Residual frequency content left after signal has passed when capturing 250MHz tone. The ADCs are running at 1GHz. From the data sheet we understand there is a small residual tone at fs/4 from the interleaving technique but it is usually bellow our system noisefloor so doesn't show up when we're not receiving signal.
A change to our system means that one of the carriers we want to demodulate is now at 250MHz and we've started noticing that content at 250MHz is now present above our noisefloor for the periods we're not capturing the carrier.
This can be recreated by just pushing a 250MHz signal from a sig gen into the ADCs. If I capture 1ms of data and do a spectragram and plot the energy at 250MHz I get the below plot which shows the pulse near the beginning with the expected large content but a gradual tail off of the energy long after the pulse has gone. The yellow plot is the energy present when the pulse is stopped .
I can't see anything in the data sheet that says capturing data at fs/4 is a problem but can imagine that maybe the DC correction stuff can get in a tizz. If I nudge the tone done to 245MHz then I do get a bit of 250MHz energy at the start and end of the pulses, which is to be expected, and the level after the pulse has finished correctly returns to the initial level for when there is no pulse.
