We are using AD8347 to demodulate an OOK signal, at a rate of several x100 Hz. The purpose is a radar-like application. The signal of a VCO at 2.4 GHz is split, half is fed to the LO of AD8347 and half is fed to the TX path and modulated with OOK. The echo is received by the RX path, which is fed to the AD8347 RFIN. The undesired return signal can be canceled out using a tunable analog loop.
When the switch is OFF the RFIN is quasi nul. When the switch is ON, the RFIN can be canceled out using the analog loop to be almost zero too (< -60 dBm). Our expectation is that the output will be stable, but it is not (check the provided doc), there is a DC introduced when the switch is ON. We are using a standard reference design with the AGC disabled (manual gain setting) and the DC filter enabled with the default values.
Thanks for helping us understand what can be going wrong.