Fender Vibrato thumping

[retrorod]

I am servicing an AB763 Twin Reverb. Initially, the vibrato functioned fine. I had a power tube start to red-plate. I replaced that as well as all new screen grid res.
Now I get a lot of noise as I advance the intensity pot. When engaging the vibrato there is thumping (not ticking). This gets louder with an increase in intensity. The opto flickers with the speed control. I have replaced 2 of the three disc caps as well as placing a cap from 10 meg res to ground.
I have subbed in tubes and touched up solder connections.
I can't seem to find the issue.
Any suggestions, please.
It worked fine until it didn,t.

[AlNewman]

Why did the tube redplate?

[tubeswell]

If it worked fine before you worked on it, it was something you did.

[DeepBias]

I am servicing an AB763 Twin Reverb. Initially, the vibrato functioned fine. I had a power tube start to red-plate. I replaced that as well as all new screen grid res.
...
Any suggestions, please.
It worked fine until it didn,t.

Did you check the bias after replacing a power tube, are they matched. When you hear the thump, I suspect that the output tubes are shutting off because too much negative voltage at the grids.
 

[stratomaster]

I am servicing an AB763 Twin Reverb. Initially, the vibrato functioned fine...
I have replaced... 10 meg res to ground.
...
I can't seem to find the issue.
...

There is no 10M to ground in the AB763 oscillator.

Recheck your work.

[SEL49]

There is no 10M to ground in the AB763 oscillator.

When you chopped up his statement, you changed the meaning. He didn't say there was a 10M to ground. He said, "placing a cap from 10 meg res to ground." That means he connected one end of a cap to ground and the other end of the cap to the 10M resistor IAW Fender service bulletin #9. See attachment...

[8wattjack]

 Check lead dress at 5a. Sometimes you have to experiment with what works there. You may have to fly some of them in the air a little. What looks prettty doesn't always work. You can also try a different tube in that position, but I find that rarely works. Make sure everything else is in good order such as bypass cap, resistors etc. Did you change anything else? What caused the tubes to redplate?

[retrorod]

Why did the tube redplate?

They are very old 5881's. I assumed that one gave up and smoked a screen res.
I replaced all 4 screen resistors. Subbed in a set of used 6l6gc and re-biased. I am waiting on delivery of JJ6L6GC's

[8wattjack]

 I wouldn't assume anything when it comes to output. I would make sure all of that is rock solid before moving forward with anything else. Difference in current draw can be the tubes, or it could be a number of other things: PI coupling caps, screen grid and control grid resistors, tube sockets etc. The first thing I do is see if the problem follows the tube or the socket. Voltages/current draw in the power stage effect everything down stream. 

[stratomaster]

There is no 10M to ground in the AB763 oscillator.

When you chopped up his statement, you changed the meaning. He didn't say there was a 10M to ground. He said, "placing a cap from 10 meg res to ground." That means he connected one end of a cap to ground and the other end of the cap to the 10M resistor IAW Fender service bulletin #9. See attachment...

Good catch. That's what I get for skimming.

I've found the SB9 to help with tick but never thump.  I've found thump to be one of three things: lead dress coupling to/from the pots or the oscillator grid, a failing roach (age of either the bulb or the LDR breaking down), or a ground noise issue.  I read an article, where exactly escapes me, that stated the speed limiter resistor off the pot should be grounded at the same ground point as the cathode network of the oscillator. 

If all other troubleshooting attempts fail, please explore moving the ground.

Note: as part of the lead dress look at the grounded wire wound around the bundle of wires running to the front panel.  If it is grounded at both ends, cut one of the connections.

Update: found a screenshot I added to my tech notes. Now just need to find the original source.

Update 2: I also remember coming across a post somewhere that mentioned a plate to cathode cap (220pF) on the oscillator triode as done in the Twin Reverb Reissue amp as a potential tool against undesirable noise.  I haven't fully thought about how that might work in the frequencies we're interested in from the oscillator though.

[retrorod]

There is no 10M to ground in the AB763 oscillator.

When you chopped up his statement, you changed the meaning. He didn't say there was a 10M to ground. He said, "placing a cap from 10 meg res to ground." That means he connected one end of a cap to ground and the other end of the cap to the 10M resistor IAW Fender service bulletin #9. See attachment...

Good catch. That's what I get for skimming.

I've found the SB9 to help with tick but never thump.  I've found thump to be one of three things: lead dress coupling to/from the pots or the oscillator grid, a failing roach (age of either the bulb or the LDR breaking down), or a ground noise issue.  I read an article, where exactly escapes me, that stated the speed limiter resistor off the pot should be grounded at the same ground point as the cathode network of the oscillator. 

If all other troubleshooting attempts fail, please explore moving the ground.

Note: as part of the lead dress look at the grounded wire wound around the bundle of wires running to the front panel.  If it is grounded at both ends, cut one of the connections.

Update: found a screenshot I added to my tech notes. Now just need to find the original source.

Update 2: I also remember coming across a post somewhere that mentioned a plate to cathode cap (220pF) on the oscillator triode as done in the Twin Reverb Reissue amp as a potential tool against undesirable noise.  I haven't fully thought about how that might work in the frequencies we're interested in from the oscillator though.

Thanks Strato, much to explore here. I also have a replacement 'bug' that I can try if all other methods fail. I will let you know 👍

[AlNewman]

The trem is tied to the bias.  As Deepbias pointed out, It would probably be best to weed out any possible issues in the power section. You replaced the screen grids, but diode, bias cap, coupling caps, grid resistors, supply resistors should be checked.  Since you had a tube go, I would personally tend to take precautions on the big picture and then narrow it down once you know you're safe.

[SEL49]

The trem is tied to the bias.

He's servicing an AB763 Twin Reverb. The trem has nothing to do with the bias.

[AlNewman]

The trem is tied to the bias.

He's servicing an AB763 Twin Reverb. The trem has nothing to do with the bias.

[/quote]

Thanks Sluckey.
So, where's the separation between the negative bias and the grids of the 12Ax7?

[SEL49]

So, where's the separation between the negative bias and the grids of the 12Ax7?

The negative bias is applied to both grids of the tremolo tube through a 2.2M. This ensures that both triodes are hard cutoff whenever the trem is disabled by the open contacts of the trem f/s. But as soon as you enable the trem by stomping the f/s a ground is applied to the 2.2M and both grids immediately go to zero volts dc. Applying the bias voltage this way ensures the oscillator will quickly snap into action. So, the bias voltage is only used to turn the tremolo off. It's not in play when the tremolo is on.

[8wattjack]

 After realizing who Sel is, whatever he says about a trem you can etch it in stone! Hope you get it sorted.

[retrorod]

After realizing who Sel is, whatever he says about a trem you can etch it in stone! Hope you get it sorted.

Whoowee!...after much probing and prodding, I discovered an open coupling cap(.1) to a plate on V4. It was allowing excessive DC to the intensity pot.
All is functioning again.
Thanks for all the helpfull suggestions and info everyone.

[AlNewman]

The negative bias is applied to both grids of the tremolo tube through a 2.2M. This ensures that both triodes are hard cutoff whenever the trem is disabled by the open contacts of the trem f/s. But as soon as you enable the trem by stomping the f/s a ground is applied to the 2.2M and both grids immediately go to zero volts dc. Applying the bias voltage this way ensures the oscillator will quickly snap into action. So, the bias voltage is only used to turn the tremolo off. It's not in play when the tremolo is on.

I had a think about it at work today, and realized that the negative voltage must shut off the oscillator, with the 2.2M as the buffer.  And then had to come back shamefacedly and admit it.
 

[AlNewman]

That being said....   If there was a coupling cap bleeding positive DC from the trem to the bias, would that not cause the power tube to melt down?

[SEL49]

If there was a coupling cap bleeding positive DC from the trem to the bias, would that not cause the power tube to melt down?

Look at the schematic. There is no possible way for a positive voltage from the trem circuit to ever get to the bias circuit. And keep in mind, the trem circuit is isolated from the bias circuit by a 2.2M resistor. There is no possible way to get any positive voltage to that 2.2M.

There is always full bias voltage on one side of the 2.2M but the other side of the 2.2M is connected to chassis ground through the F/S when the trem is enabled. When the F/S is open the bias voltage passes through two 1Ms to both grids.

While looking at the schematic, imagine that any coupling cap is completely shorted. Do this for every coupling cap in the trem circuit. Now tell me which cap you think could possibly bleed positive voltage to the bias line.

[AlNewman]

Look at the schematic. There is no possible way for a positive voltage from the trem circuit to ever get to the bias circuit. And keep in mind, the trem circuit is isolated from the bias circuit by a 2.2M resistor. There is no possible way to get any positive voltage to that 2.2M.

There is always full bias voltage on one side of the 2.2M but the other side of the 2.2M is connected to chassis ground through the F/S when the trem is enabled. When the F/S is open the bias voltage passes through two 1Ms to both grids.

While looking at the schematic, imagine that any coupling cap is completely shorted. Do this for every coupling cap in the trem circuit. Now tell me which cap you think could possibly bleed positive voltage to the bias line.

Except There's a 3M pot as a variable resistor.  So, it's basically a 50% voltage divider when the pot is turned at full resistance.  If the .02 cap on the V4A plate, and the .01 cap between the 1M resistor and the .02 cap are both failing, I don't see how there's no voltage differential at the bias, with the trem off.

[SEL49]

So now you have two failed caps! You're grasping straws!   

So what. Let's just assume both those caps are dead shorts. That would put 300v plate voltage on the 1M resistor. That 1M is in series with the 2.2M just to get to the bias line. Now on the other end of the bias line is a 10K bias pot and a 27K resistor. These four resistors make up a voltage divider. So 300v across 1M plus 2.2M plus 10K plus 27K. Do the math and you'll see that very little of that 300v can ever appear at the wiper of the bias pot. I calculate about 3.5v. Don't forget there's also -52v on that line, so you would end up sending -48.5v to the 6L6 grids. Ain't gonna melt any 6L6s.

Meanwhile, retrorod has fixed his amp. I'm moving on.

[AlNewman]

Alright, let's see about that.
So, -52V split at the bias pot = 52/32K = .0016A.
.0016 x 37K = -59.2 V
300 minus -59.2 = 359.2V differential.
359.2/3.2M = 0.00011A
0.00011 x 37K = 4.07V
Your math lines up.
Thanks for the lesson.