Hello!
From my observations, almost all modern low voltage noise bipolar opamps have some sort of bias current compensation.
If the opamp input sees an impedance greater than zero, the product of the impedance and the noise current adds an additional noise voltage to the circuit.
Because bias current compensated opamps have additional transistors connected to the input pins (those of the compensation current sources) those current noise adds to the current noise of the input transistors.
This can clearly be seen in the datasheet. So far so so good.
Some literature states, that the noise current of bias compensated opamps is at least partly correlated and this leads to a noise penalty, if the impedances seen by the opamp inputs are not equal.
Only a few datasheets (e.g. AD8597) differentiate between corraleted and uncorrelated current noise.
I have the following questions:
Is it correct, that current noise of bias compensated opamps is (at least partly) correlated?
If yes, is there a rule of thumb, how much is correlated?
Does this depend on the implementation of the compensation?
A standard high level noise model of an opamp uses two uncorrelated current noise sources. Is the potential current noise correlation usually considered in the spice models?
Best regards
