Here is a comparison of the L-arm VDC u1 drift time spectra for event types 4 (L/R coincidence based on scintillators only) and event type 5 (L-arm singles trigger). For event type 4 only, there is clear evidence of double-timing - part of the spectrum is shifted by ~100 ns to earlier times. Meanwhile, the timing of trigger 5 is very clean.
Trigger 4 does not seem to be configured very well ... perhaps an expert can look into it?
(I am doing more replay right now to get enough statistics for trigger type 6 - the one we really want.)