Crosstalk between multiple AD8307

The beauty and difficulty of a device such as AD8307 is that it can detect very small signals. This includes unwanted feedthrough. So first task is to figure out the feedthrough mechanism. I would start with the RF input. If your ac-coupling caps are close to the INP and INM inputs, try soldering down some 0 ohm resistors on the signal side of the caps. If you do this and find that the dc level from that channel drops, that is a sign that the signal is coupling on to the RF input traces. If there is no change, that suggests that the coupling mechanism is elsewhere. The next thing I would try to do vary the capacitance on the VPS line. Each chip should have at least one power supply decoupling cap located close to pin 7 (VPS). You could try to add more capacitance to each pin, say, 10nF or 1nF caps on each of the four VPS pins. You could also go the other route and pull off the caps. If the problem gets worse, that is an indicator that the problem may be with the power supplies. Another approach you could take (if practical) would be to power all four devices from different power supplies.

So I would start with these experiments.  Note also that the COM (ground) pin should connect directly to a low impedance ground plane.

Thank you for the suggestions.  I'll test and see whether it helps.

The inputs were coupling to each other due to poor grounding of my circuit board to the enclosure.   Some copper tape and all is well.  Thank you for your assistance!