Meridian flip

[Gary01978]

Hi,

I'm looking to use sequencer to perform meridian flip, I assume someone has successfully got this to work and instead of trying to reinvent anything I hope someone could share what works for them I would be great full in avoiding wasting hours of clear skies testing stuff.

I am controlling an eq6 over EQMOD and would be hoping it to work while live stacking (which I prefer to use for imaging)

Otherwise could you please advise rough sequence to try?
I was thinking the following
Stop after -1 degrees of meridian (15min after crossing meridian)
Pause live stacking
Stop guiding
Plate solve and save coordinates
Sync to saved coordinates
Sync to saved coordinates (to try again sometimes first sync isn't exactly centered)
Wait 5s
Start guiding
Wait until guiding is active
Resume live stacking

Thanks,

[admin]

Hi,

unfortunately the meridian flip is a complex thing to get right - particularly because different mounts behave differently in how they deal with the meridian. From memory, for instance, I think that EQMOD will sort of get partly stuck if it is allowed to track past the meridian - it ends up in a position where it will refuse to perform GOTO movements, but you can get it back to the other side of the meridian using the movement buttons... Other mounts (or even other drivers for Skywatcher mounts) behave differently...

Fortunately it's easier for someone like you to set up a meridian flip sequence than it is for me to try to add fully automatic meridian flip support (you only have to deal with the quirks of the mount you are using - I would have to try to cope with all the possibilities...). Basically you need a few different parts in the sequence.

1) Something to stop whatever is going on before the meridian at the right point (just before the meridian, probably)

2) steps to perform the actual flip itself - getting the mount onto the other side of the pier

3) getting back to the same target (maybe this will be handled by 2 above, maybe not)

4) resuming whatever you were doing before the flip (live stacking in your case).

For part one, you can try the 'Stop running these steps when <degrees> from the meridian'. You put the things that are happening before the meridian in the sequence inside this step, and set maybe 1 degree from the meridian which will try to stop about 4 minutes before the mount reaches the meridian (4 minutes = 1 degree of RA). Note the 'try' - this only checks for proximity to the meridian between different sequencer steps or between frames inside the 'capture still frames step'. For this to work nicely, you need to ensure that individual steps in the sequence that runs before the meridian don't take more than the 4 minutes available, otherwise the mount could reach the meridian and stop tracking between checks for meridian proximity. You should stop guiding at this point too.

Part two is best handled (at least as a first try) by the 'Meridian Flip the mount (experimental)' step - that tries to deal with at least some of the quirks of mounts around the meridian that I am aware of. It should get the mount to the other side of the pier and hopefully in the right place. Before that step, you could use a 'plate solve and save co-ordinates' to record the current position of the mount to get back to it after the flip, or you could rely on the 'target' co-ordinates of the mount being set when you did your initial GOTO - those should be preserved.

Step 3 - depending on which approach you used to save the co-ordinates is a 'move to saved co-ordinates' or 'Goto target co-ordinates', followed by a solve+sync to get to exactly the right spot

Step 4 - should just be resume guiding and live stacking.

I would suggest testing anything you build on a target that is just 10 minutes or so from crossing the meridian - doesn't have to be anything interesting, you are really testing that all the moving parts of the sequence do the right thing, and by starting only 10 minutes short of the meridian, the whole test will only take 15 minutes or so.

cheers,

Robin

[Gary01978]

Thanks, I done some testing (with camera and mount simulator), the stop x degrees from meridian always seems to just proceed onto the next step even if the mount co-ordinates are nowhere near the meridian and whether x degrees is positive or negative, maybe its just me or a problem doing it with the simulator.

I tried the "retry if errors" command which works fine (incase it fails platesolve first time), but, it would be good if the sequencer finishes with errors (i.e. after retrying x times it still failed) it would perform another command. I have a python script on my PC which I could potentially get sharpcap to run which will email notify me if the sequencer failed what it needed to do.

[admin]

Hi Gary,

did you have something inside the 'stop x degrees from the meridian' step - you need to put something in there that takes a long time, maybe repeat 100 times of wait 10 minutes or something. The key is that the 'stop at' step type only stops the contained sequence code early if it is still running when the stop condition is reached, it doens't make the stuff inside run for longer to get as far as the stop condition being reached. Hope that makes sense.

I will look into having a bit more in the way of dealing with errors in a single block of code.

cheers,

Robin

[admin]

Hi Gary,

actually, the error handling possibility is already there... Use the 'Ignore Errors' step - although it sounds like it is going to just ignore any errors, it's more flexible than that.

1) You can pick a subroutine to call if an error happens - this subroutine can contain sequencer code to either try to recover from the error or to try to send a notification

2) You can put a 're-raise the last error' block inside the subroutine - that lets the error re-propagate outside the ignore error block. This should be the last step in the error subroutine.

Effectively these give you a 'try ... catch ...' exception handling construction with or without re-throwing the exception.

cheers,

Robin

[nexusjeep]

Hi Gary,
Bit off topic but I used to use EQMod as I have an Avalon M-Uno that has the synscan board so controlled it that way. However, I did have issues with EQMod not playing nice with Windows 11 and other pieces of software i have installed so switched to a different control software called Green Swamp Server GSS for short that is modern code and have found that it works without a hitch so not sure if it would be something you would be interested in it is free the same as EQmod is and works with the same cable.

Cheers
Nick

[Gary01978]

Hi Gary,

actually, the error handling possibility is already there... Use the 'Ignore Errors' step - although it sounds like it is going to just ignore any errors, it's more flexible than that.

1) You can pick a subroutine to call if an error happens - this subroutine can contain sequencer code to either try to recover from the error or to try to send a notification

2) You can put a 're-raise the last error' block inside the subroutine - that lets the error re-propagate outside the ignore error block. This should be the last step in the error subroutine.

Effectively these give you a 'try ... catch ...' exception handling construction with or without re-throwing the exception.

cheers,

Robin

Thanks, regarding the stop at meridian instruction I understand now, I had nothing inside, I assumed it would hang inside a loop until the stop condition was reached, even if there was no operation.

I think I get the ignore errors instruction now, but not 100%, from my understanding, the ignore errors would contain some commands like platesolve then set exposure time then start capture, but if any of them sub steps contained an error it wouldn't do the next sub step then it will leave the ignore errors block and onto the next instruction outside that block (for example send a notification), however, if there was not any errors in the sub steps in that block then wouldn't it leave the ignore errors block after the sub steps are successful completed and this proceed to send the notification anyway?

I also suffered an issue with guiding commands, PHD2 is definitely linked with sharpcap (sharpcap controls dithering etc), guiding was active but when I sent stop guiding command, PHD2 said PAUSED but sharpcap report error "guiding not connected while running step".
I also tried to rerun the routine but only to restart guiding with the start guiding if necessary step and it gives the same error about guiding not connected

Thanks, your time is appreciated

[Gary01978]

Hi Gary,
Bit off topic but I used to use EQMod as I have an Avalon M-Uno that has the synscan board so controlled it that way. However, I did have issues with EQMod not playing nice with Windows 11 and other pieces of software i have installed so switched to a different control software called Green Swamp Server GSS for short that is modern code and have found that it works without a hitch so not sure if it would be something you would be interested in it is free the same as EQmod is and works with the same cable.

Cheers
Nick

Thanks, I don't really have any issues with the EQMOD control itself however but out of curiosity I'll take a look

[admin]

Hi,

you are pretty close on the ignore errors step - it will work through the things inside. On the first error it will give up on the things inside and move to the stuff following the ignore errors step. Alternatively if the things inside work without error then it moves on to the following stuff too.

The one extra thing that you need to consider is that you can specify a subroutine of sequencer steps to call as part of the 'ignore error' step - that is *only* called if an error occurs in one of the sub steps, so you can put your notification 'bad stuff happened' or some other recovery steps inside that subroutine.

On the subject of guiding - did you have a 'monitor guiding application' step at/near the start of your sequence? That's a requirement if you want to do any guiding related steps in the sequence and it ensures that SharpCap is connected properly to PHD2.

cheers,

Robin

[Gary01978]

right I got it, so of course I just need to do something to keep the execution inside the "ignore errors" instruction whether it be with a wait command for a crazy amount of time or what not, that way the proceeding "notification" steps will only be performed if an error occurs (its the only way to leave that block then).

I got you on the monitor guiding thing, I missed that and see it loud and clear now

I'll be sure to post a full working sequence once tested