High gain Transimpedance amplifier - 40 Hz noise

[arivalagan13]

I bought the OPA656 evaluation board and configured it as TIA with a gain of 10^9. Used application circuit from the opa656 datasheet.

I enclosed the amplifier in an aluminium shield to avoid EMI.

The amplifier is now connected to a CRO.

Now, I try to disturb the amplifier with body current (since I don't have a precision low current source), the amplifier is oscillating with a square wave with frequency roughly at 40 Hz. The amplitudes of the square wave are the saturating voltages of the opamp. The system is powered with +/-5V regulator which in turn powered from two 9V batteries and no transformers are involved in the design. The power supply is enclosed in a separate aluminium shield. Both aluminium shields are professional ingress protection box bought from SparkFun.



What could have gone wrong? The CRO is of analog type and with an only single trace. I checked the ground loop to my best and found to be ok with it. And by the way, for the information, the frequency at India is 50 Hz from grid power.

How do I troubleshoot such a problem?



Regards

M A

India

[tggzzz]

What could have gone wrong?
How do I troubleshoot such a problem?

Many many things could have gone wrong.

You start by providing sufficient information for us to discount the unlikely possibilities. That should include a schematic, photos of the device's interior and exterior and test setup, and a photo of the output.

[alanambrose]

Probably leakage or inadvertent coupling between output and input. A small parallel capacitor in the feedback loop (maybe only 1pF) will probably fix the problem.

Alan

[arivalagan13]

I'll try to add one and update.

[arivalagan13]

I'll update all the details.

[Vovk_Z]

Gain 10e9? Isn't it too much?
It needs only 5 nV to saturate opamp.
I think you have to lover gain to 10e6 max.

[Marco]

Now, I try to disturb the amplifier with body current (since I don't have a precision low current source)

What do you mean body current?

Any way, the input is virtual ground ... just set a lab power supply to 0.1V and use 100 Meg resistor to get 1 nA.