I have a signal in the range 2000-4000 MHz which contains a (very) small phase shift relative to another signal at 500 MHz (my reference signal). In order to measure this phase shift I use an ADL5380-EVALZ module to demodulate each of these signals and then compare them. Currently I am using the board in its default configuration, with single-ended inputs and outputs, which means there is a lower limit on the frequency of the I and Q outputs of 2 MHz, due to the balun used at the output. This would be fine, however at 2 MHz I am having trouble resolving the small phase shift difference I expect in my signal. I believe my problem would be solved if I were able to demodulate the signal to baseband, and combine the DC I and Q outputs to give a signal proportional to the phase shift (sqrt(I^2+Q^2)).
I see that the balun can be removed quite straightforwardly, giving a differential output, but the problem with this is that the output impedance is then 400 Ohms, and I would like to record the data using a digitizer with 50 Ohm input impedance.
I am looking for the best method for getting around this problem, so that I can demodulate the signal to DC and measure it easily. Are there any AD components that may be suitable to interface the two?
Note that my system needs to be as low noise as possible, as the phase shifts to be detected, and the signal I obtain, are both very small.
Your help is very appreciated!