Op-Amp oscillations--what causes it?

[John in WI]

Hoping to learn something to trouble shoot a distortion pedal I built last weekend.

The amplifier was built along the lines of a DOD250, with 741 op amps.

The problem I am having is that at certain gain and diode settings, I get a high pitch oscillation in the circuit. Oddly, it gets a lot worse when I crank my tone knob to high pitch (on my guitar). It seems to go away when I'm playing, and get worse with the strings muted. If I turn down the gain, there is a sharp break where it goes 100% away.

how do I trouble shoot this problem? I guess the best bet is to make the oscillations stop, but short of that, can I place a low-pass filter just before the output, to at least filter out the squeal? (I just don't want to fry chips ever couple hours).

Thanks for any thoughts on it. I'm very new to op-amp devices and trying to learn.

 

[ICTRock]

First, tidy up all leads and make sure they're as short as possible. I'm old school, I like shielded input and output leads too. This helps rule out crosstalk and parasitic oscillation.

 

[dsutton24]

Generally unwanted oscillation happens when you are trying to get too much gain out of a stage, or when somehow some of your output signal gets mixed in phase with your input, or finds its way into the power supply.

Trying to filter the oscillation will only result in distortion.

 

[FenderLover]

The 741 is an ancient design, the first op amp. It requires an offset null (pins 1 and 6??) so it is not internally compensated. It is also probably one of the noisiest op amps out there. I understand why you're trying the 'original chip' but I'd use an internally compensated op amp, like the TL071 instead. Radio Shack carries them.

 

[ICTRock]

The 741 was the first internally compensated op amp

 

[Gnobuddy]

Yup, fully internally compensated, and also old, slow, and noisy.

To the OP - ICTRock is pointing you in the right direction. There is a lot of gain in an op amp, even such an old one. Poor layout or wiring can make it squeal. Treble boost can make it squeal. Building it on veroboard or strip-board can make it squeal. Not using shielded wires for input and output can make it squeal.

If it's any consolation, the old 741 is usually one of the easier-to-tame op amps. It's much slower than newer op amp designs, and the slower speed (technically: gain bandwidth product) let's you get away with some things that will provoke a newer and faster model into oscillations.

-Gnobuddy

 

[Montana_Dawg]

To prevent oscillation, put a bypass capacitor from the power pin to the ground pin. A 0.1uF should be sufficient.

 

[John in WI]

Thanks for all the information. I used the 741 because it was available at the local Radio Shack. Might be time to order a better single (or maybe dual) op amps.

I did build the circuit on protoboard, but was careful to run the wires straight, and tight to the board. Nothing crossing anyway. I put a ground plane around the circuit and under the IC socket to hopefully soak up RF noise. I will try to replace the in and out with shielded wire. I clipped all the wires as short as possible so there isn't a bunch of wire hanging out the back of the board.

The +9V is connected through a large and small cap to ground, so I do not believe it is weird oscillations in the voltage input.

It might just be a matter of turning down the gain, and if more dirt is required, maybe installing an additional amplifier stage to get it.

Ideally, if if I can get this tweaked and working perfectly, I would like to re-do the project with a PCB which would certainly help the situation

Thanks for the thoughts and suggestions. I will toy around with it when i get a minute. It's close--I'd just like a little bit more dirt before the squealing starts.

 

[Montana_Dawg]

Did you try the bypass cap? Parasiitic oscillations are a common problem with amplifiers and even other circuits. Most data sheets have a typical application circuit, and they almost always have the power input pin tied to ground through a small cap. Any oscillation that occurs is then bypassed to ground.
Some transistor configurations have a bypass cap in parallel with the emitter resistor to ground for the same purpose.

If youndo place the cap on your board, be sure to place it as close the Vcc pin as you can get. The further away it is, the less effective it is.

 

[FenderLover]

The 741 was the first internally compensated op amp

Oops...Thanks guys. I can't remember the last time I've used a 741 because the TL071 is available. I use mostly duals because there are so many other drop-ins that match pinouts, and even more circuits (that I build) that require duals anyway.

 

[Gnobuddy]

The +9V is connected through a large and small cap to ground, so I do not believe it is weird oscillations in the voltage input.

As Montana_Dawg is hinting at, that small filter cap should be mounted as close to the op-amp power pins as you can manage. Even a few inches of copper track between the cap and the op-amp can have enough inductance to cause high-frequency RF oscillation with some op-amps.

I think with strip-board the problem is the parasitic capacitance between those long, closely-spaced, parallel copper tracks. That capacitance can couple all sorts of signal points together at high frequencies, and cause oscillation that way.

I'm still surprised that you're having these issues with a 741. In my (admittedly long-ago) experience with these, they are extremely stable, and will tolerate sloppy construction that will immediately provoke misbehaviour in better quality (faster) op amps

Not that your construction is sloppy - from you description, it sounds as though you were paying attention to the layout.

Is there any intentional treble boost built into the feedback network? That is one guaranteed way to turn a stable fully-compensated op-amp into a high-frequency oscillator...

-Gnobuddy

 

[John in WI]

That's a good point. I did not intentionally boost the treble. I did, however, run the signal from the first op-amp stage (the full guitar signal) through a high-pass and low-pass filter, to seperate the two into 2 seperate components, and sent these into two separate op-amps.

I should check again and see--it could be that the low-pass system is working ok, but the high pass is oscillating like crazy, but I am not detecting it since they are tied back together through a blend pot.

I can disconnect the high-pass and low pass circuits, and investigate them separately. The first stage (a "booster" stage) seems to work fine, no matter how cranked the gain is. It's either that I'm driving the inputs too hard on the second stage, or I'm inducing some weird high-frequency oscillation.

It might be that hi-pass amp stage that's getting cranky with me.

 

[John in WI]

The treble problem was an excellent call!

I de-soldered the leads from the high pass and low pass filter, and tested them individually. If I run the lo-pass signal through either op amp stage, it's fine, cranked to full gain. The high-pass signal, through either op amp, gets screechy at maybe 2/3 of the way.

Still, I get good crunch at 2/3 without squealing. I think the simple solution here is to put a limiter on the gain control, so I can safely max out the pot without screaming.

I think this circuit is pretty slick now. I think I need a bigger value on the blend pot though. I think I'm currently using a 50k, and when I turn it all the way one way or the other, it flavors the tone but I guess I would prefer it if turning it maximum one way or the other turned off the other signal. Of course that would attenuate the total signal when it's in the middle. A little more tinkering is called for, I think. But the screeching problem seems to be solved.

Thanks for all the insights--the sheilded in and outputs helped the white-noise back ground, and an extra cap at the power supply never hurts--amp circuits can start pulling a lot of juice when you put them under load. WIthout an o-scope, I'm flying blind here. Literally, playing it by ear.