One reason is that I added some checks before invoking MovePT() so as not to call it if any of the stopping action bits were set. (This hasn’t always been the case.)
When a probe strikes, we attempt to transition the MultiAxis motion state to STOPPED/IDLE while we clean up. I suspect this accounts for the long gaps around the probe strikes.
So, the AT_TARGET probably indicates that we’ve sent a stop point and nothing’s MOVING any more. We normally send motion every 10ms.
Here’s the same data for one of the other thousands of strikes.
It looks like there’s still a gap preceding the TriggeredModify of maybe 35ms (good scenario) and 66ms (bad scenario).
Nothing I’ve said previously would account for that.