A Fully Automated Imaging Observatory - WIP

[martin_h]

I also purchased the full version PinPoint 6 because I was hedging my bets between Maxim/CCD Commander and Sequence Generator Pro (PP 6 will work with both programs)I use both functions ie an accurate solve with coordinates, but if it fails to solve it will attempt to blind solve as a fall back.

[steppenwolf]

Hi Martin,

Not completely sure of what you are saying here, just to clarify:-

1. are you saying that both CCD Commander/MaxIm DL and Sequence Generator Pro will automatically carry out a blind solve if the PP6 solve fails?

2. if the answer to 1. above is yes, do you happen to know if, following a blind solve, the system will then carry out a PP6 solve using the updated coordinates?

[martin_h]

Certainly in SGP this is the case...from the help file....... 

• Blind Failover:  This option will allow the Astrometry.NET blind solver to function as a backup plate solver to the primary solver (Pinpoint or Elbrus).  Since the local (primary) solvers can fail under a variety of circumstances, it is ideal to have this option checked so you can keep a sequence on tack (especially if you rely on unattended meridian flips or auto centering routines).  You must have an Internet connection to use this feature!

[steppenwolf]

Thanks Martin - I am hoping that long-term I haven't jumped in the wrong boat with CCD Commander rather than SGPro but, time will tell! I would really like that feature but it's not enough to pull me from the CCD Commander/MaxIm DL combination.

Maybe I can write a VBScript to do it if CCD Commander can't ........

[martin_h]

I'm not 100% in the SGP camp yet, Maxim/CCD commander still have the edge I think. CCD/Maxim weather reporting is much more flexible at the moment. And the other night SGP shut down my sequence, parked the scope and shut the roof because a cloud passed over just as the guider was dithering so it couldn't reacquire the star and settle, took me a while to sort out what had gone wrong there!

[steppenwolf]

As usual, Martin, there is no 'one size solution fits all', but I really want my solution to be CCD C ! The trouble is, there are so many variables in this game!

[tdnicholls]

The lack of 'one size' is  reason I I built most of my own control hardware and wrote the main control software for my automated setup, I still use SGP but hook into the logs and custom sensor control.Just in the process of finishing off a bespoke ASCOM safety driver that hooks into the AAG as well as UPS, humidity monitoring etc. that will fire an unsafe if more than just the weather turns.

I have looked a CCD C, APC and a few others but SGP with my own software offers me the best option personally, the lack of pause for weather and target selection looks to like it may be addressed on the 2.5 release - do realise that producing custom  hardware/software is not something available to all.  

Trevor

[steppenwolf]

Thanks for your input, Trevor, I can do hardware and electronics (see my aperture controller post) but my software production days ended with writing my own order processing system in FoxPro and Windows 98!!

I'd like to get to grips with VBScript but truth is, what I really, REALLY want to do is just get out there and start imaging again!!

[tdnicholls]

Totally agree, but weather has been so dire recently the only choice seems to be building hardware and creating the software.  Must admit the electronics side is rusty but OK with the software as that is the day job though less VB and more c#

[Geordie mc]

Hoping to get started with a Pulsar obsy next summer. Your thread has been very helpful, so a huge thank you. I have Maxim and SGP but at the moment prefer the Maxim option. Full automation sounds way too complex for me. Think I would be happy with the dome tracking the scope and getting a weather warning to go and close everything manually. I have no electronics or software writing skills at all! The install will be in Charente, France, so not exactly handy for getting on site help from Gary. I will be happy if I can get a 3 or 4 hour session started manually, leave it to run through (including meridian flip) and then go back and shut everything down again manually. Maybe add more automation laterals my time allows to problem solve. It's taken me a year just to get a good setup with Maxim and Pinpoint LE working reliably. Still in UK so nights to practice have been very limited. Would be very interested in seeing your ascom settings using POTH. Thanks again.

[steppenwolf]

I am really pleased that my experiences of dome automation have been of help - I too have had some great help from others along the way.

I would be happy with the dome tracking the scope and getting a weather warning to go and close everything manually.

Full automation is absolutely not required to achieve a decent boost in your imaging opportunities and your requirements outlined here will allow such a boost.

Would be very interested in seeing your ascom settings using POTH.

Here is a screen dump of the POTH settings (The Lat. and Long. are fictitious) but if I can assist any further, just let me know what you need. I am using a Mesu 200 mount with a pier slightly offset from centre. How I arrived at those figures is shown on page 3 post #51, you will have to use your own measurements, of course and getting these right is crucial to ensure accurate positioning of the dome slit.

[Geordie mc]

Thanks Steve. That's great. Much appreciated. Hope all is going well with your set up.

John

[steppenwolf]

ADDITIONAL NOTE 10 - Plate Solve/Focus Hiccup but now sorted!

Plate Solve

A couple of weeks away and poor weather have put automation implementation on hold but I've now got back into it. However, at the same time as implementing PinPoint (full version) I also upgraded to MaxIm DL V6 as this has a couple of potentially desirable features added since V5. Unfortunately though, this exposed a strange bug where the image capture feature hung when changing filter and starting a plate solve 'action' - CCD Commander would issue an error message that just stuck in a loop. The 'action' starts but when it gets to 'Setting Filter to Ha', it just sits there and then I get the following error message:-

'An action cannot be completed because a component (MaxIm DL Pro 6 - CCD image 4) is not responding. Choose "Switch To" to activate the component and correct the problem.

Switch To                 Retry          (Cancel greyed out)

CCD Commander has a 'feature' that by default forces a dummy image to be captured prior to changing the filter as this apparently adds some stability in certain installations but in my case this caused a problem Disabling this option improved reliability for not 100% so I reverted back to my trusty MaxIm DL V5 and the problem went away!

Focus

Next up was focus after a plate solve using FocusMax and this worked most of the time but sometimes failed with narrowband filters - this is just a bit of fine tuning for exposure values and flux densities but a chance discussion with Sara Wager revealed that she only focuses using her Luminance filter and uses offsets for the other filters (this allows for filters not being truly parfocal within a particular optical system and in my case, I have different makes for wideband and narrowband). Now, I had considered using this method but was concerned about backlash in the focuser drive gearbox but as Sara uses it very successfully, it was worth a try. Last night was fairly clear so I had an opportunity to determine the required offsets and build them in to the system. I used the following work-flow that compensates for changes in temperature during the calibration process:-

1. Autofocus using L filter to check that all is working well and check focus on a bright star with a Bahtinov mask noting focus position
 
2. Use CCD Commander (CCD C) to change filter to Ha and take an image with Bahtinov mask in place
 
3. Use ‘Jog’ in FocusMax to adjust focus until perfect and then issue 200 steps out followed by 200 steps in and check focus with mask noting focus position – still perfect in my case
 
4. Enter offset in CCD C
 
5. Select 'L' filter using MaxIm DL, use Jog to return to original ‘L’ position, then issue 200 steps out followed by 200 steps in and check focus with mask, adjust if necessary noting new focus position if required
 
6. Repeat from 2. for each filter entering each offset with reference to the current ‘L’ position. After a short while, the L position remained constant so the temperature had stabilised
 
7. Use CCD C to take an image of a bright star using each filter in turn using the Bahtinov mask returning to the 'L' position first to simulate an ‘L’ autofocus and make sure each image is correctly focused – which they were
 
As a real bonus, I discovered that plate solves have now become completely reliable using the 'L' channel only.

I have one last, possible task to consider with regard to focus and that is to calibrate automatic compensation for temperature shift but as I will be re-focusing automatically every 20 minutes, I may just miss this one out!

Auto Acquire Guide Star

I rarely have an issue with auto-guiding using MaxIm DL but last night I did under CCD Commander control. It turned out to be a trivial fault caused by the first guide star capture reverting to requiring a 'Simple Dark Frame' by default but I have now switched this default off and re-saved my configuration file - hopefully this will ensure that the default is disabled in the future using the current configuration.

So, the fun (and a certain amount of frustration!) continues but the end is now in sight ........ maybe .......perhaps ........possibly .........hopefully ......... :grin:

[steppenwolf]

ADDITIONAL NOTE 11 - My first Semi-Fully Automated imaging session completed successfully - nearly there!

Why only 'Semi'? The imaging session started after the object had passed the meridian so I didn't have to complete a meridian flip and guide star re-acquire! Now, I know that the mount will successfully complete a meridian flip and during the most recent session, the system successfully re-acquired its guide star every 30 minutes following a re-focus/plate-solve but so far I haven't carried out a meridian flip and guide star re-acquire at the same time ......

So where are we with the project?

1. The system checks that it is dark and dry and if so,

2. The Dome aperture opens, the CCD camera is cooled in three stages to -20° C

3. The Mesu 200 mount un-parks automatically, slews to the object chosen for the session

4. At the end of the slew, the dome is 'coupled' to the mount and rotates to align with the telescope

5. A plate solve image is captured, the mount is synchronised with Cartes du Ciel and a fine adjustment (if required) is made to acquire the correct framing of the object

6. A suitable mag. 6 star is chosen by PinPoint Astrometry and the telescope slews to the test star

7. The camera selects the Luminance filter and carries out an auto-focus

8. The mount slews back to the chosen object, plate-solves and fine-tunes the pointing if required

9. The L filter is swapped for the Ha filter and the previously computed offset applied

10. A suitable guide star is chosen, autoguiding starts and the first of three 600 second exposures is captured

11. After the third image has been captured, the telescope slews to the focus test star, selects the L filter and carries out an auto-focus

12. The mount slews back to the chosen object, plate-solves and fine-tunes the pointing if required

13. The L filter is swapped for the Ha filter and the previously computed offset applied

14. The guide star is re-acquired and autoguiding restarts

15. The first of the next set of three 600 second exposures is captured

16. This is continued until all the images have been captured (in my case 10 sets of 3 images - 30 in total but see later for the failures)

17. After all the images have been captured, the dome aperture is closed and the dome is 'de-coupled' from the mount to stop it following the mount

18. The mount is parked and then the dome is parked as well

19. The camera is warmed up to ambient temperature in 3 stages

The following image, although not a prize-winner, was captured while I slept comfortably in bed but in truth, it was with some relief that I looked outside in the wee small hours to see the dome was indeed parked and closed!

So what's next? Well, the last 6 exposures were rejected because of cloud/mist cover as I have yet to calibrate and enable the cloud sensor part of the weather system so this final weather calibration must be carried out soon. Next, I must test the meridian flip/guide star re-acquire system. After that, I will have achieved my original goals bar the final test of a close-down through bad weather followed by an automatic re-start but, you know what?, although a nice feature to have, I'll be happy with what I have already achieved!

[steppenwolf]

Additional Note 12 - Unfortunately the autoguiding after a meridian flip failed to work correctly!

As a test, I arranged to capture some OIII data of the Eastern Veil to add to the previous Ha image, about half an hour before a meridian flip was required. The capture of the first three frames went without incident and the mount flipped right on cue having plate-solved to determine its last position. Having flipped, a second plate-solve was carried out followed by a tiny pointing adjustment, the camera was re-focused and the guide star re-acquired. So far so good! Sadly, as guiding re-started, the guide star moved steadily out of the field of view until it was lost! The system had failed to reverse the control movements following the meridian flip.......

There are so many ways of initialising and controlling a meridian flip and its aftermath and what I believe is happening here is that there are too many systems vying for control:-

1. The Mount Controller

2. MaxIm DL

3. CCD Commander

The key is to ensure that I pass all the control on to CCD Commander but I await a clear sky to determine if the changes I have made will resolve the issue. Meanwhile, although 3 subframes of OIII is far too little, I couldn't resist adding them to my earlier Ha data so here is the rather noisy result!

[Andrew INT]

This is such a brilliantly helpful thread Steve et al.

A couple of years ago I had just started to plan my own automated observatory from scratch with a £25k budget. Unfortunately fate stepped in and made my job redundant, which was a bit poo. Anyway, Ive used my time off work to gain a degree in astronomy. Im now back in work and hope to rekindle my dream project in a couple of years time once I get my house back in order!

This thread is great since it adds detail to my somewhat vague project planning!

Thanks all for contributing! This is a great community wd have here!

Kind regards, Andrew

[steppenwolf]

I'm delighted that the thread has been of some use, Andrew and I wish you luck with your own project - mine has been done on a much smaller budget than yours but it can be done I feel sure, despite the set-backs and problems that I have 'enjoyed' and detailed on the way!

I've had some really helpful assistance from the knowledgeable people on SGL and you are right - this is a great community .......

[Andrew INT]

My budget is for everything; dome, mount, telescope etc. Those three will account for the majority of the cost!

[steppenwolf]

Additional Note 13 - Unfortunately the Project is currently stalled as I battle with restarting autoguiding after a meridian flip!

I've spent far too many hours trying to resolve this issue and have tired numerous permutations of the settings, asked on all the relevant forums and Yahoo groups all to no avail!

Last night's clear sky (rather squandered on testing!) gave me a further opportunity to carry out more testing and trying to reduce the time for each test led to a strange issue - if the meridian overlap was too small, the mount would just stop, not flip!! So, I am back to a 03:45:00 overlap.

I currently control both the mount and the dome via POTH but today, I have decided to try and take POTH out of the equation thus putting control more directly in the hands of MaxIm DL so I guess it *might* mean that some of the guiding permutations in MaxIm DL may have some more relevance.This seems to work very well during daytime meridian flip tests but only time will tell if it makes any difference under night skies.

Sadly, if I don't find answers to this issue shortly, I will have to start looking into using a different orchestration program or just accept that after a meridian flip, manual intervention will be required!

[steppenwolf]

My budget is for everything; dome, mount, telescope etc. Those three will account for the majority of the cost!

It's horrendous how the costs mount up isn't it? :grin: I have to say though that despite the problems that I have slowly knocked off one by one, I have had a blast with this project. The thrill of resolving seemingly insurmountable problems is immense although I am rather losing the will over this current autoguiding issue!

[swag72]

I hope you get to the bottom of this issue soon Steve - Battling myself and it really is soul destroying......... Good luck

[steppenwolf]

Thanks, Sara, we do put ourselves through the mill don't we? :grin:

[Andrew INT]

Isn't it a mesu200 mount you have? I though they went a couple of hours past the meridian before flipping.

[Andrew INT]

Im luck that in work (I write plans to mitigate mass fatalities), I have also been given a big project to manage which involves relocating a control room into police HQ. Im slowly picking up some skills which I can apply to my own future observatory project.

[Sp@ce_d]

Additional Note 13 - Unfortunately the Project is currently stalled as I battle with restarting autoguiding after a meridian flip!

I've spent far too many hours trying to resolve this issue and have tired numerous permutations of the settings, asked on all the relevant forums and Yahoo groups all to no avail!

Last night's clear sky (rather squandered on testing!) gave me a further opportunity to carry out more testing and trying to reduce the time for each test led to a strange issue - if the meridian overlap was too small, the mount would just stop, not flip!! So, I am back to a 03:45:00 overlap.

I currently control both the mount and the dome via POTH but today, I have decided to try and take POTH out of the equation thus putting control more directly in the hands of MaxIm DL so I guess it *might* mean that some of the guiding permutations in MaxIm DL may have some more relevance.This seems to work very well during daytime meridian flip tests but only time will tell if it makes any difference under night skies.

Sadly, if I don't find answers to this issue shortly, I will have to start looking into using a different orchestration program or just accept that after a meridian flip, manual intervention will be required!

Well, if you do have to look else where Steve, I have to say this has been well taken care of for me with ACP... I cannot recommend it high enough really (I know we've discussed before and oh yes those costs do mount up!!). It is so well documented and the support is excellent. I've worked for big companies in IT and worked for big customers of IT. So I've seen support from both sides. The support for ACP is at that level in comparison, if not higher. There are a multitude of components to make work together in what we are doing.

Now, having said that I've noticed a couple of times recently that my meridian flip has stalled. ACP thinks its slewing but the mount hasn't moved. I haven't had the time to go back through all the logs & see if theres anything.. I certainly noticed nothing at the time. Luckily it was when i was watching, I think it may have been on the same target. I was so desperate to maximise clear sky I just resubmitted the plan & off it went happy. If this has happened before during the night when I'm asleep then I've been unaware. ACP must have sorted it out on it's own.. there must be a timeout. Certainly as far as guiding goes ACP will restart as necessary.