. Bias was -1.62 volts. Fidelity
was
good up to about half power then distortion crept in until at max volume
control (2V P-P drive) it sounded like a PA system.

4- Put 6SN7 in both sockets, returned second 6SN7 to values on
schematic,
increased 1500 ohm to 2200 and used only half of first 6SN7. First 6SN7
bias
is -3.8 volts. Without feedback gain was way too much but with input
adjusted fidelity was good up to clipping. With the 470-10 ohm feedback
network gain was still more than needed but fidelity good.

Looks like Patrick is correct. It is hard to make a bad triode amp.
BTW
the amp is stable with only resistors in the feedback path. There is no
sign
of oscillation or ringing with a square wave input. A square wave looks
good
down to 100hz and is still recognizable as a square wave up to 15 kHz. A
better output transformer might help. I am using a transformer from a 50
year old Hammond organ amp that had push pull 6L6 tubes.

Applying overall loop FB around tube amps with resistance loads is
usually
OK and without LF or HF oscillations if the FB is below 12dB which
is usually plenty with a triode output stage.

But if you have a capacitor load of say 0.22 uF, and no other load,
any tube amps will oscillate badly at say 200kHz,
or whenever the phase shift reaches 180 degrees caused by leakage L, the
load capacitance,
and shunt C between tube stages and where open loop gain exceeds 1x .
But if RL is always a resistance, then the amp will take the 12dB of
NFB, usually.
If an ESL type of speaker is used, perhaps there could be trouble with a
peaked response
or oscillations because of the C part of the ESL load.

That 470k between anodes of the 6SL7 and 6SN7 is a form of slight local
NFB
and it also loads the LH SN7 triode more than the RH triode, and its an
attempt
to balance the outputs.

Total loads for each half of the LTP are the dc carrying loads, 82k, and
the
following grid bias R for the 6BX7.

The LH SN7 load should be 82k plus a parallel resistance equal to the RH
triode gain x the total common cathode resistance.

Its simple this to work out and remember, and the higher the carhode R
value, the nearer
the loads need to be for balance.

At normal low volume levels an imbalance of 50% isn't often audible.
But if you want to set your amp up technically right, then ac balancing
should be done right, and once done, it doesn't matter if the halves of
the 6SN7 are unmatched;
balance will stay balanced as long as the tube gain remains within 10%
of the new tube value.
The main balance setting issue with an LTP is that the signal current
change in the LH load =
signal current change in the RH load + current change in the common
cathode R.
Hence to get the balance to be exact the LH load must always be a bit
lower value than the RH load,
so the imbalance of tube currents generates the same output voltage
amplitudes.

Plenty of schematics with LTP drivers at http://www.turneraudio.com.au

Patrick Turner.

Thank you for the very thoughtful replies.