Plate-solve failure during intermittent clouds

[BucketDave]

I'm writing a sequence to image multiple targets during the night. For each target, i'll do a 'Goto target' followed by a 'plate-solve, synch and Goto'. And then image for an hour or two. Like all good code, if there is an error, I want it to fail gracefully.

If auto-focus fails, i understand that the focusser reverts to the starting position. If the guide star is lost, the image capture will stop and restart. If a plate-solve/synch/goto fails, right after a slew, we appear to be in big trouble. Imagine the following scenario:

There are three targets. Intermittent clouds appear during the imaging of target #1. As I'm using the 'Capture images while Guiding' step, the sequencer waits until PHD2 is guiding again. The clouds stay for an hour and the sequence for target #1 hits the buffers.

Onwards to target #2... The mount slews to target #2 but, in reallity, is 2 degs out. The plate-solve fails (still cloudy). Ten minutes later the clouds clear........

1) If the whole sequence stops when the plate-solve fails, it's a disaster. I lose all the imaging of targets #2 and #3. The worst outcome.
2) If i simply ignore the plate-solve error, i will start imaging an empty patch of sky for a few hours before proceeding to target #3. Not ideal but better than option 1 - at least I image target #3.
3) the only way round this, that I can think of, is to set-up the sequence for each target to include a plate-solve every half hour (and ignore solve errors). A bit wasteful of imaging time but in the event of intermittent clouds, I would eventually get a successful solve, synch and goto. This is a modest improvement on option #2. I'll use some (but not all) of the clear sky available for target #2.

So I was pondering some more elegant solutions:

A) Enhancing the plate-solve command with an optional 'require guiding' checkbox. If the clouds are thin enough for PHD2 to guide, then Sharpcap has a sporting chance of a successful plate-solve. Not guaranteed but might be simple to implement.

Really, you want to keep trying until you get a successful solve (within an overall time limit). Maybe something like:

B) A 'Repeat until successful or 3am' type command. This step might contain a plate-solve/synch/goto followed by a 1 minute delay. It keeps trying until it eventually succeeds. It it hasn't succeeded by 3am, the sequencer can move past the 'image target #2 until 3am' steps and can slew to target #3.

Or is there already a neat way to avoid the failure of a plate-solve/synch/goto?

Regards

Dave

[BucketDave]

Just realised that 'plate-solve and synch with guiding' option shouldn't include the goto (because that would need guiding suspended). So the proposed command structure might look something like this:

Set target name
Goto target
Do these steps until 2am
Monitor guiding
Plate-solve and synch WITH GUIDING
Stop guiding
Goto previously saved co-ordinates
Start guiding
Capture images
Stop guiding

Repeat for the next target....

Regards
Dave

[admin]

Hi Dave,

the trick here is to build up the functionality that you want by combining various steps available in the sequencer. For instance, if there are clouds then the plate solve may fail, but there is a retry block that you can put the plate solve command in that lets you specify how many retries.

So, for instance the following sequence fragment will retry plate solving up to 1000 times with 10s delays between each attempt

Code: Select all

    RETRY ERRORS UP TO 1000 TIMES
        DELAY 10
        MOUNT SOLVEANDSYNC
    END RETRY ERRORS

You can't put the delay and the solve/sync in the other order because then the error in solve and sync would cause the retry to go straight round and try again without ever running the delay.

You can then wrap that code in a time limit block if you want to limit the retries to maybe 60 minutes or whatever value is suitable in your situation.

Hope that helps,

Robin

[BucketDave]

Thanks Robin. I knew i must be missing a trick. I'll go and try that in half an hour.
Dave