Paraphrase Phase Splitter

[Toalpo]

Hi!

Working in a Peacemaker Custom 100 which is not quiet at all when the standby switch is not even engaged. The sound is not perfect, but still you can hear the guitar and the master pots do control the signal amplitude. The standby swith has nothing to do with the HV, indeed all the tubes have HV on their plates as soon as the power switch is on. What the standby does is to short the phase inverter outputs wich should produce a cancelation of the signals but that is the problem. The phase inverter is a paraphrase phase splitter circuit and it works great when the standby is pressed but no so good when it should be quiet.
So far what I have seen is that the second triode does not invert in standby so the shorted outputs do not cancel and you can hear it since it is amplified by the power stage. I have isolated the paraphrase from the signal chain just wondering if the masters configuration got something to do but the result was the same, changed the tube with no change in the performance....

I am drawing an schematic since I did not find any info about this amp, sorry it is so poor any hint is welcomed.

[terminalgs]

paraphase inverters I've seen have R5 to ground, and R4/R6 aren't connected to R5, but are a 2 resistor divider to attenuate signal for the V2 grid input to match the amplitude of what V1 grid was fed. 


with what you've drawn and described I would guess is that the network of three 1M's don't properly attenuate V2's input to match what V1's input happens to be, so the output of each plate isn't matched amplitude wise.   If V1 & V2's signal output isn't a close match, (1) you'd have some audible signal when in stand-by and (2) when not in stand-by,  push-pull output wouldn't be all it ought to be.


I would expect  R6 to be about 20K or so...

[PRR]

Where is this "Master Volume"?

There are many things called "paraphase" (note not paraphrase). This is one of the floating or "self-balancing" schemes. Indeed it may not cancel well.

http://www.valvewizard.co.uk/paraphase.html

How loud is the "Standby" volume? If not loud, I'd put it down as a unique quirk and leave it be.

The R4 R5 R6 network is already a too-high resistance for power tube grids. If a MV pot adds more, you are at risk of burning power bottles.

[Toalpo]

Thanks for the feedback. This is the original circuit, so I guess this was working fine until now that this guy told me to review it. The sound can be loud enough to be anoying, it shouldn't be like this. The master is placed before the inverter. I did isolated from the master and still the same result.

But what I do not get is how come it works perfect when the standby switch is pressed and in standby V2 is not inverting and the two outputs are perfectly equal and in phase. It seems as if V1 output would go through R5. I have attached a pic of the circuit, as you might see both outputs are shorted by the switched. I did disconnect R6 with no change.

If still uneable to resolve this issue.....I am pondering modifying the circuit and using the standby swith to ground the V1 input signal through a 1K resistor. What do you think? dumb idea?

[PRR]

> the two outputs are perfectly equal and in phase.

I asked the Idiot Assistant (simulator). Yes, shorted you get identical (of course) outputs at about 1/50th of the amplitude of "Normal". Maybe a "Dim", not a "Mute".

Quite radical changes of value and topology make hardly any difference.

The "50:1" is the clue. This is the gain of a 12AX7. The inverter should cancel all output, but only as its gain. I inserted an ideal gain stage of 10 between the 1Meg network and the inverter tube. Now it gave 500:1 rejection. But there is no simple way to do this in practice, and it is a radical change from the simple Floating Paraphase.

The Power stage "should" take two equal but opposite inputs and *cancel* in the output transformer. That may only be a 10:1 null. But a 50*10= 500:1 reduction should be awful quiet. I wonder if you have a serious unbalance in the power stage.

That appears to be a double-pole switch? Using two poles to ground *both* power tube grids should kill signal. But I see other wires? Even so, 3-pole switches exist. Or if the 2nd pole is for a lamp, it might be fed otherwise.

[Toalpo]

Thank you PRR, really good reply. That makes completly sense, I have to change my focus and review the power stage. It uses four EL34s but have to check the circuit to extract the schematic.

The standby-switch is a 2 pole, one is used to short the paraphase outputs and the other one to feed the siwtch lamp. So far, finding this standby "method" pretty interesting, never seen it before.

Hopefully in the following days I will find some time to check the power stage and let you know.

[Toalpo]

PRR you hit the nail on the head. The power stage was badly unbalanced. Just after the first bias adjustment the improvement was noticable. Still some output when shooting up the preamp which reduce after doing fine adjustmen of each tube. Cranking gain and master up will still pass some low audible distorted signal which I think is acceptable.