EV-ADF5356SD1Z problem

I first purchased one EV-ADF5356SD1Z evaluation board, which showed off spec performance by generating 3rd harmonic with a higher amplitude than that of 1st harmonic, i.e. about 2 orders of magnitude off the specs (-14 dBc for 3rd harmonic), and revealed a spectral breakdown to odd harmonics of approximately equal amplitude: 1st, 3rd, 5th, 7th, and so on. I mainly tested at the minimal frequencies from 53 MHz to about 1 GHz and traced the outputs with a 50 Ohm and 1 MOhm switchable inputs of a 200 MHz bandwidth oscilloscope, and traced higher harmonics with a 3 GHz limited spectrum analyzer. The free outputs (RFoutB and the unused RFoutA) are terminated with 50 Ohm. I only occasionally see correct frequency at the outputs, and NEVER see a single harmonic generated, i.e. a visually pure sine! Moreover, suspecting a damage, I bought another identical board from another vendor and observed exactly the same behavior. I only used the SDP-S/USB2.0 and the v1.4.4 evaluation software, along with the 6.0V DC power supply, and proper grounding. I also do cautiously use the board to avoid static electricity or touching conductors. I checked a variety of settings and tried to terminate SMA REF inputs with 50 Ohm stubs too. Nothing looks right: neither the actual frequency (it does not coincide with the set value), nor the spectrum which is supposed to be single line. Can you please advice, what can be done in this case?

attached images
[edited by: tormondo at 5:29 AM (GMT 0) on 15 Feb 2019]