Hi,
I'm in the process of modelling an increase in cycle length along a corridor from 80 to 100 seconds; it is the CBD, so the queuing and delay results are important.
At one intersection in particular I ran into a strange bug. The splits were previously 55/25 (total 80) for main street/side street. When I attempt to increase the main street split to 75 seconds -- leaving side street alone at 25 sec, the model shows total delay and queue length doubling for main street in the southbound direction only (northbound is unaffected). However, if I code the new splits as 74/26 or 76/24, the delay and queue length for main street in both directions are reduced as one would expect. It is only when I use the value of 75 seconds that I obtain the bad results, and then it is only for one movement!
Is this a known issue with Synchro 6? In any event, it seems to be one worth fixing...

Cheers,
Campbell