Why do I get this spontaneous Tremolo? (And how to get rid of it?)

[voskarp]

This homemade "AC4" where I've combined the preamps from two different amps has a weird quirk. When the volume knob of the EF86 channel is set over 3/4 a tremolo effect starts to set in as the note fades (sample 1). Sample 2 is playing though channel two (ECC83) at low volume, with the EF86 channel volume set on max.

https://soundcloud.com/oskar-pohjola/ampvibsample1

https://soundcloud.com/oskar-pohjola/ampvibsample2

I believe this effect started when I vent from version 1 to v.2. (In v.1 the gain from the EF86 channel was very low, so maybe it "was there" but never showed because it comes with high gain.)

AC4v.1.png

•AC4v.2c.png

This is not a huge problem since there is no reason I need to dime the gain of the EF86 channel, it gets crunchy enough at about 1/2 anyway. But if I can get rid of it, that would be the best.

[ChopSauce]

May that show the limits of your power transformer; what is it capable of?

NB: I have just started investigating that issue, but it seems that the 5M6 screen grid resistor matches the tremolo. Without the tremolo, this resistor - which you set at 1M, rather - might well need to be altered, yet I still can't figure exactly what happens in both cases. Did you made the calculations, as you wrote about some hi gain from the EF86, or is it just sound-wise?

[voskarp]

If it was the power transformer, there would be sag at max signal, but this effect comes quite opposite when the signal wanes.

[ChopSauce]

I had to Google/translate "signal wanes" but still I don't understand: doesn't the effect appears when pushing the EF86?

[voskarp]

It occurs when the volume (gain) is more or less maxed, but the signal from the guitar is fading. (As in the sound sample).

[ChopSauce]

Can't that R13 create some feeback between preamp tubes - which could explain - if it's not a power-supply related matter?
(just curious of its use, really a naïve question)

EDIT: afterwards, I realize I didn't really understood what your circuit actually does.

[martin manning]

Maybe try powering the EF86 stage from B+3 instead of B+2?

[voskarp]

Yes, that could be it. I guess I should reduce R25 then.

[ChopSauce]

Why that? The original schematic shows R19=1k & R23=22k. The departure with your schematic being one reason why I asked about the PT being possibly slightly weak?

[voskarp]

I reduced the resistors to get higher voltages at the valve plates, I found them to low.

The PT is from an AC4C1 with a top boost circuit, so it used to power two ECC83s and an EL84.

[voskarp]

I moved the EF86 power supply to B+3 (and changed R25 to 10k), and voilà! -tremolo effect gone!

I'm glad I could solve this without tearing the whole amp apart!

[martin manning]

Great! So it was a power supply rail decoupling problem.