how important are feedback capacitors to prevent oscillation in active crossover?

[mazurek]

I see many places where capacitors are inserted between the opamp output and input to prevent capacitive oscillation. I had bought linkwitzs crossover boards to learn from (decided I can't use for my application), and in several places, but not all, he uses feedback capacitors. Is there any easy (for mechanical engineer) rule about this to decide whether they are needed in the design before I make it?

I think he might use one every time there is just a resisitor in the feedback loop. Thanks for any help.

 

[I_Forgot]

For this sort of info, I suggest you pick up a book like "the IC Op-Amp Cookbook" or "Active Filter Cookbook" or even "The Art of Electronics". It is more difficult to read and understand the material yourself, but you will be getting the info from engineers who know what they are talking about. There aren't many of those who are able to give you accurate info here.

The value of capacitor used depends on the op-amp and the gain at which it is operated. The data sheet/ap notes from each manufacturer usually recommend the proper value to use.

I_F

 

[richie00boy]

It is also possible that the capacitor is used to perform some filter or integration function, not necessarily stability.

 

[mazurek]

Thanks for the feedback, I will check to see if my engineering library or public library has those books, and check the datasheet.

richie- I am reasonably sure it is not part of the filter, he has for example a 220pf capacitor in the feedback loop of gain and delay circuits.

 

[RetroAudio]

mazurek - I think you might be referring to a bandwidth limiting cap that's in parallel with the feedback resistor. Excuse my misunderstanding if this is not what you are talking of. the cap's value in conjunction with R forms a -3db point which rolls the BW off. The cap also forms a phase lead which is "supposed" to correct any lag found in the open loop characteristic. There's more to the story, but hopefully you get the drift.

I_Forgot is correct about getting the info readily from here by many,..you'll probably get more misinformation than anything. The thing to do is to pour over app notes and data sheets to get the lowdown on how opamps work, these can be very informative and even application specific. Granted,it can be daunting at first, but just keep cracking away and eventually you'll get it soaking through. It's not a real easy subject to understand which is why audio can be "interesting" with lots of play room. My apologies for things not being any easier, but sometimes we have to bang our heads against concrete walls to enlighten ourselves and not many really seem to do that.

 

[RetroAudio]

also, Googling on specific opamps will go a long way, but by all means read anything you can.

 

[lineup]

I see many places where capacitors are inserted between the opamp output and input to prevent capacitive oscillation. I had bought linkwitzs crossover boards to learn from (decided I can't use for my application), and in several places, but not all, he uses feedback capacitors. Is there any easy (for mechanical engineer) rule about this to decide whether they are needed in the design before I make it?

I think he might use one every time there is just a resisitor in the feedback loop. Thanks for any help.

If you use unity gain Op-Amps, I see no reason for using external compensation cap.
They are already using internal comp caps.
Most used active crossovers for audio are gain=1 circuits.

OPA2134, TL072, NE5532, OP27 are all stable in unity gain ( gain=1 ).
NE5534 and OP37 are examples of OP-Amps needing a bit external comp at unity gain.

JFET input Op-Amps are popular in active crossovers.
This is because they have high input impedance, and does not need much input bias current.
Because of this, they will not effect the crossover network very much.

Most used JFET input Op-Amps are for example: OPA2134, TL072
TL072 for low cost non-critical circuits.
OPA2134 when you need higher quality.


If datasheet for an OP-Amp says 'unity gain stable'
I would not use any unnecessary capacitor.

 

[mazurek]

Thank you very much lineup. I am using OPA2134s in my crossover, and I have all unity gain or attenuation. This will make building the prototype board easier. Sounds like I will only need to worry about them if I want to build gain boosts or possibly linkwitz transforms.

Thank you all very much for the help.

Also if any other uneducated people like myself read this thread, I found that the bandwidth limiting frequency cap can be chosen using the intuitive lowpass formula f = 1/(2piRC) such that f is above audio bandwidth, and R is the feedback resistor.

 

[jan.didden]

Re: Re: how important are feedback capacitors to prevent oscillation in active crossover?

If you use unity gain Op-Amps, I see no reason for using external compensation cap.
They are already using internal comp caps.
[snip]If datasheet for an OP-Amp says 'unity gain stable'
I would not use any unnecessary capacitor.

Hi Lineup,

I agree with you, but even in these cases you still find sometimes a small feedback cap. Reason is that even unity-gain stable opamps can oscillate if there is for instance a little capacitance from wiring or PCB layout at the inputs. Just a few pF could be troublesome. Still, 220pF seems excessive to me.

Jan Didden