Dear All

Sorry for making my point unclear in the title. Could I get some help on this question?

Must the amplitude of oscillation go to power rail of an OpAmp circuit?

Here is what we have done:
1. We designed a preAmp for a photodiode. And we observed frequency harmonics of 100kHz, 200kHz, 300kHz,...and so on. We are not sure if they are noises from the laser or they are oscillation.

2. We designed a band-pass filter after the preAmp. We observed harmonics of 2kHz, 4kHz, 6kHz and up to 10kHz at the output of preAmp. We think they are circuit oscillations, since they still exist even when we turn off the laser. However, the oscillations do not go to power rail of the OpAmp circuit....

Could I have some comments on following question?
1. How to differentiate noises with oscillation?
2. Must oscillation go to power rail of the OpAmp? If not, what decides the amplitude of the oscillation? It is interesting that the amplitude of the oscillation we observed is constant through repeated experiments..

Any comments will be highly appreciated.

Regards,
Richie

 

How to differentiate noises with oscillation?

Noise often has a spikey or apparently random waveform whereas 'oscillation' may be more regular or sinusoidal. That is not always the case. Any unwanted signal is regarded as noise, so unwanted 'oscillation' can be 'noise'.

Must oscillation go to power rail of the OpAmp?

Not necessarily. In fact many opamps can't drive their outputs within a volt or two of either power rail.

 

It seems all the psychic electronics guys have left the forum, so first you need to post your schematic. A picture of the build may also help. Then either snap pictures of the scope image or sketch the waveforms you see.

That way someone might see something beyond asking "did you add bypass caps close to the amps?"

 

What does this "oscillation" look like on an oscilloscope?

 

My psychic foo on the missing schematic is detecting that there are not enough PSU decoupling caps.

 

Noise often has a spikey or apparently random waveform whereas 'oscillation' may be more regular or sinusoidal. That is not always the case. Any unwanted signal is regarded as noise, so unwanted 'oscillation' can be 'noise'.
Not necessarily. In fact many opamps can't drive their outputs within a volt or two of either power rail.

Thanks for your reply. Could I ask if oscillation must go to the power rail-margin?
I mean, is it possible to have a, say 1V oscillation, on a +-12V supply OpAmp band pass filter?

 

Dear all, about the caps:
We put 5 caps for every OpAmp, (10u,0.1u) for each power rail and one 0.1u between +_ power rails.

The attached two pictures show the condition of the BandPass filter:
Picture 1: Time domain diagram
Picture 2: Frequency domain diagram

Thanks a lot for your help..

 

is it possible to have a, say 1V oscillation, on a +-12V supply OpAmp band pass filter?

Yes. Any amplifier of finite gain will provide a low output amplitude 'oscillation' if the input signal is an even lower amplitude oscillation.
We really need to see your schematic to help further.

 

The high frequency oscillation amplitude of an op amp can be limited by the slew-rate limit of the opamp.