Wild plate-solve get sync'ed, which sinks my mount

[ChrisR Oz]

Hi Robin,

Using the sequencer to acquire Mosaics with dual exposures for a HDR Composition. Worked well, except once, when the plate-solve was way off (>15 degrees out). Despite such an unlikely error, SC requested the sync to the mount anyway. The next goto then crashed my mount, with the OTA dipping below the horizon. Seems the iOptron CEM120 is short on hard stops. Some sequences attached.

A second plate-solve to the same spot worked fine. But this has not been a once off. I have seen it now 3 times on Gotos. In all cases a repeat exposure and plate-solve works fine (still with shorten exposure for plate-solve option). My FOV is about 59' x 38', so nearly a full degree.

So, my issue is that this just gets sync'ed without checks, which creates a huge problem. The sync needs to satisfy a plausibility test. For a system that has been aligned, you don't expect errors bigger than about 1 degree or so.

Could you add an error check to the options (currently none) for the "plate-solve, sync and re-goto" option? An entry field of degrees, above which SC throws an error and does NOT sync would be good.

An additional option tick-box could be to "on error retry an exposure and plate-solve" if the error tests fails. This would mitigate against a sequence simply failing and stopping.

Cheer, Chris.

[admin]

Hi Chris,

OK, interesting one... Normally the solve and sync procedure if you run it from a button press will prompt you to confirm if the offset is considered 'large' (more than 5 degrees). However, that is suppressed when run from the sequencer, since otherwise you risk a sequence sitting there for hours doing nothing under clear skies waiting for a button press.

There are a couple of things to think about here...

* what should the behaviour be in the sequencer - probably an error not a prompt - is the 5 degree limit OK for that without configuration?

* why are we getting plate solve results that are so far off? Were you using SharpCap's internal plate solving tool, in which case I should be looking into that too!

cheers,

Robin

[ChrisR Oz]

Hi Robin,

Yes, there are two issues. Firstly, the internal plate-solve that I am using, while fast and generally very reliable, can get it seriously wrong occasionally. And secondly, the use of it in the sequencer needs some reality checks and retry options.

The reality checks I would envisage are: (i) do not Sync if beyond a certain error, which the user can specify, and (ii) have an option for retries, either as a simple one-off check-box or as a retry count. I think 5 degrees is still way too large, for an error check. Even before my first sync, my mount is seldom more than 0.7 degrees off. After that, seldom more than 0.3 degrees off. I would set an error bound of about 1 degree. The retry would work most of the time, it seems from experience. So the sequence is unlikely to fail completely.

A simple retry will probably work most of the time. Alternatively, you could nudge the exposure up (gain down) from the fast/high gain for plate-solving for the retries to help reduce noise.

Cheers, Chris.

[admin]

Hi Chris,

if you turn on the option for additional logging of plate solving then it will record a snapshot image and detailed log every time the SharpCap plate solving engine runs - this would allow us to catch the instances when the solution is considerably out and I could investigate why that happens. I don't expect to see that sort of thing - as part of testing I ran the plate solve engine many thousands of times on the whole sky with random patterns of stars to make sure that it didn't find false matches with any significant chance. I guess there must be something slightly different about the circumstances you are encountering...

I think I will stick with the 5 degree limit, but adjust the code so that the sync fails under the sequencer if the distance is beyond the limit. You can build retry into the sequencer (there are various retry and error handling steps), so that is handled if you want it. The reason for a larger limit is that when you swap sides of the pier, the offset will typically be roughly twice any cone error that you have. That could easily be 1 or more degrees, meaning that a 1 degree limit would keep prompting people in those situations.

cheers,

Robin