AD8676 popcorn noise

Hello,

I have some questions about the AD8675 (or 8676) and its popcorn noise.

I plan to use an AD8676 (or AD8675) in order to measure [0,1Hz;10Hz] noise of a low noise voltage source. I have chosen this amplifier as, according to the datasheet, it has great performances at low frequencies.

Here is a simple view of the circuit:

A capacitor and a resistor are used for AC coupling (fc ~0,8Hz). The opamp amplifies the signal with a 60dB gain. Then, another stage filters the output (not represented here).

In order to test my circuit, I short the input and look at the output. I then filter the output (fc=40Hz).  In theory, I should measure aproximately the amplifier (+resistors) noise times the gain.

You can find below some captures made with an osciloscope (voltages are in mV, 500ms/div).

We can see some noise alternating in between two levels (don't forget that the voltage has been amplified by 10000). This behaviour is strange and I tried to remove it by changing the resistors and the capacitors, none of it worked. Then I changed the opamp with another low noise reference (not an AD's one) and this noise disapeared !

This type of noise clearly is a problem for my circuit as it limits the noise levels I'll be able to measure.

So, I have a few questions for you :

- this noise looks like popcorn noise, is it correct ?

- do you have any data/information about this noise in your products ?

- do you have any advice to remove/reduce this type of noise

Thanks a lot for your time,

Have a nice day,

Jean-Baptiste