I am trying to use the ADF4107 chip to frequency and phase lock a slave laser to a master laser (already locked), as performed in previous experiments (see http://iopscience.iop.org/article/10.1088/0957-0233/20/5/055302). We hope to modulate the slave laser frequency by taking the error signal output from the ADF4107 and beat that with the master laser. This will then be input back into the ADF4107 to act as a feedback look, instead of using a VCO as the board is initially intended.
However, after some initial testing using a signal generator (N5183B microwave signal generator) I have found that the board doesn't produce the expected outputs, for any frequencies lower than about 4.7GHz. I then observed the N divider output using the MUXOUT port on the ADF411xSD1Z and found that the board wasn't dividing the input signal as expected at these lower frequencies (see fig.1). It seemed to behave as though there were some other high-frequency signal driving the N divider in the ADF 4107. Indeed, when the input signal was removed all together, we found that the board appeared to be dividing some signal of about 7.35GHz.
Fig.1. The N divider output frequency as a function of input frequency, for different values of the predivider P, N=1000 and R=10 throughout. The 'Expected' line is simply the input frequency divided by N (Input/1000).
Fig.2. This is the setup of the ADF4107 chip on the ADF411xSD1Z. However, we are using the onboard 10MHz TCXO instead of the input into REF IN that is shown here.
Does anyone have any ideas about the origin of this mystery signal and what we can do to get rid of it? It will eventually be necessary for the beat frequencies to be around 3GHz and so we need to try and sort this out.