I have a 64QAM waveform that is generated in tx HDL and looped back to rx HDL and can properly demodulate with zero BER. The first plot shows the digital loopback and capture the output of the FFT and the output of the Corrected LMSE. The begining of the waveform shows a 13-bit preamble and the center shows a reference 32-bit BPSK symbol. As you can see the Q path of the BPSK section should be all zeros. The same waveform is sent through the AD9361 TX1A and RF loopback to RX1A. The results show there is an IQ mismatch or some other distortion on the signal. I have made sure the IQ cal quad/dc tracking is enabled on the receiver. Is this a matter of LO synchronization?