It all became pretty obvious once I thought about it but at the time it was something I did not consider.
I was driving a domestic sewing machine with a steel reinforced, T5, synchropower, timing belt. I set the correct belt tension and the driven load was well within the belt manufacturer’s specification. What I, rather stupidly, failed to consider was that the driven mass increases exponentially with acceleration.
At 300 rpm and with an acceleration, from stationary, of 0.6G’s everything was just fine but in the quest for speed I increased the acceleration to 0.8G’s and then the belt started to stretch.
In the interest of science I re-adjusted the belt tension a number of times and it continued to stretch until such time as the teeth pitch failed to match those of the pulleys.
Basically, with this type of belt the steel reinforcement is helically wound during manufacture and there is no real bond between the wire and the polyurethane so what started out at say 11 complete turns of the wire (within the belt) soon became 10.8 complete turns as the polyurethane stretched.
What I have learnt – I choose the wrong type of belt for the intended application.

Yet more knowledge to file away in my repository.

Tweakie.