I just received the eval board for the 9518-1 and it has very bad phase noise performance. 1st off I have tried to lock it to a 10,20,50,100 or 200 MHz clock in both single ended, on ref1 or ref 2 and as a diff signal on both inputs. I am trying to generate a 2.4GHz VCO and output this to one of the ports. The device will sometimes say it is locked up, but the phase noise varies from about 100KHz wide with no discernible center frequency using a 10 or 20MHz reference to the expected tone with about 20-30dBc noise using a 200MHz reference. I have seen RC oscillators more stable than this is(think 555 timer for the old folks). It seems to fall out of lock if I make the lock criteria too tight. If I turn the PLL off and just let it run free, the phase noise is great, but of course the signal quickly walks away. The source is an HP synthesizer...very stable with excellent phase noise.
If I input an external clock of 2.4GHz and turn off the VCO, the output is just as I expect with essentially the same phase noise as the signal generator. The internal VCO is just horrible.
This is the rev b eval board with no changes, hot off the press as it were. I just want to make the eval board work. I have contacted tech support and received the typical questions, is it powered up, is it connected to the software, have you read the documentation...etc. The answer is yes to all of these. I have tried every register setting I can think of. I have run the phase detector any where from the recommended 10MHz up to 50 MHz and down to 50KHz with no change.
Do I have a bad board or is this part just not designed to generate a stable signal?