Dear all,
designing with AD8421 is really a nice thing, because this device has superb performance characteristics making it a very good part. However looking at following questions related to it shows it has some hidden pit falls, too.
This is similar to its faster and more current hungry equivalent AD8429.
ad8421-output-oscillation-with-no-input-signal
ad8429-gain-peaking-at-high-frequency
ad8421-ad8429-gain-selector-switch-drawbacks
What to make sure, that such things won't happen?
- Both inputs (+IN, -IN) are sensitive to RF signals > 1MHz
- Limit possible input frequencies with proper input filtering -> well described in datasheet
- RG inputs are sensitive to RF signals similar to the inputs (+IN, -IN), however here it is hard to filter, because any capacitance at these nodes change the gain and frequency response
- don't add much parasitic capacitance to these nodes, if you want a smooth frequency response without excessive gain peaks
- keep any high voltage AC-signals away from this signal traces
- that is true even for the output signal, as this introduces a unintended/parasitic feedback loop
- thus short traces keep things simple here, the longer they get, the easier is it for other signals to couple into RG
- As RF signals don't need large structures to crosscouple -> even short copper traces in parallel or crossing each other on adjacent pcb layers can introduce trouble here
- Same is true on bread boards and so on
- Tuning frequency response of this part is a hard thing (?), filtering with additional circuits is much easier
How about you do you found any other point for these InAmps, or are some of my hints not hitting the right point?
Best regards,
tomspe
Format correction
[edited by: tomspe at 1:01 PM (GMT -4) on 25 Oct 2018]
