Do guitar amps tend to self-oscillate?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Not sure I described this clearly enough or not. My scope, like most, ties the probe ground to safety ground at some point. My signal generator does the same thing. In the DUT we can call some point ground, say the junction of the filter caps. It may or may not be connected to safety ground, but will certainly have an AC connection to that point via a "ground breaker" circuit.

The problem is that the wiring to the speaker ground has some finite resistance. When the speaker is driven, a voltage will be developed at every point on that wire, referenced to the amp star ground. That voltage will also be sourced at a very low impedance.

When I hook scope probe ground to the speaker terminals, that voltage gets applied to safety ground, and ultimately modulated the signal generator slightly- even safety ground isn't zero ohms!

If the phase is additive, at some level of drive the system will go into oscillation due to positive feedback.

My usual cure is to float the signal generator, but what I didn't catch was that I often have it connected to a frequency counter that isn't floating. The ground path is thus complete and all heck breaks loose!
 
Member
Joined 2010
Paid Member
.......Note that the wire soldered across the backs of the pots in Marshall amps is NOT a ground bus. It is just a convenient place to tie wires and makes it inconvenient to replace a pot.....
I think this is the traditional and often effective means of shielding pot. elements from EMI. It's a particular problem with large pots in tube amps but all high impedance circuits are vulnerable. The chassis earth and metal construction of pots doesn't provide reliable shield grounding, so this wiring was once universal practice in quality audio control circuits.

Miniaturisation and PCB techniques reduce this necessity but it's still good practice in old, hard wired circuits. Even now, I see ALPS pots wrapped with foil and grounded to reduce noise in hi-fi gear and even the whole input section of guitar amps shielded top and bottom with tinplate.
 
soundguruman is correct. Its more and more common the higher the gain in the amp. I just got a Peavey classic 50, and it's very versatile. But if you're dumb enough to turn up all the volume knobs on the 'dirty' setting (not really channel-switcing, it switches in 2 more tube stages) and turn up the 'presence' it will make a high squeal.

A big clue is whether the input device has any effect on the oscillation, like does it change if you turn the guitar's tone control or unplug the input or short it? That might indicate the problem is indeed in the first stage(s).

On the other hand, if minor changes to the 'presence' control or master volume makes the squeal go away (not jsut change volume but stop oscillating), the problem is in later stages (what stages does your feedback loop involve?). Less often the output stage itself can be unstable with some settings of a 'presence' control. On some, 'presence' could be re-labelled 'oscillate'...

It's pretty creepy, and can be hard on the amp and speaker voice coils, but its not impossible to eliminate or live with. Many people consider the possibility of occasional oscillation during setup preferable to limiting the range of the controls. You may want a ton of gain and treble for some stages, but just never on all at once. For instance, I like to boost treble in an early stage, then generate a bit of distortion, then cut the treble in a late stage, so that the guitar comes thru about the same after treble was boosted then cut, but the distortion only has its most irritating treble cut. Some even use bigger-diameter speakers for that final treble-cut after boosting the heck out of the guitar treble via a bright switch or bright input in a ver early stage. For the opposite case, you might want a clean sound to some thru unaltered with no toen alteration until the presence control. So all the wild ranges of each control is useful in some condition, but there are some combinations of settings that just don't make a good circuit together! A gelding plowhorse is easy to control, a thoroughbred is more finicky and requires some finesse. It's probably possible (and a good idea) to minimize or eliminate the squeals and keep the versatility, but amp mfgrs can't safeguard against things as simple as tube manufacturer variations, so they make them robust enough to survive an ocassional bit of oscillation. It's usually so high that it's already getting damped by the band-pass limits of the amp, so it's not usually putting the full power of the amp into the voice coils either. When my Peavey 'rescue amp' squeals, its not really all that loud and not really all that much power; it certainly COULD be worse. So you'll have to judge for yourself whether it's really out of control enough to be dangerous to the amp or speakers. Then again, if the PA picks it up and the mixer has add a lot of EQ it can really be unpleasant for the horn diaphragms and audience. So you decide whether to address it.
 
OK all that said I am getting kind of tired of putting up with the squeals on my 'rescue amp' though they're not very loud...neither is a harmless burp, just unwanted, mildly bothersome, and crude in certain company or settings. On my bluesy amp, no big deal, it's commnon on bluesy amps, but not very professional. Drunken sex with a toothless waitress is bluesy and usually harmless, but not something to be proud of. So I'm going to take the advice somebody just gave:
Installing a 7 pf, 1000V cap, between plate and grid (between pin 2 and 1 OR pin 7 and 6 on a 12AX7) is a very widely used snub circuit....The cap should be a high grade silver mica, soldered directly to the pins of the socket.

On mine, the problem configurations involve multiple stages, the treble-boost being cumulative. Doesn't seem to invovle the very first stage though. But if I choose a high knee at the top of my pass band, its not giong to affect the range I want to reproduce. So first I'm going to try something really small like 3p on the 2n and 3rd (which are inserted by relay in the "dirty" mode) and 4th stage.
 
Last edited:
perhaps the plate to grid is a phase cancellation of the passed frequencies? I've got an old projector amp that uses a cap from plate to plate across the 6v6s in such a way, and in the preamp across the plates of a 12ax7. since you have many options, if its easy to get around the amp and you have the time, you could try a variety of things, though best would probably be to determine where the squeal starts and nix it there.
 
The tube changes phase by 90 deg, from grid to plate. That's not 180. I still dont' get it.

It definitely involves the 2nd, 3rd, and 4th stages as well as the feedback in the output stage. I can stop the squeal by turning down the gain between stage 1 & stage 2, or by turning down the gain between stage 3 & 4 (there is no control between stage 2 & 3) or by turnign down the master volume or turning down the presence control.
 
OK I think maybe I understand. Placing a capacitor between the anode and the grid applies negative feedback of very high frequencies to the grid, and also appears in parallel with the miller capcacitance, so its value is multiplied by the gain of the stage, due to the miller effect. So you get away with a smaller capacitor than you'd need from grid to ground or grid to cathode.
 
Hi Guys

Adding a plate-to-grid cap is distinctly a feedback loop. This works against the grid circuit resistance and rolls off high frequency response, which in turn reduces the propensity to oscillate.

The same effect is attained using a cap across the plate resistor, or one from plate to ground before or after the coupling cap. If you want a solution that is not feedback, then use one of these methods.

The other standard method to eliminate oscillation is to add or increase grid-stop resistance. Most tube amps lack grid-stops as that is the "engineering ideal" - and also makes the schematic simpler and the supposed cost lower. Especially in a guitar amp, grid-stops are a necessity as they prevent grid rectification of the signal, as TUT5 clearly illustrates. For any technology, a grid-stop, gate-stop or base-stop works against the internal capacitance of the active element to create a HF rolloff. The stop resistor must be placed as close to the control pin as possible.

Have fun
Kevin O'Connor
londonpower.com
 
Hi Guys

It's a lot cheaper than going to college and acquiring decades of experience. The practical experience embodied in the TUT-series books keeps many hobbyists and builders from wasting a lot of time or money reinventing the wheel.

Besides, you can get TUTs for free when you ask others to give them to you as birthday or Christmas presents.

Have fun
Kevin O'Connor
londonpower.com
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.