Re: New NINA plugin for APCC now in beta: Astro-Physics Tools
On Fri, Nov 26, 2021 at 11:55 AM, Ray Gralak wrote:
Can you provide a link to your zipped log files? (Please use the APCC Log zipper to zip up the ASCOM, APCC, APPM, and your PNT files).Sure. The APCC files are large because it was a whole night run. The model was built about 6:30pm local (EST) and the go-to for the slew-to-zenith sometime after 7:30am local (just after an unpark from park 2). I spent the night guided, but guiding was really really good, so I do not think the model was invalid near the target, I think it actually helped.
I also want to share this screen shot. It's a small model made just to test the link with NINA the night before. This is not the problem model. The reason I'm including is to show that the non-perp numbers are not wacky in a whole sky model. The west is still worse than the east for some reason, but nothing like the huge numbers in the dec arc model.
In retrospect I am not really surprised at what happened. I think the dense but tiny model is being applied to other parts of the sky, where it is not valid (really not applicable). In this case I had too few points (if you need 12) east, but I think the west pointing correction was so huge for the slew-to-zenith that it caused a flip (my meridian limits are at meridian).
Assuming I didn't just do something wacky, there seem to be three possible mitigations:
- Human avoids doing that (using a dec arc model for a slew to other parts of the sky)
- APPC somehow notices, and when pointing more than X degrees from the closest point ignores the model
- Provide a "turn off model corrections" function in the API and in NINA's sequence we can turn them off when done with a target.
The second sounds ideal. :)
But... the chances of this being operator inflicted remain high, so I welcome your feedback.