Can you test this specifically with some frequencies which are whole divisions of the 250mHz clock. Like 50.00, 41.66, 35.71, 31.25, 27.77, 25.00 MHz
I did the measurements at the requested frequencies, a lot more carefully than the previous day's rough look, and took the average over four separate trials, after a 30 minute warm up period. The Ch1 trace was centered exactly on the crosshairs at 25MHz, and the Ch2 lag time was measured at each frequency, the results being:
25.00MHZ 1.00nS
27.77MHz 1.00nS
31.25MHZ 1.10nS
35.71MHz 1.00nS
41.66MHz 1.10nS
50.00MHz 1.20nS
In fact, the relative phase shift was constant from at least 5Mhz to 50MHZ, but both traces kick laterally by about 2nS at 20MHz - otherwise their positions are quite stable.
One thing I did notice today, however, is that the Ch2 amplitude is about 0.15v less than Ch1's when set to 5v at 25MHz, when the channels are supposed to be in sync, and the difference appears to get bigger as the frequency increases. I'll have to have another look at that to see exactly whats going on over the whole range, and see if other waveforms are affected.
Regards,
Dave