I tried your setup with less sophisticated signal source:
CSA803, internal triggering, 1 ps/div, 1024 points. FG504 set for triangle waveform approx 100 kHz. FG504 trigger out connected to A input of DC510. CSA Calibrator output connected to input B of DC510. DC510 set for measurement of frequency ratio B/A. Attenuators where appropriate.
With FG504 fine tuning I could obtain a just slowly moving display (somewhat erratic of course, moving backward and forward), good enough to freeze with the Run/Stop button. The frequency ratio was 0.9989-0.9990, or about 1 promille difference. The frozen displays nicely showed one triangle period. Cursor measurements of delta-t between nearly endpoints with equal vertical value scattered around 10 ps in repeated measurements.