We accidentally ran the ultimate test of how well the CenterTrack belt works under very low tension.
After taking our de-coupled tandem on a small charter plane during our New Zealand tour (we were going from the end of one dead-end road to another - they don't put roads over every mountain range there like they do in Europe), we hurriedly put the bike back together at the grass airstrip to ride the couple of miles to the campground in the nearby village. The next day, we rode out to do a hike in the afternoon, and I noticed that something felt a bit funny with the belt - there was a slight lag whenever going from coasting to putting the power down. I didn't check it until we got back to the campground at the end of the afternoon, when I discovered that the belt had been almost completely slack, I believe for about 20 km of riding, including up some really steep pitches on a dirt road.
It appeared that despite the lack of tension and the power I must have put through it at times, there was no skipping, as shown by the pedals still being 90 degrees out of phase, as I had set them initially. At the time I didn't think to check the pedal phase too precisely to know whether or not it had skipped one or two teeth because I was too busy rectifying my foolish mistake, but I certainly hadn't felt any sudden skips, so I don't believe it had moved at all.
This happened because when putting the bike back together, I had played with the eccentric to get it to the right tension (using the iPod Touch App), and when doing so I only ever bother to half-tighten one of the four bolts on the eccentric before checking the tension each time (it normally takes a few attempts before getting it right). When I'd got the tension right, I forgot to go back and tighten the eccentric properly because I was in a hurry. The eccentric had therefore quickly moved into the loosest position.
I was a bit surprised that apart from losing the feeling of direct connectedness between the cranks, there didn't seem to be any other effects of having run the belt totally loose. After I tensioned it properly, it performed flawlessly for another 800 km of touring. Even so, I don't plan to run this experiment again, and I'll try to keep the belt at the advised tension, ignoring the fact that this might be robbing us of 1 or 2 watts of power.
Last edited by Chris_W; 03-02-13 at 04:22 PM.