If you do a manual temperature conversion on the AD9361, it produces a small glitch in the phase noise of the Rx VCO when the conversion completes. Typically about 22 deg for about 2us at 2GHz.
Are you using Evaluation board or custom board ?
Do you see any variation in power supply during this period.?
Many thanks, its a custom board. We are using a recommended design for the power supplies to the VCOs. Phase noise is otherwise excellent. I think any spike in voltage is going to be too small to measure. It not a problem as we have now re-timed the manual temperature conversion start and end so it does not co-inside with anything critical. Just more an observation. We captured the event by injecting a CW into the Rx and recording the digital I and Q values.
Thanks for letting us know.
One comment. The temperature sensor readings are not accurate even with a one time factory calibration. The temperature sensor should only be used by non-critical functions that are not sensitive to the wide tolerances .
Yep we calibrate and use only to keep the RF power at the output of PA devices roughly constant. We are using one LDO for the 1v3 rail. I think you now suggest using two LDO's keeping the synth supplies separate. This may well be why we see the phase noise glitch.
Retrieving data ...