Weird signal instablity of an opamp amplifier

[VanitarNordic]

I have configured a well known TL071 opamp in single supply and AC coupled inverting amplifier. with 10K and 1M resistors, it produces around 100x gain.

I inject a 10mV sinewave signal to the input. for the 10mV signal,  I should get around 1V after amplification, it is almost okay, but look at that weird instability (the video has attached). The frequency is just 300Hz and if I increase it, this instability would be stronger.

The supply is regulated. oscilloscope has no noise. The signal generator is the Siglent brand and no noise. I have tested the circuit also with a battery but the same results. with this setting, The GBW is about 300 * 100 = 30Khz, which is far away beyond opamp limits, but why this happens?   

Has anybody experienced this? is this a fake opamp?

[StillTrying]

It looks like mains hum to me, see what it looks like with the scope's trigger on AC LINE.
I'd start with 1Vpp and divide it by 100 on the board right at the op amp's input to give a clean 10mVpp.

[VanitarNordic]

It looks like mains hum to me, see what it looks like with the scope's trigger on AC LINE.
I'd start with 1Vpp and divide it by 100 on the board right at the op amp's input to give a clean 10mVpp.

I'll try your suggestion but if it is the mains hum, then why still exists when I use a battery?

see what it looks like with the scope's trigger on AC LINE.

You mean I should put the scope input coupling on Ac? it is already on the AC coupling.

[StillTrying]

"You mean I should put the scope input coupling on Ac?"

No, Trigger Input to AC LINE to see if the wobbles are synchronized with the mains, if I stop the vid and look at 2 very similar peaks they're often 20ms apart.

"why still exists when I use a battery?"

Mains noise maybe coming from the signal generator, cables or GND connections.

[VanitarNordic]

I'd start with 1Vpp and divide it by 100 on the board right at the op amp's input to give a clean 10mVpp.

I did. I did set the generator output to 1V and divide it by a multiturn pot to 10mV to handle 1V at the opamp output. The instability is still present.

Yellow: channel two (1V from the generator)
Red: Opamp output

[StillTrying]

It still looks like mains and SMPS noise to me, I often get quite bit worse than that and have to use "Waveform Averaging" to steady the trace.

I'm disappearing for a few hours, only meant to have a quick look in here! Hopefully you'll get a more expurt opinion/solution.

[VanitarNordic]

No, Trigger Input to AC LINE to see if the wobbles are synchronized with the mains, if I stop the vid and look at 2 very similar peaks they're often 20ms apart.

The scope itself had AC Line triggering inside its trigger menu. I put on that and here is the results

[Conrad Hoffman]

Looks like hum pickup. That's significant gain- do you have everything shielded and are using bypass capacitors?

[Wimberleytech]

If it is hum, (which it appears to me), slow down the sweep and measure it to be sure.

[SiliconWizard]

It doesn't seem synchronized to mains.

Why don't you select a longer timebase so you can actually figure out the exact period of the apparent modulation? That will help diagnosing the source of noise.

[SiliconWizard]

It doesn't seem synchronized to mains.

Why don't you select a longer timebase so you can actually figure out the exact period of the apparent modulation? That will help to diagnose the source of the noise.

Actually, this instability intensifies when I increase the signal frequency, which I think should not if the noise is the low-frequency 50Hz hum. It does not get fixed when I use a battery.

so far I could reduce this instability by choosing a different value for the input capacitor (opamp AC coupling). reducing from 10uF to 2.2uF reduced the instability significantly.

Does the modulation change in frequency too?
Is the opamp powered by a single- or dual-supply? Can you post a schematic maybe?

[StillTrying]

There's more than enough evidence to show it's mains hum + SMPS noise, and not the op amp for me, I've seen it too often!

[KE5FX]

There's more than enough evidence to show it's mains hum + SMPS noise, and not the op amp for me, I've seen it too often!

I have seen some misbehavior from otherwise-trustworthy opamps when using high source impedances.  That waveform looks a bit like what I observed a while back on the LT1677.  (Link goes to a Usenet thread where I asked for some peer review.)

The takeaway from that experience, at least from my point of view, was "don't use opamps with R-R input stages if you don't actually need that particular feature, and don't use high source impedances when you do."  I don't know if the TL071 would be vulnerable to anything like that or not.  The datasheet does say that the common-mode range includes the positive rail.  Try (much) lower resistances and see if that changes the behavior?

[VanitarNordic]

There's more than enough evidence to show it's mains hum + SMPS noise, and not the op amp for me, I've seen it too often!

You mean the small noises in valleys? it happens because of the connected voltmeter. otherwise, if I disconnect the voltmeter, the peak and valleys are clean, but this something like signal dancing effect (I don't know what this specific signal behavior called) still exists. Anyway, for a project, I work with small signals amplification and if I could not remove this annoying phenomenon, I can not go to the next steps and examine other parameters in practice.

[VanitarNordic]

Try (much) lower resistances and see if that changes the behavior?

I have tested a few more part numbers such as LM318, TL072, TL081/2 and all show this behavior. I think it might be the hum noise, but I should remove it before I go to the next steps

[SiliconWizard]

Have you tried putting your circuit away from potential noise sources?
What's the set output impedance of your signal gen?

[HB9EVI]

If you're not limited by the output impedance of the signal source, you should try lowering your system impedance; like 100k/1k

[VanitarNordic]

Have you tried putting your circuit away from potential noise sources?
What's the set output impedance of your signal gen?

If you're not limited by the output impedance of the signal source, you should try lowering your system impedance; like 100k/1k

The signal generator is the Siglent 1025. The output can be adjusted to 50ohm or High-Z. I swapped between these two options, but no difference

[HB9EVI]

You cannot adjust the output impedance - this function only allows to have the reading of the signal level related to the system impedance - on the physical signal itself it has no influence

[VanitarNordic]

I reached this conclusion also that this problem causes by the 50Hz mains hum. Thank you for all comments and contributions 

if there are some practical suggestions to minimize it, I would be glad to know since sometimes we have to deal with small signals and sensitive circuits

[StillTrying]

In a room with mains, a PSU, a SG and a scope all connected, 0.5mV of total noise on a 10mV signal might be about the best you can get.

Have a good thick GND.

Decoupling caps very near the op amps supply pin(s).

Make sure there's no noise on the inverting amps +ve input.

Some of the fast SMPS spike noise won't be real - just picked up by the scope's probes and GNDs, at these frequencies you could use the probe on X1 to filter some of it.

Lower all the impedances quite a bit.

After that it's a full GND plain and screened box.

That's all I can think of!

[HB9EVI]

Depending on your setup, it's very likely common mode noise. If you use the scope probe with long ground tail clip, then you very likely have culprit.

The only way to be sure, that the circuit is clean, is keeping the ground connection between circuit and scope as short as possible