I have been using a standard ADC12D1600 with great performance in the DESIQ mode. I am sampling with a 1.2GHz clock and an RF either in the 1st or second nyquist zone. The auto cal feature brings all spurs below -60dBm 100% of the time and typically less than -70. I populated a second board with the ADC12D1600RF part expecting to see the same or better performance especially above 2GHz RF input.
What I have found is that the noise is less, the spur level is generally higher, but still around -60dBm. The image around the FS/2 is -58 to -60 after cal in the normal DESIQ mode, but climbs to -30dBm to -50dBm when I go to the DESCLKIQ mode. The high frequency response is 3-4dB better above 2GHz, but the image is horrible. I have tried adjusting the DES timing adjust and I have found that a value of 0 for DTA is best, but that only gets me to around -50dBm.
Is this poor performance expected of this new part or is there something else that can be done?
I have also adjusted the FSR with hardly any change.
Thanks,
Dan