Reverb Circuit has now killed two reverb drivers. Please help!

[Gaz]

As others mentioned, the primary of the 022921 should have resistance in the 1.7K to 2K range.
You mentioned in the first post the new one had about 800 ohms. This makes me question the type and quality of the replacement transformers you are using.

They are both Fender replacements from Heyboer.

[Gaz]

Is that a typo on the second grid of the 12AU7 driver tube, where it's connected to the source side of R68 56k grid stopper resistor. Is it instead connected to the other grid directly? Just curious.

Yes it’s a mistake, the grid stopper is before both grids.

[66 Kicks]

As others mentioned, the primary of the 022921 should have resistance in the 1.7K to 2K range.
You mentioned in the first post the new one had about 800 ohms. This makes me question the type and quality of the replacement transformers you are using.

I remember those 022921's having a 2K DCR decades ago, but now-a-day P-TF22921's and Hammond 1750A's run out at 1K. These things are always + or - 20%, so a Heyboer with 0.8K is right in there.

[Helmholtz]

Yes, but I am fully bypassing the 820R as well, so could that be too much potentially?

The bypass cap increases gain but not available current and power. It does increase the risk of overdriving/clipping and this might cause high voltage spikes at the primary. So Nick's proposal with the MOV seems reasonable.

Another cause could be HF oscillation producing excessive voltages at the primary. Also in this case the MOV can protect the transformer, but the oscillation itself needs to be adressed with different means.

Did you consider to send in the defective transformers to the manufacturer for an analysis?

[Gaz]

The bypass cap increases gain but not available current and power. It does increase the risk of overdriving/clipping and this might cause high voltage spikes at the primary. So Nick's proposal with the MOV seems reasonable.

Another cause could be HF oscillation producing excessive voltages at the primary. Also in this case the MOV can protect the transformer, but the oscillation itself needs to be adressed with different means.

Did you consider to send in the defective transformers to the manufacturer for an analysis?

Yes, plan on sending them in just out of curiosity. Has to be from some spikes, right? One other odd thing I forgot to mention is that adding the .0022uF cap in the recovery side across the 220K to ground the reverb became brighter, not darker. I thought for sure the .0022uf would dump highs!

[Helmholtz]

One other odd thing I forgot to mention is that adding the .0022uF cap in the recovery side across the 220K to ground the reverb became brighter, not darker. I thought for sure the .0022uf would dump highs!

No, we had this in another thread here before. The cap would reduce highs, if the signal source was resistive. But as the reverb recovery transducer is inductive, you get an LC low-pass with a resonant peak that actually boosts high mids.

[Gaz]

No, we had this in another thread here before. The cap would reduce highs, if the signal source was resistive. But as the reverb recovery transducer is inductive, you get a parallel LC resonant bandpass that actually boosts high mids.

Oh thanks for the explanation! I couldn’t find any info online (except one site that erroneously states that it cuts highs). I found this after trying everything to brighten up the shorter 3 spring tank I’m using after comparing to a SF reverb I had with the same tank). I was surprised it brightened it up but it works great with the tone control, giving it more range and getting the 3-spring tank closer to the zing of a common 2-spring.

Now if I could just puzzle out my “primary concern”

[Gaz]

Ok, got a replacement driver (different brand, and measures 1K DCR on the primary btw). I’m gonna reduce the drive and put an MOV across the primary, but wanna check for oscillation. What should look for scoping the primary? What should a healthy wave form look like with the tank attached? Should I use a 1K resistive load to test?

[nickb]

Ok, got a replacement driver (different brand, and measures 1K DCR on the primary btw). I’m gonna reduce the drive and put an MOV across the primary, but wanna check for oscillation. What should look for scoping the primary? What should a healthy wave form look like with the tank attached? Should I use a 1K resistive load to test?

Use an 8 ohm resistor load - the same as the tank. Feed the amp with 100mV 1Khz from a signal generator of your choice. Probe the secondary and look for something that is very close to a sine wave. Any sign of flat tops or large spike is not good. You never want to drive a tank so hard that it clips as, IMHO, it doesn't sound nice and is likely the primary reason for big spikes that kill your transformer. Never, ever, run it without a dummy load or tank connected, just like the main output.

[Jazz P Bass]

From the Deluxe Rvb schematic:
60 Vac signal on the primary should induce a 900mv signal on the secondary. (with the tank connected)

[Helmholtz]

Use an 8 ohm resistor load ..

Substituting the transducer by a resistor changes the system behaviour. Saturation and oscillation are much more probable with the inductive transducer load. As the impedance of the transducer increases with frequency, also transformer primary voltage and gain do. Not to mention the phase shift involved.

[Gaz]

Use an 8 ohm resistor load - the same as the tank. Feed the amp with 100mV 1Khz from a signal generator of your choice. Probe the secondary and look for something that is very close to a sine wave. Any sign of flat tops or large spike is not good. You never want to drive a tank so hard that it clips as, IMHO, it doesn't sound nice and is likely the primary reason for big spikes that kill your transformer. Never, ever, run it without a dummy load or tank connected, just like the main output.

Thanks, Nick. I did this and here's the scope shots. First three are at power amp clipping, the gain half way, then dimed. You can see there is a spike when overdriven. This amp gets run this way a lot too because there is a PPIMV.





Next shot are the same, but with a 1nF cap across the primary, and it take care of the gain spike for the most part.





I'm not sure if this takes care of the problem, and what voltage cap would be most appropriate here (Fender used a 560pF/3kV volt cap). Also, I forgot to scope before and after the transformer to see if the driver itself is actually clipping or just the preceding stages...

(And apologies the shots are sideways. Grrr!)

[nickb]

You need to measure on the primary of the Reverb driver transformer using a x100 probe. The secondary doesn't tell the whole story. I see nothing in those scope shots that concern me. We're looking for HUGE transformer destroying size spike. Since these sorts of drivers have been around for donkey's, my money would be on a bad connection on the secondary. All I see is a little ringing.

Also HH made a very important point. You do need to test into the actual tank to really see what is happening. The resistive load is a first step. It's a way of taming things for an initial look.

[Gaz]

From the Deluxe Rvb schematic:
60 Vac signal on the primary should induce a 900mv signal on the secondary. (with the tank connected)

Thanks

[Gaz]

You need to measure on the primary of the Reverb driver transformer using a x100 probe. The secondary doesn't tell the whole story. I see nothing in those scope shots that concern me. We're looking for HUGE transformer destroying size spike. Since these sorts of drivers have been around for donkey's, my money would be on a bad connection on the secondary. All I see is a little ringing.

Also HH made a very important point. You do need to test into the actual tank to really see what is happening. The resistive load is a first step. It's a way of taming things for an initial look.

Ok here are the shots at different gain levels on the primary with the tank attached using a 100x probe. This is with 820R unbypassed cathode resistor (B+ is only about 350vdc as well), and 1n/3kV cap from plate to cathode.

Gain at power amp clipping:



Gain half way:



Gain maxed:




I don't have anything to compare to, but does it look okay? It seems to me it's "clean" and nothing off the charts. The wave shape changes smoothly as the gain is turned up if that makes sense.