5E9-A Depth control and PI bias

[stratomaster]

All,

I'm working on an April 1957 Tremolux but am new to Tweeds.  I've gotten the circuit to where the voltages are matching the annotations on the layout for the most part, but there's some odd behavior in the PI when the tremolo Depth knob is adjusted regardless of if the oscillator is active or not.

Cathode voltage climbs as the Depth is turned up along with an audible background thump.  Additionally, plate voltage goes way up and the drop across the plate resistors goes to about 15v from closer to 100v (don't have exact numbers in front of me at the moment). As you can imagine the sound gets very compressed and thin.  This is also very visible on a scope.

I posted about this on the telecaster forum first and spoke with who I believe is our pdf64 over there about it.  We seemed to have arrived at this being the nature of the beast.  I wanted to pick the brains here of those that have experience with this circuit if this behavior is expected. 

Would a fairly large capacitor at the Depth pot output lug help keep DC from the tremolo circuit off the PI cathode?  I'd imagine it would need to be fairly large due to the low frequency oscillation.

Finally, even a clean sine looks odd at the output on a scope. The negative swing looks compressed (squatty and round) at nearly all volume and tone settings on the high input of both channels (input: 400Hz and 1kHz at 108mV).  It sounds awesome, but doesn't look healthy on the scope.  Is this also the nature of the beast, or have I missed a fault?

[tubeswell]

The LFO wiggles the bias of the PI, which is a paraphase inverter. Paraphase inverters are characteristically set up so that the output signal is more balanced when the signal is ‘clean’, but a side effect of the design is the balance varies as the signal increases. This makes for a nice bluesy amp under ordinary circumstances but when tremolo is inserted to vary the bias (in this case, on both sides of the inverter), this changes the output balance as well as the gain. The 5E9 didn’t last as long as the 5G9 in the Fender catalog

[HotBluePlates]

I'm working on an April 1957 Tremolux but am new to Tweeds.  I've gotten the circuit to where the voltages are matching the annotations on the layout for the most part ...

The more I compare real vintage amps to the schematics/layouts, the more I find the voltage figures do not match.

They seem to be "close enough" numbers most often, and sometimes are mathematically impossible.

... even a clean sine looks odd at the output on a scope. The negative swing looks compressed (squatty and round) at nearly all volume and tone settings on the high input of both channels (input: 400Hz and 1kHz at 108mV).  It sounds awesome, but doesn't look healthy on the scope.  Is this also the nature of the beast ...

This is normal IMO, and one of the reasons I like the paraphase inverter.

[shooter]

It sounds awesome, but doesn't look healthy on the scope.

when I showed up here from the world of precise RF waveforms, and digital timing in sub-microseconds I was quite shocked at what a "great sounding amp" looked like on a scope, both time based n frequency based, signals like that woulda got me fired on day one in my field 

[Willabe]

This is normal IMO, and one of the reasons I like the paraphase inverter.

The hot amp for a while now, back ordered for a year or more (?) is the Benson amp, uses a paraphase inverter.

Schematic in 1st post;

https://el34world.com/Forum/index.php?topic=27092.msg297284#msg297284

[stratomaster]

The Depth pot was bad.  I replaced it with a 250k log pot (instead of linear) and tweaked the tapering resistor to 220k to retain the signal at the upper end of the rotation.  Now the effect is audible, easier to dial in, and the Depth pot can't bias the PI basically into cutoff.

The minimum Depth is almost too much effect. I'm considering a 500k pot (since it'll be a non original replacement anyway) in its place with experimentation to select the tapering resistor--or a resistor in series with the pot, say 100k for starters.