Fender Reverb, splitting 12AT7 Driver Question

[Thisismyname]

So I have a single channel, no fx AB673 Deluxe built into a Deluxe Reverb chassis. I recently added the reverb circuit back in. It now has a super high frequency oscillation that I can't exactly hear, but it makes my ears ring when it's on. It showed up some time after the reverb circuit was finished and being used. I tweaked the preamp since then, so I probably did something the amp didn't like. I have troubleshooting and potential remedies I'm doing on my own before I consult the forum. It's also a good occasion to finally learn how to use my oscilloscope.

However, in my research, I came across a thread about Fender reverb oscillation where a member said the 12at7 as a reverb driver is a bad design choice due to unbalanced triodes and can induce oscillation because of how the grids and cathode are in parallel. He recommend splitting the cathodes and grids and adding grid resistors. The plates are the only section that can be in parallel. I understand splitting the cathodes, and the plates in parallel, but I'm unsure about the grids.

This is the only implementation I can think of is splitting the signal after the 500pF highpass where half goes through a grid resistor into 1 triode and the other half goes through it's own grid resistor and then goes into the other triode. So the grids are still in parallel, they just have their own grid resistors. Is this right?

I am wondering if this is the correct way to wire it and if the poster just wasn't as specific as he needed to be when he said only the plates can be in parallel.

[dogburn]

I built a single channel AB763 Deluxe Reverb (no tremolo) derived from Rob Robinette's Blackvibe Reverb, and did spend some time trying to make the reverb behave. One thing I tried was putting in separate resistors and caps for the cathodes, but that didn't make a big difference. Otherwise, having it all in parallel is pretty much the standard Fender design. Did you try other AT7 tubes in that socket? And have you tried to measure the frequency of that high pitched sound?

Otherwise, some photos (and schematic) would likely help people troubleshoot.

[Latole]

I have few amps with 12AT7 as reverb driver and repair many customers amps with this reverb ; never see a issue.
Amp built ; more chance is a wires / resistors dress issue

[Merlin]

This is the only implementation I can think of is splitting the signal after the 500pF highpass where half goes through a grid resistor into 1 triode and the other half goes through it's own grid resistor and then goes into the other triode. So the grids are still in parallel, they just have their own grid resistors. Is this right?

That's right, although I think it is unlikely that is causing your issue. But it won't hurt to try it, it's only a couple of resistors (a few hundred ohms to a couple of k ohms).

[tubeswell]

I came across a thread about Fender reverb oscillation

Got a link?

[Thisismyname]

If I can't figure it out on my own, I'll start a new post with more details. I'm not out of ideas yet. Thanks for the advice so far!! So I must be correct that the grids are still in parallel, they just have their own grid stopper.

https://ampgarage.com/forum/viewtopic.php?t=36771

Bepone is who I'm referencing.

[HotBluePlates]

... in my research, I came across a thread about Fender reverb oscillation where a member said the 12at7 as a reverb driver is a bad design choice due to unbalanced triodes and can induce oscillation because of how the grids and cathode are in parallel. ... The plates are the only section that can be in parallel. ...

Fender has been doing it for 62 years now.  I wonder if they've topped 1 million amps built this way?

Point being, there is a massive number of amps out there with 12AT7 sections in parallel without any apparent issue.

https://ampgarage.com/forum/viewtopic.php?t=36771

Bepone is who I'm referencing.

Ugghhh...  Sometimes the loudest voices are not the ones you should listen to.

Instead, take Martin Manning's advice on that thread.  Everything I've seen from him is spot-on.  I notice he had a tip for wiring the 12AT7 exactly as Fender's layout showed (connect grids & cathodes directly over the top of the socket, but have a wider looping connection between plates that lays on the chassis).  See the photo at bottom; 12AT7 is the 3rd tube from the left.

... I recently added the reverb circuit back in. It now has a super high frequency oscillation that I can't exactly hear, but it makes my ears ring when it's on. ... It's also a good occasion to finally learn how to use my oscilloscope.

Use that scope to determine if/where the oscillation is happening.  Make sure you have the amp Volume all the way off when you do this test, so you're not distracted by background noise.

You may also need to show some photos of how you added the reverb back into the amp-circuit.  If you're having problems, it could be due to the implementation not-matching how Fender laid things out originally.

[Lectroid]

Instead, take Martin Manning's advice ... he had a tip for wiring the 12AT7 exactly as Fender's layout showed (connect grids & cathodes directly over the top of the socket, but have a wider looping connection between plates that lays on the chassis).  See the photo at bottom; 12AT7 is the 3rd tube from the left.

@HotBluePlates,
That's a great observation.  Might be elementary to you, but a great observation to those of us still finding our way.  I've looked at Blackface amps' pictures at least dozens of times without noticing that before.  Fender engineers were focused on noise-reduction even back then.

So now (just for a double-check) I looked at the '64 Deluxe Reverb gut shot on Sluckey's site.  Yes, it shows the same wide looping wire connecting the reverb driver tube's plates, and pressed along the chassis floor.  They were serious.

I've read here many times that good lead dress is essential and I believe it.  But this kind of layout planning seems next-level to me.  I think I need to really study Sluckey's DR shots, there must be other secrets of the layout masters hidden in plain sight.