USRP 2921 phase drift at high Tx gain

[hkhalili]

Dear NI RF experts :-),

I am transmitting a QPSK signal on one USRP 2921 and receiving it on another USRP 2921 at 2.4 GHz. The two USRPs use the same reference signals, and hence, there is no frequency offset between them.

There seems to be a phase drift in long symbols. Therefore, the phenomenon is worse in low signalling rate. It is much better at low transmit gain.

Please find attached a snapshot of the received signal (I and Q) and the demodulated symbols demonstrating this phase drift. This was captured when the transmit gain was 30 dB. 

Could you shed some light on this? Is this related to amplifier saturation or possibly to Tx signal processing?

[muffin.vi]

Are you using a MIMO cable, or are you using external 10 MHz and PPS signals into the SMA connectors? 

[hkhalili]

Hi,

I have used both setups (MIMO cable and external signals), and the phase drift occurs in both. 

It almost looks like a capacitance affect that settles after about 100 samples at an IQ rate of 25 MHz. (see attached)