The data sheet says "featuring... low current noise" and then never says what it is. Also, "featuring... low input voltage" should probably read "featuring... low input voltage noise."
Unfortunately, data is not available for the current noise density. However, the AD864x is a CMOS amplifier, so we can assume that the current noise will be in the fA/rtHz to pA/rtHz range.
Any chance this data will become available? I have the same question. fA to pA is quite a large range - 3 orders of magnitude.
See the noise writeup in the ADA4661 data sheet. Most current noise for CMOS amps is suspect, but the good news is it usually doesn't matter in most circuits.
I have perused the ADA4661 noise writeup, and also MT-035. It was partially helpful, but I have further questions:
1. If the current noise is mainly due to the ESD protection diodes, does this mean that it is shot noise (white) associated with the reverse leakage current of those diodes? Is there also a significant flicker noise component? Even if this is not rigorously specified in data sheets, is there some generic information that one can use to understand device behavior?
2. If the current noise is mainly shot noise from the ESD diodes, does this mean that there is the usual relationship between leakage current and shot noise, as commonly known for PN junctions? Can this leakage current be assumed to be the dominant input bias current, or are there other effects? Is there significant bias current cancellation even in a CMOS input, so that the relationship between bias current and current noise doesn't hold anymore?
3. What do you mean when you say "current noise for CMOS amps is suspect"? Do you want to say it is poorly controlled and therefore has to be assumed to vary greatly from part to part?
Whether current noise matters is of course dependent on the source impedance. In fact, once both current noise and voltage noise are known, it is easy to determine the crossover impedance, i.e. the source impedance where both noises contribute equally. Any further increase in source impedance makes the current noise dominant. Now think about a capacitive source, i.e. a sensor. The lower the frequency, the higher its impedance becomes. Somewhere, current noise WILL dominate. Hence the question does have some importance, considering that OpAmps with CMOS inputs will be chosen for such high source impedance applications exactly because of their low current noise.
Current noise for CMOS parts is extremely difficult to measure. While we cannot say exactly what it is because of physical limits of measurement equipment, I can tell you that it is below 50fA/rtHz at 1kHz.
bolerm - See answer above. Also, thank you for the suggestion - we are actually in the process of measuring and releasing that information for our CMOS parts as well.
Retrieving data ...