I am using AD9954 DDS to provide RF signals for RF device in sensor application. The physical properties is obtained by measuring the phase variation between the signal from the RF device and the exiting signal from the DDS (Fig. 1). With this approach we can achieve a sub 10 mili-degree standard deviation of the phase.
We want to now to use AD9958 to provide the used signals (Fig. 2) without splittering the exciting signal. To test this approach we use AD9958 demo board and the same AD8302 board as used in the first application. The results obtained with the second configuration show a non-coherence of the phase between the to channels :
@ 125 MHz (same frquency as used for AD9954 tests) we obtained a standard deviation of the phase about 50 mili-degree.
@ 200 MHz we obtained a standard deviation of the phase about 200 mili-degree.
@ 20 MHz we obtained a standard deviation of the phase about 300 mili-degree.
It's look like the LUT is not read in the same way for the two channels which implies phase fluctuation between the two signals and so on the AD8302 output.
I am right?
Is there a way to obtain real coherent phase between the to RF outputs ?