The frequency drift needed to be addressed. I modified one of my original LiteVNAs to include a external clock reference input. I was planning to modify the latest revision of the LiteVNA I have but it turned out there appears to be a problem that causes 50Hz dips in the frequency. I repeated the frequency stability test using my third LiteVNA which is also the older revision of hardware and it also does not exhibit the problem. While I have raised the issue, I doubt it will be addressed so future tests will be based on the older hardware.
Shown is the modified VNA. I pulled the oscillator. Added an SMA with a TVS and termination. The LiteVNA requires the clock to be 24.0MHz. For now, I am just using my RF generator to derive it. With everything now locked to the GPS, we can measure the resonance of the cavity over an extended time.