op-amps - phase response?

[Dominique]

Hi everyone,
I'm new to this place and I hope that my first question isn't too stupid for this forum so I will be excluded!
I read some stuff on op-amps, but couldn't find an answer to my problem...

While looking at a few OP-amp datasheets,
I was wondering about their phase characteristics...

In most datasheets you can find a diagram with "open-loop gain" and phase over frequency.
Together with a few other parameters (noise and thd as the most important), I used that as a criterium to filter out OP-amps with unwanted characteristics. Namely: in case a considerable shift of phase between 20Hz and 20kHz appeared I excluded them from my list of possible audio op-amps.

But in some datasheet I found such a diagram each for open-loop AND for closed loop, and the phase over frequency curve was pretty different for both cases, so I imagine that the whole phase characteristic would change with the gain? Could somebody explain this bevaviour in more or less simple words?

How do you guys see if an op-amp is good for audio from the specs? I guess that some of the higher GBW op-amps (gain x bandwidth) not definitely designed for audio would be fine too, but how can I distinguish them from those which aren't - phase-response wise - and at a given gain?

Cheers
Dominique

[SY]

My first suggestion is to read up on feedback and opamp theory in a good basic text like Horowitz and Hill's "The Art of Electronics." The mysteries of Bode plots will become much, much clearer.

To get to your specific question, in a minimum-phase device like an opamp, the phase response is intimately tied to the frequency response. In fact, it's basically the first derivative of the frequency response with respect to frequency. (derivative = rate of change) So in a region where the response is flat, the phase shift will be zero. In opamp open-loop plots, you see the effect of the dominant pole, that is, a deliberately induced low-pass rolloff. The reason this is built into the opamp is to stabilize it when the feedback loop is closed. And indeed, you'll notice that the closed-loop plots show a much flatter, wide-band frequency response and correspondingly, a low degree of phase shift through the passband.

As a passing comment, don't get too excited about fantastic bandwidths- high speed circuits are a pain to get stabilized and any audible benefits are highly debatable, to say the least.

[Dominique]

My first suggestion is to read up on feedback and opamp theory in a good basic text like Horowitz and Hill's "The Art of Electronics." The mysteries of Bode plots will become much, much clearer.

that's what I already did. I probably am missing too much of the basics...

To get to your specific question, in a minimum-phase device like an opamp, the phase response is intimately tied to the frequency response. In fact, it's basically the first derivative of the frequency response with respect to frequency.

Wow, that's great to know!
That helps me a lot!

As a passing comment, don't get too excited about fantastic bandwidths- high speed circuits are a pain to get stabilized and any audible benefits are highly debatable, to say the least.

No, I won't! In fact, like in all applications a little surplus is soothing, but if I get e.g 5 or 10 times what I need, I fear that I have to pay that somehow somewhere else...
But that increases the choice of op-amps greatly... I'll look through the datasheets tomorrow

After all, I DID get a lilttle excited, but temporary excitement is the fate of the diy'er, I think!

Thanks a lot, SY
Dominique

[jcarr]

Dominique: Note that it is also possible to use two opamps in a composite circuit for improved phase response. James Wong of Analog Devices wrote an application note on this.

See:

Duplicate opamp inside feedback loop?

And if you can't find AN107, study the OP-271 data sheet, also from Analog Devices.

http://www.analog.com/UploadedFiles/...557OP271_a.pdf

Unlike SY, I like wide-bandwidth amplifier circuits, as they tend to give me a sound that subjectively I prefer. But it is true that they can be a pain to stabilize. See:

http://www.diyaudio.com/forums/showt...867#post205867

http://www.diyaudio.com/forums/showt...957#post205957

http://www.diyaudio.com/forums/showt...000#post206000

I definitely recommend low-pass input filters.

hth, jonathan carr

[peranders]

Originally posted by Dominique
How do you guys see if an op-amp is good for audio from the specs? I guess that some of the higher GBW op-amps (gain x bandwidth) not definitely designed for audio would be fine too, but how can I distinguish them from those which aren't - phase-response wise - and at a given gain?

Most opamps with a speed over 10 V/us are more or less suitable for audio but in some cases you must look more carefully at the performance such as noise, current capability, speed, input bias currents etc.

[sreten]

http://www.dself.dsl.pipex.com/ampins/webbop/opamp.htm

some useful info.

sreten.

[Dominique]

thank you all! Now that gives me SOME more things to read!
I'm very glad - the responses were more helpful than I dared to expect!

Best wishes!
Dominique