I'm using the AD9129 in a custom board. It is being clocked at 2.8 GHz and is working well for the most part. The initialization sequence in the datasheet is being followed.
The data paths appear solid as verified through the SED.
After initialization, the FIFO thermometer status registers 0x13 through 0x16 read back data values of 1, 1, 1, and 3, respectively. I would have expected them to be the same value.
Post initialization, the FIFO reset procedure using the FRAME signal as described on page 43 of the data sheet was also tried. No change.
The spurious performance is not as good as expected and my hunch is it may have to do with the FIFOs being out of sync. With a data pattern input set to produce a single tone at one frequency, an aliased tone approximately 20 dB down appears at the opposite end of the same Nyquist zone.
The DAC is being clocked by an ADF4351.