Guiding pains

Questions and answers about the best way to get started with Prism including: est hardware to use, software settings, etc..
Post Reply
jason
Posts: 44
Joined: Thu Apr 18, 2019 3:53 am

Tue Aug 27, 2019 12:58 am

Recorded a desktop session of this:

https://www.youtube.com/watch?v=6i8bXHFS_HI

Manual calibration was working earlier in the evening and gave me a perfect trapezoid, automatic never gave a trapezoid.

After calibration, starting would yield a crash of the process and then I would have to restart prism as the device could not be disconnected from.

Later on, not even restarting prism was enough, pretty much any guide action gave a crash somewhere.

Managed to get it working fine in phd2, after calibrating at 0 declination but I've had a ton of phd issues in the past, which I why picked up prism. I've figured more things out (mainly the scale dependence of the pulse guide duration) but still, would love to use prism and integrated software for these tasks.
jason
Posts: 44
Joined: Thu Apr 18, 2019 3:53 am

Mon Sep 02, 2019 6:03 pm

still monkeying with it.

I switched it to full frame guiding, with cover on (daylight hours) trying to duplicate the read exception.

The guide loop seems kind of running, if we allow running to include crashing:
error -> access violation at address 004c9d8e in module prism write of address 006e1174 - this was on automatic star finding. There's no stars so the input could be exceptional. I suspect guiding wouldn't work if there were stars though. Manually selecting the star does seem to work though, I'm not seeing the invalid access/write on manual vs automatic star finding.

I was trying to duplicate the read exception only prism gets in what seems like the ROI / guide window based mode. I think there's bugs in prism for ZWO (including ASCOM) ROI - dovetails with the ROI based star focusing issues). The error is "Process exec after readout error". At this point it guiding cannot be stopped but it isn't running, I have to restart prism. If selecting manually picking the guide star, the star chooser step works fine, but the immediate next frame gives the "Process exec after readout error". What does this error mean? Is it a typo for process exit?

@Hamza Touhami / @Cyril Cavadore please provide some feedback / confirmation on bugs. What do you guys think about an issue tracker, by the way?

I'd love to help you guys debug these problems prism is having... There's very many null pointer or invalid memory accesses in prism + speed boosts.
Hamza Touhami
Site Admin
Posts: 564
Joined: Sat May 20, 2017 6:05 pm
Location: Phoenix,AZ
Contact:

Mon Sep 02, 2019 6:15 pm

Jason

Cyril has actually just finished working on an issue coming from the framing inside the driver of the ZWO cameras, a fix is in place already and an update will be pushed in the next couple of days. This is, potentially, what is causing the exception error.
Hamza Touhami | Site Administrator
Hyperion-Astronomy.com
rottielover
Posts: 128
Joined: Thu Aug 23, 2018 8:22 pm
Location: Missouri, USA
Contact:

Fri Sep 06, 2019 10:00 pm

Hamza Touhami wrote:
Mon Sep 02, 2019 6:15 pm
Jason

Cyril has actually just finished working on an issue coming from the framing inside the driver of the ZWO cameras, a fix is in place already and an update will be pushed in the next couple of days. This is, potentially, what is causing the exception error.
Great news this is going to help me out as well!!!
jason
Posts: 44
Joined: Thu Apr 18, 2019 3:53 am

Sat Sep 07, 2019 11:40 pm

@Hamza Touhami / @Cyril Cavadore great - sounds like the region of interest is coming out wrong and your claim is that this is due to the driver internals of zwo? I find that kind of interesting since I did some ascom work with it (in python) and didn't see an issue - sharpcap et all works in ROI modes but maybe I'm not looking carefully enough at the issue source. Can you briefly mention what your observations are of it and zwo driver versions - and whether or not this is a bug zwo should fix? With knowledge of what is wrong and reproducible demonstration I can help get zwo to fix the problem so you don't have to work around it - zwo can have good turn around times on fixes like this.

Anyway I'm very eager to try out any fix - if correct this could be responsible for all prism's focusing methods being broken for me as well as local star guiding - I can only hope (and even then it could be jumping to conclusion on autofocus). In effect short of using the skymap, I basically cannot do automated shoots with prism. Please let me know if you can make available a test build sooner than later (it's been 1 week since potential fix was mentioned) so we can at least learn of the effect on the issues I have even if other things aren't polished yet. I'm having to lean alot on SGP in the mean time and while it has many parts where prism is currently completely broken for me, it is not without it's own hiccups and instabilities.
jason
Posts: 44
Joined: Thu Apr 18, 2019 3:53 am

Sun Sep 08, 2019 5:59 am

unfortunately even with whole frame guiding, even while the guiding window stays open, it is still crashing (memory access/segfaults) and guiding is completely unusable... meanwhile phd guiding, while having it's own problems works.
Cyril Cavadore
Site Admin
Posts: 74
Joined: Sat May 20, 2017 6:15 pm

Wed Sep 11, 2019 8:41 pm

You can try with a release we have put here

https://www.dropbox.com/s/v7l420sam68sm ... e.zip?dl=0
Cyril Cavadore
Principal Prism Author,
Alcor-System
www.alcor-system.com
jason
Posts: 44
Joined: Thu Apr 18, 2019 3:53 am

Wed Sep 18, 2019 1:29 am

I cannot get a calibration going - I am running a new guide scope OTA (90mm maksutov @940mm FL , which is working alot better on the same camera) but I can only do a drift alignment (via phd2) on my balcony (only viewing southern sky) - so the polar alignment isn't perfect, but it is pretty close. Still working on the process for this.

I can calibrate phd2 interestingly but I haven't been able to achieve a trapezoid in prism and it always rejects the calibration. Generally it isn't finding enough movement in one axis. Is there a better way , maybe a way to make larger movements so it can distinguish the principal axis better relative to the bias/noise in the system?

at this rate I won't be able to test guiding until I go to a dark site next which is the worst time to look into these issues...

Is it sufficient to just write down the xy positions of a star I move myself with manual/hand control or do I need to manipulate the az/el positions by fixed amounts rather than pressing the button till I achieve something that resembles a trapezoid? Hm - the trapezoid is in ra/dec space and I record the xy locations of a star at the trapazoid corners? Hm, need to figure out something to at least provide the manual calibration override.
jason
Posts: 44
Joined: Thu Apr 18, 2019 3:53 am

Sun Oct 13, 2019 8:42 pm

I'm still reliably having issues with guiding, the calibration step in particular.

It seems like it's either related to having a fresh telescope connection or whether or not I have my main zwo camera connected. I've only been able to track it down to those 2 sensitivities.

It seems like it breaks the guiding too, the odd bit is the ROI prism is extracting clearly contains a crisp star, but the ui reports no stars found. Similarly for no star selected / auto mode - it will just say something to the effect of no stars detected or star at 0,0.

It seems like it continues to work after getting it going or calibrated given the above dance (which involves head scratching and hope).
rottielover
Posts: 128
Joined: Thu Aug 23, 2018 8:22 pm
Location: Missouri, USA
Contact:

Mon Oct 14, 2019 6:35 am

I had some of that last night, but with as bright as the moon was I wasn't really surprised, all that extra light bouncing around didn't help anything at all.

I've just switched to the ZWO OAG and eliminated a separate guide scope. Calibration went pretty smoothly, but I did that really quickly after it got dark enough and just before the moon came up over my neighbors house.

You should join us on one of the Saturday night google hangout sessions sometime, if you join the Facebook group there are links all over there, I think I posted the link around here on the forums someplace as well.

I kept messing about with things and after the moon came up I ended up doing the calibration by unchecking the auto-find box and clicking on a star myself. Seemed to work fine and I was able to put it into Full-Frame guide. I slewed a little farther over toward the meridian and it auto-calibrated fine then.

If I recall correctly are you the person that also has a 290 mini for a guide cam? Did you update to the latest version of prism? They improved ZWO support.
jason
Posts: 44
Joined: Thu Apr 18, 2019 3:53 am

Mon Oct 14, 2019 9:52 am

I did the update check before beginning, yes I have a 290mm and 183mm pro. In hyperstar mode I have to use a guide scope for multiminute + exposures on my cem60. Having said that I have an easier time with the guide scope vs the celestron oag in f/7 and f/10 because we're talking about ~2000-3000mm fl with a small fov.

Anyway one important detail I noted while the calibration is working vs isn't is that it seems to just drift in one direction while it is having the calibration issue (due to connected to 2 cameras or a scope connection seemingly not being fresh enough despite controls working). Then when it works, perhaps seconds later, I notice the calibration routine is making the star move up down left right.
rottielover
Posts: 128
Joined: Thu Aug 23, 2018 8:22 pm
Location: Missouri, USA
Contact:

Mon Oct 14, 2019 10:13 am

Did you setup a dark image for the 290? I had that same thing I think Prism might have been trying to calabrate/guide on a hot pixel or something. I can now tell when I'm going to get a good calibration and when I am not based on the numbers I see in the box, I know if I see less than 10 it won't calibrate, but if I start seeing 20's and 60's etc I'm going to get a good one.
jason
Posts: 44
Joined: Thu Apr 18, 2019 3:53 am

Sun Oct 20, 2019 11:33 am

yes, problem happens with and without the dark frame, I mean when you select the pixel and do it manually that shouldn't matter at all anyway except for enhanced centroiding... but really even that should be robust to pixel errors. The part that is very interesting and a key detail I think is that it seems to lose the ability to move the mount in the cross/trapezoid and instead looks pretty much to move in only direction in one dimension (like tracking gets disabled?) I've seen this happen if you don't connect to the telescope at and try and run this but I don't understand why it can happen with the link good and working even from inside prism.

I noticed frequently that when prism starts up it seems to have forgotten the calibration (it makes a popup) but sometimes it recovers it... only thing I can figure is that it depends on the order the cameras are connected inside prism (i'm not changing hardware ports).

Once I had guiding on tonight and working, I connected my 183mm and it seems like that killed the ability to track the star despite it finding it. So weird.

Still reads like some weird hard to understand issue(s) of prism's handling of the driver, best I can tell @Cyril Cavadore

and man prism crashes so many times a night still... possibly while opening and closing camera devices. May be coupled.

Despite the use of delphi, perhaps you can dump core files so you can narrow in on the crash causes more readily? You don't have to distribute debug symbols to make that work.
John Monroe
Posts: 40
Joined: Thu Dec 06, 2018 7:17 pm

Sun Oct 20, 2019 8:51 pm

by Cyril Cavadore » Wed Sep 11, 2019 8:41 pm
You can try with a release we have put here

https://www.dropbox.com/s/v7l420sam68sm ... e.zip?dl=0
Cyril Cavadore

Did this help? I c licked on the link and the file is gone.
Post Reply