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.
I've got a solid state bass amp that, when you boost the treble, will self oscillate if the gain is high. The severity depends on what impedance is connected to the Hi Z input, but would this be common or acceptable? I'd think a manufacturer would be very reluctant to put out a product where some control combination could cause self-oscillation. OTOH, with the amount of boost available, it might be hard to prevent. I know a lot more about hi-fi amps than instrument amps, so enlighten me!
 
Ultimately you are dealing with a manufactured product. Things happen in the manufacturing process far from the eyes of the people that designed it. Parts get substituted where price is the most important specification. Checkout procedures are written so people without technical training can do functional checkout of finished products. The people manufacturing amps may not understand what it is or how it is used. This happends frequently when manufacturing is moved offshore and the parent company receives amplifiers boxed for shipping on pallets. Sometimes they are shipped straignt to distributors. I'm surprised more don't burst into flames when you plug them in.
 
I've got a solid state bass amp that, when you boost the treble, will self oscillate if the gain is high.

I have an old 80s solid state Marshall (type 51-something) guitar amp that also peeeeps when you turn the gain fully open.
Later versions of this amp had a small ceramic cap (few pF) somewhere in the preamp to cancel this. So yes, it sometimes happens.

YMMV
 
I had a Sunn Coliseum Slave that did that.
It did to the owner and only once for me.
So, I cleaned it and Sprayed the controls out and reseated all of the connectors to the board.
As far as I know it has never done it again.
I don't know exactly what the cause was but it never came back.

I also had a Peavey Renown 400 the destroyed a speaker because of ultrasonic oscillations.
The cause of that was a bad Molex connector to the board.
So, I cut the connector off and soldered the wires directly to the board and it never did it again.


FWIW

jer :)
 
Ah ha, so it does happen. This is an older amp and it oscillates with a sig gen on the test bench, so acoustic feedback is highly unlikely. I've got a message in to the manufacturer to see what's normal. The amount of gain is tremendous when you boost the treble, and there are three ways to do that- treble control, "bright" button and a 5-band equalizer. I'm not above adding some non-factory shielding if it would help. Thanks!
 
It Is hard to tell without seeing it or a schematic, But my guess would be maybe a cracked trace or bad solder joint somewhere.
If it as a Graphic eq or any eg for that matter if the negative feed back path is open this could cause to much gain and cause it to oscillate.

I had described a little about this here,

http://www.diyaudio.com/forums/instruments-amps/215066-little-help-here.html#post3071431

http://www.diyaudio.com/forums/instruments-amps/215066-little-help-here.html#post3070996

Check the control to see if maybe the element has worn out or had opened on one end.
As this could cause it to be separated from the negative feedback path making it only have a positive amount of gain.


jer :)

P.S. In some eq circuits there there is a positive feedback return path and a negative feedback return path unlike the link above.
 
Last edited:
I've got a solid state bass amp that, when you boost the treble, will self oscillate if the gain is high. The severity depends on what impedance is connected to the Hi Z input, but would this be common or acceptable? I'd think a manufacturer would be very reluctant to put out a product where some control combination could cause self-oscillation. OTOH, with the amount of boost available, it might be hard to prevent. I know a lot more about hi-fi amps than instrument amps, so enlighten me!
It is normal and quite common to have squealing oscillations in a tube guitar amplifier. These types of oscillations most commonly originate in the first preamp stage.
Sometimes this is cured by selecting the proper grade tube for V1A, selected for low microphonics.
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.
This design was used by both Fender and Marshall to prevent squealing oscillations.
That's what you were looking for.
The cap should be a high grade silver mica, soldered directly to the pins of the socket. Production amps use a ceramic disk capacitor.
This is ONE of the techniques used by manufacturers to stop the ringing.
 
Ex-Moderator
Joined 2011
It is normal and quite common to have squealing oscillations in a tube guitar amplifier. These types of oscillations most commonly originate in the first preamp stage.
Sometimes this is cured by selecting the proper grade tube for V1A, selected for low microphonics.
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.
This design was used by both Fender and Marshall to prevent squealing oscillations.
That's what you were looking for.
The cap should be a high grade silver mica, soldered directly to the pins of the socket. Production amps use a ceramic disk capacitor.
This is ONE of the techniques used by manufacturers to stop the ringing.

Why do the caps have to be rated for 1000V? To prevent HV shock in case of a capacitor short? But as most preamps have relatively "low" voltages on the plate, even if we double the rating, it is still not close to 1000V, unless for some un-common designs. I know many amp manufacturers use them, but never understood why, curious that's all...:confused:

Jaz
 
In addition, how close does the value need to be to 7pf? Will 10pf be OK?

I'm experiencing the same issue with a 5W "pre" amp (Champ with master volume), when the gain and tone are pushed to the top of their swing. What aspect of the preamp tube or wiring related to it or the tone control causes this to happen?
 
I can't speak to the 1000v part, though I've seen it used several places, including between plates in the output section, to snub or roll off high frequencies.

The amount of capacitance is a variable - ideally enough to pass the high oscillating frequencies and yet leave the desired lower frequencies untouched. This can be effective in tone shaping in an otherwise "brittle" amp well. In Herzogs, the basic idea is applied in a couple of places to roll off upper harmonics so that 2nd harmonics can be emphasized.

In otherwords, try the 10pf and see if it works for you! or series 2 together for 5pf total capacitance at twice the voltage rating
 
Thank you blue, I'm just waiting on some parts and I'll be playing around with different values. My project is indeed based off the Herzog.
Does tube microphonics come into play here? (does it have anything to do with self-oscillation) I've found that the no matter which 6V6 I use I still get mechanical noise through the circuit.
 
i've read that all tubes are microphonic by design, but obviously not generally to the degree of what people call "microphonic". But a Herzog can be pretty high gain and thereby susceptible to microphonics simply because of how much it amplifies the signal - so a little microphonic noise in the first gain stage gets amplified again and again . . . Depending on what you are building, you might be able to move the tube that causes problems further down the line to reduce the microphonic effect. This same tube may be the source of your squealing, though, and swapping it with another may be what you need. Still, I wouldn't necessarily toss it out.

I built a Herzog with a 12ax7 and a 6AQ5. It has a very slight mechanical noise when tapped, but does not self-oscillate even without the roll-off caps. I put in a three way switch with two different cap values (third setting Off) so that I could have roll-off options. It helped get the Herzog out of the one-trick pony category.
 
Hi Guys

Guitar amps whether tube or solid-state or high-gain or low-gain should not oscillate with all the controls dimed. If a tube amp does this, part of the reason may be microphonic tubes and/or microphonic construction - both can be eliminated and remedied. If it is solid-state then there is a serious layout flaw on the board.

It is really no different than laying out hifi gear. The high gain stuff in solid-state is built around opamp circuits which often incorporate diodes and nonlinear elements in the feedback loop. These are all local to the opamp, but a poorly designed board can make any circuit problematic. Occasionally this is mixed with bad circuit design, as in the case of most Crate amps, especially the GX series. Their designer is a mental case according to the 800-number help line.

Then you have mesa telling you explicitly in their owner's manuals that setting the controls certain ways may cause oscillation and "this is normal". They leave out grid-stops almost everywhere and have a pretzel-shaped signal path, so stability is compromised wherever they can do it. Their simulclass amps will oscillate if you probe them with a scope. Maybe that is part of the IP defense system?

Microphonics can occur in solid-state assemblies, as well. Caps can be microphonic.

As far as cap ratings, many low-value caps are only produced in high-voltage forms, so you see 1kV parts in 50V circuits. The only caps to be wary of are the low-voltage tiny ceramics, which do not seem very reliable if they are operated with more than half their rated voltage.

Have fun
Kevin O'Connor
 
Having done a lot more of these, I now understand the problem. As is often the case, it was operator error. The usual test bench setup involves a signal generator and a scope, both with safety grounds that are often common to the signal ground. What happens is that a guitar amp has such high gain with the controls boosted that even the slightest modulation of ground on the scope will wiggle the input enough to cause oscillation.

The thing to be careful of is measuring anything across the speaker terminals. The minute you hook a scope ground to the speaker ground, it sees a voltage due to the non zero resistance of the ground line. That drives the signal generator via the safety ground and causes oscillation. We're talking extremely small voltages here, but at high frequencies and with the amount of gain available, it's enough. The cure is never to ground where current is flowing.

IMO, an amp that oscillates without being provoked in this manner has something wrong with it.

Regular hi-fi equipment never has enough gain for this to be a problem, though it might cause slightly high THD readings.
 
Last edited:
Hi Guys

It sounds like what you are seeing is AC hum in the room picked up by your own body and amplified through an open input. For the amp to be sensitive to grounds associated with input, AC or anything else suggests a fault in the basic grounding of the amp - usually a cold solder connection.

Modern ground-isolation places a low-value R like 50-100R in parallel with antiparallel diodes, all between the circuit ground and the safety ground. The R provides a low-frequency ground tie so the circuit is referenced. The diodes limit the ground fault voltage and must be high-current types. Usually a bridge is used with its DC pins linked and tied to one end of R and the AC pins linked and tied to the other end. About half the time you see a 100nF across the whole affair and this provides an RF short.

Have fun
Kevin O'Connor
londonpower.com
 
now that you mention grounding . . .

Hey Kevin! Funny you should mention the "ground breaker" circuit. I've seen it drawn two ways - one with the circuit ground going through the diode/cap/resistor network you described with NO chassis connection save for the Earth end of that network, and another where the circuit ground is connected to the chassis and a wire runs from that connection to the diode/cap/resistor circuit to ground. Which is correct? Or do both work? I've only used star ground or multi-star points in a bus ground.
 
Hi Guys

The chassis is only supposed to be connected to the AC safety ground. It conduct fault currents and provides electrostatic shielding. That's all.

Chassis ground should be tied to the audio ground via one point. This should NOT be a star ground.

The Galactic Ground we use in our amps and advocate in our books is a set of local stars on a common bus. The bus is tied to the chassis at only one point. Star grounds guarantee that signal currents will mingle that should otherwise be far apart, and so is a waste of time in all Fenders, Marshalls et al.

Note that the brass ground plane used in Fender amps is NOT a ground bus. It is just easier to solder to brass than to steel.

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.

The modern ground link is both a coupler and an isolator, and can be viewed as a two-terminal device unto itself. As I said above, one end ties to the audio ground and the other to the safety (chassis) ground.

It goes without saying that the mains cord is a 3-wire type and plugs into a properly-wired wall receptacle.The AC ground wire should be hard-connected to the chassis. otherwise it provides diminished safety.

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