I use the ADF4350 to synthesize a 1282 MHz local oscillator signal for a mixer. After configuring the device (details see below), the Lock Detect output goes high, and the device produces the expected output:
(This signal is measured by placing an antenna close to the synthesizer output track; since I'm not interested in the amplitude of the signal currently, this seemed to be the least invasive and simplest way to observe it)
However, after waiting for 20 to 30 seconds, the output gradually turns into a broad frequency comb as shown below:
I suspect this has something to do with device temperature, since turning it off and on again immediately (1 second off) retains the error, while turning it off, waiting two minutes, then turning it on again makes it return to the state shown in the first image.
What could be wrong here? Below, you can find the schematic, register setup, and layout of the board.
It is possible there is a soldering problem, since this is a hand-soldered prototype, although I checked quite carefully and also wouldn't know what kind of soldering problem could cause the observed behaviour.
Thanks in advance for any help!
Register setup (input is from 16 MHz crystal oscillator):
reg_0 = 0b00000000101000000000011110000000
reg_1 = 0b00000000000000001000111100000001
reg_2 = 0b00000000000000001001111011000010
reg_3 = 0b00000000000000000000000000000011
reg_4 = 0b00000000100101111000010000100100
reg_5 = 0b00000000010110000000000000000101
Board Layout (this is a 4-layer board, the second layer is solid ground but is not shown here):