...and today it's back to being a bit dodgy (not as bad as Thursday). After switching on, the TrapKAT was hardly responding to the pedal closing and, after retraining, the training data look odd (very narrow range).
Start training, manually press the trigger tight closed, tap a pad, release the trigger, tap a pad. So that's max force range applied. (If you like, even lift the rubber off the sensor for "open".)
Now, start up a MIDI monitor and start gently pressing down on the trigger. There's a little bit of "give" before the TrapKAT starts emitting any events in response to pressure on the trigger. Press all the way down, like when training... and events stop before you get there.
Next, let's try setting the pedal up to control the pressure on the trigger between these two limits. So, loosen the chain so the pedal footplate drops freely and slide the trigger under (I used a bit of paper to prevent the Velcro catching). You should start seeing events as the pedal pressure increases. What we're after here is the pedal hitting the hard stop on the trigger just at the point the TrapKAT stops emitting events - so that'll feel like "fully closed". Once there, it's a bit fiddly, but you need to get the trigger anchored on the Velcro (*) so that the pedal touches down the same. Once the "closed" end is sorted, raise the pedal by pulling the chain back up until events stop again, then tighten it.
(*) I'm starting to suspect it's actually the Velcro that's failing rather than anything else, to be honest. It's as if the trigger is slipping out of its ideal position. But... it doesn't seem that simple.
If you DO NOT train after this, it works pretty well -- apart from the velocity sensing for CHICK events. (And probably splash won't work at all, either.)
However, if you train... you'll again find there's a zone both at the open end and closed end where the pedal moves without events being emitted. This is the bit I don't understand at all. It's as if you need to "over train" the pedal in order to get a proper response.