signal stability super reverb ab763

[pbiagi]

While researching why the amp tone was raspy - I believe its a speaker which I'll change out tomorrow -
I fed a ~400Hz sine wave and noticed that B wave coming out of the phase inverter (see photo) jumps up and down while the A wave remains stable.
Maybe that's nothing but curious what it's about. Appreciate any info.

[guitician]

The B side .1uf cap could be leaking DC. How stable is the DC on that side of the 220k resistor pair?

[Enzo]

This was using the dual trace display? Use one channel of the scope and watch the point that jumps, does it still appear that way now? Also, in dual mode, try adjusting the trigger threshold. The scope triggers on just one of the waveforms, so the other waveform could be occasionally losing sync.

That may have nothing to do with it, but you need to rule it out.

[Jazz P Bass]

That trace looks like it is not syncing up.

[g1]

I think that is 2 channels of scope showing out of phase relationship of PI plates, but lined up one on top of the other?
It would be better if you could show one trace on the upper half of scope, the other trace on lower half.
Indeed it does make it look like a single trace mis-triggered.

[Tom Phillips]

Also swap the scope probes to the other signal phase and determine if the same signal still jumps when it is monitored by the other scope channel.

[pbiagi]

Thanks for the advice.

Correct, the previous photo was 2 channels showing out of phase relationship of PI plates.
Swapping the probes displayed the same results.

To better describe, after the .1 uF capacitor from the plate on the B+ 100K resistor side (as opposed to the 82K side) the wave form super-imposes several traces which are not aligned on top of each other.
The 82K wave traces more closely align themselves on top of each other.
Here's a photo of the 100K side by itself.


Taken from the same spot the DC voltage does not drift much, constant around -51 VDC (I didn't write it down).

To measure for DC leakage, should I insert the DMM between the .1 uF capacitor and the power tube?
Or disconnect one side, and jumper in another .1 uF capacitor to view the difference?

[Tom Phillips]

...To measure for DC leakage, should I insert the DMM between the .1 uF capacitor and the power tube?
Or disconnect one side, and jumper in another .1 uF capacitor to view the difference?

Just lift the sides of the exiting 0.1uF caps that feed the 6L6s. Then turn the amp on and monitor the voltage on that free end of the cap with respect to chassis ground. For a good cap you will see a voltage indication as the cap charges and then the reading will settle to 0V. (The indication may flip between + & - too). You will have two caps to compare.

That's how you quickly check for coupling cap DC leakage with the cap in the circuit. The amp's power supply provides the DC test voltage.
However, I don't feel that your symptom would be caused by DC leakage.

[guitician]

Yeah, AC is more apt to "wobble" that other trace, so then a bad tube? 60hz or 120hz wobble/ripple?

[g1]

Maybe from the NFB loop?

[pbiagi]

Quick update - I tried another tube yet the exact same wobbliness from the same PI plate is displayed.
The trace from the PI plate's grid is fairly stable.
Therefore it seems the wobble can only come from the .1 uF cap or 100K resistor (which in circuit measures 92K ohms) or - thank you g1 - the NFB loop.
It may be a day or two but I'll start by disconnecting the NFB loop.

[loudthud]

The signal frequency looks to be about 400Hz. Slow the frequency down to about 2x the line frequency to see if it syncs up and the wavyness freezes or moves slowly across the wave. Try adjusting the "Holdoff" control on the scope to see if the wavyness will sync up. Do you have a good ground lead on both probes?

[Enzo]

Did you ever go back and using just one probe, watch the offending channel to see if it still hops around?

[pbiagi]

Update - sorry for the long post!

For a single probe, see my post May 17.
This is still the case - using a single probe on the PI after the .1uF cap from pin 6, the wave is stable, not traveling across the screen, but the traces are not aligned on top of each other. See the May 17 post.
The PI .1uF cap from pin 1 (82K side) aligns better but does not perfectly align either.
I'm not sure this is an issue or not, but I also noticed today that the auditive quality of the sine wave is drifting as well. I don't think this sounds well?

Some background -
This is a vintage '65 super reverb amp (with the original Oxford speakers), which I've always felt did not sound well.
Over time I've found some major - broken resistor - and minor things wrong and fixed them.
The tone has improved but is still not deep and the high end is squashed, a little annoying, cannot pick out the individual strings.
One of the Oxford speakers have been reconed and I suspect it might be part of the problem. Next I may substitute a 10W 8 ohm resistor for it.
However I've put the amp through a bassman cabinet which has the same tone issue.

Work done since my last post -
The amp has some hiss which gets much worse if the NFB is removed.
For that I replaced the plate load resistors with metal film, including the PI resistors except for the 82K ohm resistor I only had another carbon comp. No real change.
Then tracing the signal through the Vibrato channel, starting at the preamp I grounded the grid at each stage to see if the hiss would go away.
The hiss is not from V2, but at V4 when the grid, pin 7 is grounded, the hiss is entirely gone.
Does it sound likely the 3.3M ohm grid stopper is the culprit?

Back to the wave stability issue -
In France 50Hz, this is the bias ripple.
Cap is 100uF 150V.


Bias ripple vs sine wave drift. (One trace on top of the other)


Both sides of the PI plates, with a 3rd line, the addition of the waves, displaying the drift.


Since there was 3mV DC measured on the grid of V4 pin 7, I replaced the .02uF and the 500pF caps on V2. Note the previous .02uF cap was microphonic.
I also changed out the .1uF coupling cap on V4.
Result was 0mV DC on V4 grid.

Leaky caps -
I didn't write these down (should always do that!) but from memory -
PI coupling caps were disconnected on side going to power tubes and measured to ground.
.1uF from pin 6 jumped around 50 - 90mV, never going to 0.
.1uF from pin 1 jumped around 30 - 60mV, never going to 0.

I jumpered in an orange drop .1uF cap and it also similarly jumped around, never settling to 0..
What does this mean?

Voltages -
V1 (normal channel), V3 (reverb) and V5 (tremelo) are removed.
Some voltages were taken before and some after the change-out of the 3 coupling caps.
V2
Pins
1: 258
2: .3mV
3: 2.34
6: 269
7: 0.0mV
8: 2.16

V4
Pins
6: 287
7: 0mV
8: 2.32

V6
Pins
1: 258
2: 62
3: 99
6: 252
7: 64
8: 99

V7 & V8 (7581A)
3: 457
4: 459
5: -51.0

Thanks in advance for your advice.

[pbiagi]

Sorry my previous post was long and covering different subjects.

Since then, improvements.

The one side of the PI plate sine wave rising and falling for me is an interesting curiosity. I'm not sure there's a real issue though.
I must note that the oscilloscope traces were done with the speakers plugged in.
The volume was low, a little over 2 1/2, treble, middle and bass all at 5.
Later I ran the same trace with a dummy load and raised the volume to 3 1/2 (from memory). Both sides of the PI plate sine wave appear more stable although that could be because the wave form is much larger, more concentrated on the screen. I did not zoom in on an upper edge..

Replacing the 3.3M grid stopper (?) with metal film appears to remove hiss, although there's still a little left.
I could go through grounding the grids to determine where the remaining hiss is introduced however there's a new issue.

The amp would not turn on. After searching, at the wall socket side of the power cord, 119 VAC, but power cable side in the amp was 25 VAC. ?
Turns out the neutral on the power chord was no longer making a connection.
It's a 3-prong chord but very old and cheap. When I opened the plug, the blue line fell out and was not well screwed in. The wire had solder on it (although it's a screw in socket) but the solder/wire appeared dirty. Likely it was not making a good connection for a very long time.

Frustrated. Similarly a year ago a plate resistor was found to be intermittent, making no connection at all.
I'm taking a day or two break from it.. but I know patience will result in good tone.