We are finding in a lot of 10 modules that a very narrow phase noise problem occurs with the ADF4155 when locking to 4.8 to 5.6GHz. When a unit has a problem it may only be 5MHz wide or perhaps 40MHz wide, usually just 20MHz wide. The phase noise will go from -105dBc/Hz in 99% of the band to -50dBc/Hz at 30kHz offset at the one particular bad spot. The bad spot is the same place consistently(+/-2MHz) on any given unit, but unit to unit, the bad spot can be found from 5GHz to 5.5GHz. We were able to duplicate this problem on our ADI Eval board for the ADF4155 as well. Its not related to input RF feedback level(varied that by 15dB to no effect on either good spots or bad spots), or loop constant values, or dithering. Cold spray on just the PLL chip can shift the problem 200MHZ lower than its original location, but the problem does not go away.