OK, I have a weird situation. I am modifying a guitar preamp. The preamp is a module with 4 tube stages: 3 gain, 1 cathode follower, and there is also a common stage (not on module) to make 4 gain stages in total. I've connected a gain boost on the stage 1 cathode by adding another resistor, along with a bypass cap in parallel, to the original cathode resistor.

I've installed toggles to bypass any number of stages, at a time. I've successfully done this before, many times, but something is different here and I can't figure it out.

Here's how the toggles work:
clean channel-bypass stages 2 and 3.
rhythm channel-bypass stage 3.
lead channel-nothing bypassed.

Here's the problem. When I engage the gain boost, the rhythm channel gets a nasty squeal and harsh distortion noise. The squeal will change frequencies when I adjust the gain up/down and it will also go away when I turn the mid range pot down far enough. The squeal can also be heard on the clean channel, but not as badly. The lead channel doesn't have any squeal.

I've already tried 6 different tubes, so no microphonic tubes. I've also changed the cathode resistor on stage 3 and plate resistor of 4 (and 5). I've rewired it with new wire, and I've resoldered all ground connections.

It seems like a ground problem to me and maybe a bad component that can't handle the gain boost. But, I hate shotgunning individual components, so I was hoping for some insight from here.

Thanks a lot!

 

The rhythm channel is the only setting that does not invert the signal.
Feedback is always about the gain being too high, and with vacuum tubes, the impedances are high so signal can leak from the wiring to the tube inputs.

Try using coaxial cable to do the wiring and be willing to lower the gain of some stages if that isn't enough. You can also put "grid stoppers" in. That's a resistor connected to the grid pin and the coax connected to the resistor. I used some 100k ohm 1/8th watt resistors soldered directly to the end of the coax and held in place with shrink tubing.

There is some stuff to think about.
It can be maddening. Thousands of tube amps have been scrapped for layout problems.

 

Your modification is causing a feedback oscillation to occur. This is probably due to a filter element(cap)somewhere on the output side of the tube not functioning. if this amp is old try replacing a few of them with new caps. That should stop the oscillations. Also try a cap with 10 or 100 times more capacitance in your mod circuit and see if the symptoms change. this would verify the oscillation diagnosis.

 

The fact that a given modification has been successful in other applications does not guarantee that it will always work. Possibly this amplifier may not tolerate much increase in gain because of existing feedback paths. In addition, the physical arrangement may be different, so that the parasitic effects produced by the added wiring are more serious.

The previous poster's suggestion of a dead filter capacitor causing power supply decoupling issues may be the root of the matter, but there are other possibilities such as stray capacitive couplings or ground loops. That said, since you say that there is no squeal until you bypass some stages, it sounds just as likely that stray couplings from the bypass wiring may be the real problem.

You might want to try the amplifier without that extra wiring to see whether the gain boost is a problem in itself. If you then obtain stable operation, you can set about trying to improve the layout, perhaps using coax. as somebody else suggested.

 

Thanks a lot for the responses! I'll have a chance to work on it again, today. I'll post back afterward. One thing though. I doubt coax is going to work, but only because the work space is so limited and trying to make short cuts and turns would probably prove to be a huge PITA.

Thanks again!

 

For this purpose you would expect to use very thin coax or screened audio wire, obviously not something like ethernet cable or antenna downlead!

If your layout is so physically constrained, then maybe you need to rethink it, or even accept that you cannot have all these switched options.

 

For this purpose you would expect to use very thin coax or screened audio wire, obviously not something like ethernet cable or antenna downlead!

I figured as much, but thanks for clarifying. So, a simple shielded wire like a RCA cable could work, or no?

If your layout is so physically constrained, then maybe you need to rethink it, or even accept that you cannot have all these switched options.

This sounds like an invitation to push limits even further...LOL. You're probably right though. I have to build the switches first with all of the components/wiring and then install them, rather than mount a switch and build to it.

 

yes. The wire out of RCA cables will work but...RCA cables are not the thinnest coax you can buy.

The last tube project I did was a pair of 12AX7's to a 6V6 and I did it all in a 2 inch by 2 inch square tube, about a foot long. I'm talking "ship in a bottle" here. It can be done with the correct forceps, but I'm answering you because of the feedback problems. Main problem was that I put my switching at the end of the tube so it would be easy to get at. I should have put each stage switcher between the tubes. It was all that wire to the switching board that radiated signal all over the place.

 

Hey guys, I found the problem. It was a lack of bright cap across the gain pot. I have a 3-way bright switch installed, but it wasn't until tonight that I actually attached the caps to the switch. The "normal" position on the switch had nothing attached and there is where the feedback and crap was coming from. When I switched to the 470p or 4700p caps, the feedback was gone.

Another thing too is, the feedback was only there when the gain pot was between 6 and 8 on the dial. Any lower or higher and the feedback disappeared.

Thanks everyone for the replies. As kermit pointed out about using a cap 10 or 100 times larger, well, that proved to be the solution.

 

Gee, I would have never guessed you didn't finish assembling it before asking why it didn't work.
I learn somethng almost every day on this site.

 

Gee, I would have never guessed you didn't finish assembling it before asking why it didn't work.
I learn somethng almost every day on this site.

That's a reasonable assessment, but I typically test my work as I complete it, rather than build to completion and then test.

The case with the bright cap(s), my idea was to mount them on the bottom of the pcb, and I had forgotten they weren't installed yet since they're not always in view..