I am building a frequency translator with ADL5380/5375. I first demod the input to baseband with ADL5380 in ~2.5GHz(f1), then filter the BB signals with active lowpass filters (a few MHz passband). The the baseband signals drive ADL5375 to ~700MHz(f2). The output is amplified by the PA.
I am seeing this annoying high order intermod spurs that fall in my output band. they are exactly as 4Xf2-13f1 and 15Xf1-4f2. The spur level is about -85dBm at the output of ADL5375, so after the amplification it got quite high (~50dBm).
I first thought that it was my LO's harmonics problem even I filter them with LPF. But even when I replace the LO signal with the one from agilent SigGen, it still shows exactly the same spurs. Putting on the metal shielding can (has individual chambers for 5380/filter/5375/PA) doesn't change the spurs at all, so they look conductive.
The whole chain is powered by the same 5V rail with only ferrite beads for some isolation.
I would appreciate any suggestions on where to attack the issue.