A strange signal artefact with a 1970 Fender Vibrolux Reverb.

[Helmholtz]

Scope pics of the effect/artefact (was tricky with my old camera):

Signal at amp output close to clipping, note upper peaks:



Square wave signal at amp output:


Signal at rev. driver grid, note negative peak:



Square wave at rev. driver grid:



Signal at rev. driver plate:



Square wave at rev. driver plate, note positive signal collapsing after a short high peak.



At pictures taken with same input level and amp settings.

[Chuck H]

Regarding Pete's consideration...

When a master volume is in use there will be considerably less voltage sag mitigating available preamp gain. In my experience this can make a noteworthy difference in preamp voltage swing AND how it peaks.

[strato56]

In the SF135 Fender puts a 560pF/3kV capacitor between plate and cathode of the 12at7 : could that be a solution ?

[Helmholtz]

In the SF135 Fender puts a 560pF/3kV capacitor between plate and cathode of the 12at7 : could that be a solution ?

I tried that. It makes the artefact lower and smoother.
But I'd expect some treble attenuation in the reverb drive with the cap.

I don't consider the effect a problem since I verified it's not a stability issue.
It's probably typical with BF and early SF reverb amps.

[g1]

Agree with you that it is not an issue but just an interesting artifact. Fender reverb is pretty much the gold standard, not something you would want to mess around with if there is no audible problem.

[strato56]

You tried it with no load ie tank disconnected.
Have you tried replacing the tank input with a 8 Ohm resistive load ? ( Assuming your reverb tank has an input impedance of 8 Ohm )

[pdf64]

…
What makes you suspect reverb driver clipping?
…

If a Type 4 master vol is fitted to a typical Fender black panel reverb channel, my finding is that at lower settings, the earliest clipping point in the preamp is the reverb driver valve grid. The clipping there affects the dry signal path and sounds bad.
If the reverb driver valve is removed (or a suitably high value grid stopper is fitted, or the cathode bypass is removed), the nasty clipping disappears (and the earliest clipping point moves to the 3rd stage grid). The clipping is asymmetric and reduces as frequency is reduced

[Helmholtz]

You tried it with no load ie tank disconnected.
Have you tried replacing the tank input with a 8 Ohm resistive load ? ( Assuming your reverb tank has an input impedance of 8 Ohm )

With a resistor in place of the transducer everything is clean, no artefact - as expected.

The artefact is caused by single sided transducer core saturation, which makes the load impedance drop from 8 Ohm to 1 Ohm (i.e. the DCR of the coil) for positive plate signals.

I would not expect the artefact with a non-saturating core. In fact the artefact disappears as soon as the transducer comes out of saturation when lowering the input signal.

[Helmholtz]

If a Type 4 master vol is fitted to a typical Fender black panel reverb channel, my finding is that at lower settings, the earliest clipping point in the preamp is the reverb driver valve grid. The clipping there affects the dry signal path and sounds bad.
If the reverb driver valve is removed (or a suitably high value grid stopper is fitted, or the cathode bypass is removed), the nasty clipping disappears (and the earliest clipping point moves to the 3rd stage grid). The clipping is asymmetric and reduces as frequency is reduced

Are you sure, you're actually seing rev. driver grid clipping?
At some frequency the core saturation effect actually looks like grid clipping (on negative grid peaks with my tanks) and that starts way before grid clipping.
In case of doubt try disconnecting the tank.

Real grid clipping would show on positive peaks of grid signal only.

What exactly is a type 4 MV and why does matter here?

[Helmholtz]

Rev. driver grid signal with transducer saturation effect, signal level about same as with earlier pics but frequency lower:
This is NOT grid conduction clipping.



No idea why the pic is rotated.

[Pedro Vecino]

What exactly is a type 4 MV and why does matter here?

Master volume type 4 (it is the denomination that appears in the Trainwreck pages) is the one inserted between the output of the treble control and the input capacitor to the phase inverter. The most conventional.

[Helmholtz]

Ok, curiosity won.

I lifted the cathode cap: No more transducer saturation, no more artefact.
But drive level lower by a factor of 2.6 or 8dB.
Means that recovery gain needs to be increased by the same amount, which will increase hum and noise.

But listening tests will have to wait.
First I want to find out if the 2nF caps at the power tube grids are necessary for stability.
They seem to cause slew rate distortion above 3kHz at high output.

Meanwhile I think that the single sided saturation is not caused by permanent core magnetization (checked with my Gauss meter) but rather by the asymmetric plate signal, which in turn causes asymmetrical core magnetization.

[pdf64]

This shows a Type 4 master vol, ie fitted between preamp and power amp https://el34world.com/charts/Schemat...mv_noboost.pdf
It matters only in that fitting one would demonstrate the effect I’m referring to.
Due to that effect, the preamp overdrive of this design sounds bad, yes, even worse than the later pull boost version.

[pdf64]

Ok, curiosity won.

I lifted the cathode cap: No more transducer saturation, no more artefact.
But drive level lower by a factor of 2.6 or 8dB.
Means that recovery gain needs to be increased by the same amount, which will increase hum and noise.
…

Ok, rather than totally lifting the cap, try 470ohms in series with it.
I’ve seen the viewpoint that later silver panel reverb is better than that of the black panel version, with a fully bypassed cathode.

[Helmholtz]

Due to that effect, the preamp overdrive of this design sounds bad, yes, even worse than the later pull boost version.

Yes, I know it sounds bad.
But are you sure it's due to rev. driver grid conduction?