Ampeg B-18 issue/debug

[aborgman]

Current flow through the tube would rise, and the tube is then a cathode follower when the pin is shorting the plate resistor.

You'd have to move the coupling cap to the cathode side for it to be a cathode follower.

The signal across the cathode resistor is in phase with the stage input, and would be way higher in the fault case than the normal feedback from the OT. Both are good candidates for an oscillation condition.

Yeah - any in phase signal getting injected into that gain stage is likely to induce oscillation.

Good call on the point contact behavior.

Yeah, I've seen that whole "poor contact makes a diode" thing a few times before.

 

[aborgman]

Well, I re-wired the beast.

I stripped the chassis bare and re-wired everything. All new components.

Only thing I re-used were the tube sockets, input jacks, the ext. amp jack, and the volume/tone pots.

Replaced the speaker jack with an isolated model.

Replaced every other cap/resistor.

Replaced all of the shielded wire with better quality shielded wire, and re-routed the tone/volume wires so they avoided going directly over the top of the output transformer on their way to/from the 6SL7s.

Added grid stopper resistors to both grids of both 6SL7s and a grid stopper to the grid of the pentode section of the 7199.

....and there is absolutely zero change in the problem.

None.

Not a bit.

$100 worth of parts, a dozen hours of work... and nothing.

B+ voltages look good (512/500/480/405), power tube bias voltages look good (-52V).

Swapping power tubes has no effect. Swapping/removing preamp tubes has no effect.

Bad tube? If so, seems like it must be the 7199 because moving/removing other tubes has zero effect.

Bad tube socket? Bad OPT? (Ged I hope not) Any other ideas?

As of right now, my next move looks like a full tube replacement... but I can't say I have a lot of confidence.

 

[BassmanPaul]

Sorry but I think you have shorted turns within your OPT. Do you have a spare OPT of some kind laying around. You can temporarily bypass the existing OPT and see if the problem goes away. Mercury Magnetics will rewind and repot the old one for not all that much money.

 

[aborgman]

Sorry but I think you have shorted turns within your OPT.

Uggg... the one thing I was abjectly hoping wouldn't be the problem is the OPT.

Is there any reasonable way to test it?

Do you have a spare OPT of some kind laying around. You can temporarily bypass the existing OPT and see if the problem goes away.

I have a few small SE OPTs sitting around, but no PP transformers of any sort.

Mercury Magnetics will rewind and repot the old one for not all that much money.

Oh, for the days when I will think $250 isn't all that much money...

 

[BassmanPaul]

Oh, for the days when I will think $250 isn't all that much money...

To get custom work done it isn't that much and you get a better transformer coming back than you sent.

 

[aborgman]

To get custom work done it isn't that much and you get a better transformer coming back than you sent.

Yeah, its really is reasonable for custom work... but it's still a lot of money. Almost as much as I paid for the amp.

...never mind that I might be replacing a perfectly functional OPT, unless I can come up with some reasonable way to test it and determine it's bad.

 

[BassmanPaul]

Feed a low voltage ac signal into the secondary and scope what is coming out of the primary. You can use a signal generator or the output (fused) from a 6.3V transformer. Be careful doing this as the OPT, even if faulty, will boost the voltage. Measure the voltages from CT to each side of the primary, they should be the same.

 

[aborgman]

Feed a low voltage ac signal into the secondary and scope what is coming out of the primary. You can use a signal generator or the output (fused) from a 6.3V transformer. Be careful doing this as the OPT, even if faulty, will boost the voltage. Measure the voltages from CT to each side of the primary, they should be the same.

Tubes removed
Scope grounded at OT primary CT
OT secondary driven with function generator, 1KHz, .1V ->8V.

Output voltage across primary are ~10x (peak to peak) the voltage being driven into the secondary.

i.e. - If I drive the secondary with 5 Vpp, I see ~50 Vpp from each primary winding to CT, one winding out of phase with the other.

I see equal voltage on each side of the CT, and I don't see any odd distortions or other strangeness.

This sort of seems to imply the transformer is ok, unless I'm just not reaching high enough voltages for an internal short to present itself.

 

[BassmanPaul]

It does seem that way but, as you say, the voltage IS low. It would seem that perhaps substitution may be the only way to be certain in your case. It's what I would try however I have a lot of suitable iron floating around in my workshop. It really is hard to help when I can't see the unit.

 

[aborgman]

Here is my next thought for debugging -

Swap the OT primary connections on the power tubes.

I currently only see the distortion on the low half of the wave. Swapping tubes doesn't move it - so it isn't a particular tube.

If it's part of the circuitry or tube socket - it should stay on the lower half of the waveform. If it's the OPT - it should shift to the other half of the waveform.

 

[BassmanPaul]

In order to do this you'll have to disconnect the NFB loop otherwise all you'll have is an oscillator.

 

[aborgman]

In order to do this you'll have to disconnect the NFB loop otherwise all you'll have is an oscillator.

No doubt... nothing like turning the negative feedback loop into a positive feedback loop.

 

[aborgman]

New and strange revelation.


So I was disconnecting the NFB loop in order to allow swapping of the power tube plate connections - and I decided to see what the waveform looked like with the NFB loop disconnected.

No other change, JUST the NFB loop disconnected.

It'll hit ~52 watts before it starts flat topping the wave. No sign of the goofy distortion/oscillation on the lower half of the waveform.

So there is definitely SOME sort of issue involving the NFB loop.

This unit does not have the original speaker cable, and is wired so the speaker jack is connected to the 8 ohm tap.


It's wired as follows:

OPT yellow (8 ohm tap) -> speaker jack +

OPT black (common) -> speaker jack -, PI gnd (for NFB loop reference)

OPT green (16 ohm tap) -> 22K resistor -> 7199 pentode stage cathode

 

[JimmyM]

beats having a problem with the transformer, eh? can't help you diagnose it but i recognize good news when i see it.

 

[aborgman]

beats having a problem with the transformer, eh? can't help you diagnose it but i recognize good news when i see it.

I'm still not sure it isn't a problem with the transformer.

It may just be a problem that falls somewhere between the 8 ohm center tap on the transformer and the 16 ohm end tap, and thus only effects the 16 ohm output.

 

[JimmyM]

guess i can't recognize good news when i see it

 

[beans-on-toast]

It'll hit ~52 watts before it starts flat topping the wave. No sign of the goofy distortion/oscillation on the lower half of the waveform.

So there is definitely SOME sort of issue involving the NFB loop.

This unit does not have the original speaker cable, and is wired so the speaker jack is connected to the 8 ohm tap.

It's wired as follows:
OPT yellow (8 ohm tap) -> speaker jack +
OPT black (common) -> speaker jack -, PI gnd (for NFB loop reference)
OPT green (16 ohm tap) -> 22K resistor -> 7199 pentode stage cathode

Your wiring looks correct. You mentioned that you installed an isolated speaker out jack. I would connect the speaker return (-) and the OT secondary ground at a spare eyelet, then a short wire to the PI at the R13 1.5K ground point.

One thing that strikes me is that they have several versions of the PI for this amp and others like the B25B or B22 that use the 7199 tube. I would look at all the schematics you can find that have this PI tube and see what the differences are. No PI is perfect and some oscillate at higher power levels. pF caps are usually there to tame oscillations. Since they were evolving the designs, maybe there was an inherent problem that they were trying to address. Add a filter on the feedback resistor by paralleling a pF range cap with the 22K resistor. You can also try changing the value of the feedback resistor.

You mentioned swapping the OT primary wires to the plates. If they are connected backwards you would hear a howl or oscillation at high volume levels. Never hurts to try it. You can swap the PI plate connections or the OT primaries.

 

[aborgman]

Your wiring looks correct. You mentioned that you installed an isolated speaker out jack. I would connect the speaker return (-) and the OT secondary ground at a spare eyelet, then a short wire to the PI at the R13 1.5K ground point.

That is basically what I have.

OT secondary ground is attached at speaker jack negative.

OT secondary 8 ohm tap is attached at speaker jack positive.

Wire jumper from speaker jack negative -> eyelet where R13 is grounded.

One thing that strikes me is that they have several versions of the PI for this amp and others like the B25B or B22 that use the 7199 tube. I would look at all the schematics you can find that have this PI tube and see what the differences are. No PI is perfect and some oscillate at higher power levels. pF caps are usually there to tame oscillations. Since they were evolving the designs, maybe there was an inherent problem that they were trying to address. Add a filter on the feedback resistor by paralleling a pF range cap with the 22K resistor. You can also try changing the value of the feedback resistor.

You mentioned swapping the OT primary wires to the plates. If they are connected backwards you would hear a howl or oscillation at high volume levels. Never hurts to try it. You can swap the PI plate connections or the OT primaries.

All good thoughts I considered last night.

I'm going to try:

1) Swapping plate/cathode connection on PI (AKA swapping plates on power tube) with feedback attached - in case it was wired out of phase previously.

2) Taking feedback off 8ohm tap instead of 16ohm (potential OPT issue downstream of 8ohm tap?)

3) throwing a filter network on the NFB loop

 

[aborgman]

1) Swapping plate/cathode connection on PI (AKA swapping plates on power tube) with feedback attached - in case it was wired out of phase previously.

Result:

Original wiring correct, swapping power tube plates created the expected oscillator.

2) Taking feedback off 8ohm tap instead of 16ohm (potential OPT issue downstream of 8ohm tap?)

Result:

Sized new feedback resistor to give equivalent gain when driven off 8ohm instead of 16ohm tap.

Works perfectly. Runs right up to 50 watts with no issue at all.

3) throwing a filter network on the NFB loop

Result:

Nothing. Throwing a cap to ground creating a LP filter with a 20KHz -3dB point had no effect.


So it looks like I can get away with just using the 8 ohm tap to drive the feedback loop, and it looks like their probably is an issue with the OPT between the 8R and 16R taps.

 

[BassmanPaul]

I had a feeling that your problem might lie in the NFB loop once the OPT seemed to be OK. I was unaware that the speaker wiring had been messed with. You still have a problem but if the 8Ω workaround works for you i'd leave it as is.

One last question: if you take the speaker out from the 16Ω tap with the NFB on the 8Ω tap what happens?? There is something tickling the back of my mind about Loops connected outside the speaker connections. Damned if I can remember precisely what it was though.