Some more data on this issue. I wrote a test program that used the ASCOM driver to do the following loop:
Set Filter position (rotating through RGBL)
Wait 10 seconds
Verify filter position
Wait 5 seconds
After 350 cycles of this sequence there was no failure. This test seems to preclude a mechanical issue and points to a ASCOM/SGP interaction problem. I suspect two things:
1. SGP is polling the ASCOM driver for position and it fails (perhaps driver does not respond)
2. SGP fails to note the problem and timeout.
If these suspicions are true, #1 indicates a driver problem or an SGP polling issue. #2 indicates an SGP problem.
If you can tell me more about SGP's interaction with the ASCOM filter driver, I can try to reproduce the problem with my test program. This failure happens most often on the B to L filter change which is the longest rotation of the filter. My filters are LRBGHa in that order. So in a RGBL rotation, the BL transition is more than one filter rotation which suggests a timing issue.