Platesolve works then doesn't?

Questions, support and discussions
jst200
Posts: 27
Joined: 20 Mar 2018, 12:43

Platesolve works then doesn't?

Post by jst200 » 21 Sep 2019, 18:53

Hi Han,

I wasn't sure whether t post this here or in the APT forum?

Anyway, I was trying to use the latest version of ASTAP last night to help with a platesolving issue I have with APT. All seemed to be going well and then the following happened...
  1. APT started the Auto Meridian Flip process, which it does by Solving the location before the flip. ASTAP solved this location fine (in 2secas always).
  2. The flip then occurred. Again ASTAP solved the flip location in 2sec and showed the flip to have been 8192 pixels off target
  3. APT started a GoTo++ process to home in on the target and slewed to a more accurate location. ASTAP would then not solve this new location, returning the error "-1:Maximum search limit exceeded'".
Exactly the same thing happened twice. SO I switched back to using PS2 and it worked fine.

Here is an edited extract of the APT log showing ASTAP successfully doing the pre-flip Solve, the 1st post-flip solve then failing on the 2nd post-flip solve:

2019/09/20 21:58:28 (UT 2019/09/20 20:58:28) Op Starting Automated Meridian Flip.
2019/09/20 21:58:28 (UT 2019/09/20 20:58:28) DBG Alt = 81:49:30; Az = 173:48:13; RA = 20:52:46; Dec = 45:12:48
...
2019/09/20 21:58:29 (UT 2019/09/20 20:58:29) Op Determining the exact position before Automated Meridian Flip...
...

2019/09/20 21:47:21 (UT 2019/09/20 20:47:21) Op Exp 1/1 : Bin(1x1), Gain(4000), Duration(5") CCD T(-9°C)
2019/09/20 21:58:39 (UT 2019/09/20 20:58:39) Op PointCraft: Image solved in 3s using 999 stars
2019/09/20 21:58:39 (UT 2019/09/20 20:58:39) Op PointCraft: Detected FOV size (arc-minutes) - 52.25 x 35.04
2019/09/20 21:58:39 (UT 2019/09/20 20:58:39) DBG PointCraft: Found Ra 20.6906483395, Dec 45.2777679703
2019/09/20 21:58:39 (UT 2019/09/20 20:58:39) Op Target locked before flip. RA=20:41:26.33, Dec=45:16:39.96.
...
2019/09/20 22:08:28 (UT 2019/09/20 21:08:28) Op Starting flip via GoTo.
2019/09/20 22:08:29 (UT 2019/09/20 21:08:29) DBG Generated coordinates for APT Dithering.
2019/09/20 22:09:25 (UT 2019/09/20 21:09:25) Op Flip move finished.
2019/09/20 22:09:25 (UT 2019/09/20 21:09:25) Op Settling after flip move.
2019/09/20 22:09:40 (UT 2019/09/20 21:09:40) Op Target Re-Centering started.
...
2019/09/20 22:09:57 (UT 2019/09/20 21:09:57) Op Exp 1/1 : Bin(1x1), Gain(400), Duration(5") CCD T(-15°C)
...
2019/09/20 22:10:02 (UT 2019/09/20 21:10:02) Op PointCraft: Image solved in 2s using 999 stars
2019/09/20 22:10:02 (UT 2019/09/20 21:10:02) Op PointCraft: Detected FOV size (arc-minutes) - 52.24 x 35.03
2019/09/20 22:10:02 (UT 2019/09/20 21:10:02) DBG PointCraft: Found Ra 20.5005614428, Dec 45.3460903526
2019/09/20 22:10:02 (UT 2019/09/20 21:10:02) Op PointCraft: Step 1/5. Target is in 8910 pixels.
2019/09/20 22:10:02 (UT 2019/09/20 21:10:02) DBG RelativeGoTo - SRa:20.701898706, SDec:45.348989362, TRa:20.690648340, TDec:45.277767970, RRa:20.500561443, RDec:45.346090353, NRa:20.891985603, NDec:45.280666979
...
2019/09/20 22:10:23 (UT 2019/09/20 21:10:23) Op Exp 1/1 : Bin(1x1), Gain(400), Duration(5") CCD T(-15°C)
2019/09/20 22:10:56 (UT 2019/09/20 21:10:56) OpErr PointCraft: Solving failed. PlateSolve2 returned '-1:Maximum search limit exceeded'
2019/09/20 22:10:56 (UT 2019/09/20 21:10:56) OpErr PointCraft Error: Can't continue.

Are there any ASTAP log files that could help identify why the 3rd solve failed, e.g. couldn't see any stars (i.e. starts movement streaked) or couldn't
match any catalogue stars.

The problem I'm trying to fix is that EQMOD might have a problem working out which side of the pier it is in,during the flip process. So I was wondering if ASTAP might be trying to match stars in the eastern hemisphere, when actually the image is in the western hemisphere and the 180deg search wasn't wide enough.

Any thoughts greatly appreciated!

Thanks, Jim
Last edited by jst200 on 21 Sep 2019, 19:19, edited 3 times in total.

jst200
Posts: 27
Joined: 20 Mar 2018, 12:43

Re: Platesolve works then doesn't?

Post by jst200 » 21 Sep 2019, 19:14

Hi Han,

I've just discovered APT (or perhaps ASTAP?) created an image of the failed solve.

The failed solve was at 22:10:56 and here is a .jpg created at that exact time (AfterFlip_22_10_56.jpg)
AfterFlip_22_10_56.jpg
AfterFlip_22_10_56.jpg (121.49 KiB) Viewed 1696 times
I've now tried solving this image in ASTSP outside of ATP and it still doesn't solve. Weird, because I can't see anything obviously wrong with it.

The large star in the Image is Deneb.

Cheers, Jim

jst200
Posts: 27
Joined: 20 Mar 2018, 12:43

Re: Platesolve works then doesn't?

Post by jst200 » 21 Sep 2019, 19:41

Sorry! A further update...

I have just changed the search criteria within ASTAP to Search Radius of 180 and the image solved. Which makes APT invoked ASTAP failing to solve even more strange.

Does APT set a search criteria of 180 or does it use the setting already set in ASTAP?

Cheers, Jim

han.k
Posts: 289
Joined: 23 Nov 2017, 15:04
Location: The Netherlands
Contact:

Re: Platesolve works then doesn't?

Post by han.k » 21 Sep 2019, 21:37

Hello Jim,

The sarch range is set by APT but is is pretty large so that is not th problem.

The problem is the image itself. The stars show a severe tracking problem. Normally you need 10 or 20 seconds before the tracking stabilise. The gears need to settle after a slew. ASTAP will interpretate these stars as galaxies or double stars. So only a fraction of the stars are detected and the high compression doesn't help either. Is there a mount settle time you could set or adjust? If so set it a little longer. For my mount HEQ5 is differs but 15 seconds settle time works for 99% of the case

Cheers, Han
Untitled.jpg
Untitled.jpg (85.94 KiB) Viewed 1688 times

han.k
Posts: 289
Joined: 23 Nov 2017, 15:04
Location: The Netherlands
Contact:

Re: Platesolve works then doesn't?

Post by han.k » 22 Sep 2019, 12:14

Hello Jim,

I looked it up and assume you should increase the APT setting:
GoTo++ Pause
Defines how many seconds to wait after GoTo++ before taking the next image for position confirmation. Some mounts can require time to settle down after a movement.
From your experience I can only conclude that ASTAP is more sensitive to tracking problems the PlateSolve2

jst200
Posts: 27
Joined: 20 Mar 2018, 12:43

Re: Platesolve works then doesn't?

Post by jst200 » 26 Sep 2019, 10:58

Hi Han,

First an apology... sorry for not replying sooner. I was waiting for an email notification from the forum, that someone had replied, but didn't get one and I also forgot to check. I'm sure the fault was mine, as I hadn't got my forum notifications set properly. I've set all notifications types to email now, so hopefully it won't happen again.

Re: The stars show a severe tracking problem. Normally you need 10 or 20 seconds before the tracking stabilise.

Yes, I can see that now. I think I had forgotten about the GoTo++ setting, with all the other Pause options around the Auto Meridian Flip option. Also ASTAP seemed happy with non Auto Meridian Flip GoTo++'s, so assumed it was an AMF problem. Anyway I have an HEQ5 Pro with a Rowan Belt mode, so the same (or much the same) as you and I have my GoTo++ Pause set to 10s. I will up it to 20s to see what happens. Perhaps it's something to do with backlash.

Re: The problem is the image itself.

But if that were the full problem, why will the image solve now using ASTAP directly? Checking the image with the "Test button to show tetrahedrons" it solves in 0,9 sec and shows a good collection of tetrahedrons. I was going to capture a screen capture, but I'll leave you to do it, unless it doesn't solve for you but does me!

Re: ...and the high compression doesn't help either.

Ah! The image I've posted is the .jpg image APT creates after a failed Solve (I think - see the extract from the APT log below). As I was using APT v5.75.4 it uses (or is supposed to use) the the original ccd taken .fit file. Unfortunately, I don't have that anymore, as it is overwritten each time GoTo++ does a move. Could it be that the .fit file fails to solve, but the.jpg, converted from the .fit file solves fine. So in fact the compression is helping!

Re: your screen capture

Always on the lookout for analysis tools, I was wondering how you created the image with the yellow boxes?

Thanks for your help, as always, Han.

Cheers, Jim

Extract from APT log:

2019/09/20 22:08:28 (UT 2019/09/20 21:08:28) Op Starting flip via GoTo.
...
2019/09/20 22:09:25 (UT 2019/09/20 21:09:25) Op Flip move finished.
2019/09/20 22:09:25 (UT 2019/09/20 21:09:25) Op Settling after flip move.
2019/09/20 22:09:40 (UT 2019/09/20 21:09:40) Op Target Re-Centering started.
...
2019/09/20 22:09:57 (UT 2019/09/20 21:09:57) Op Exp 1/1 : Bin(1x1), Gain(400), Duration(5") CCD T(-15°C)
...
2019/09/20 22:09:58 (UT 2019/09/20 21:09:58) DBG File saved: D:\Documents and Settings\All Users\Documents\APT\CameraCCD_1\Focusing_CCD.fit
'''
2019/09/20 22:10:02 (UT 2019/09/20 21:10:02) Op PointCraft: Image solved in 2s using 999 stars
2019/09/20 22:10:02 (UT 2019/09/20 21:10:02) Op PointCraft: Detected FOV size (arc-minutes) - 52.24 x 35.03
2019/09/20 22:10:02 (UT 2019/09/20 21:10:02) DBG PointCraft: Found Ra 20.5005614428, Dec 45.3460903526
2019/09/20 22:10:02 (UT 2019/09/20 21:10:02) Op PointCraft: Step 1/5. Target is in 8910 pixels.
2019/09/20 22:10:02 (UT 2019/09/20 21:10:02) DBG RelativeGoTo - SRa:20.701898706, SDec:45.348989362, TRa:20.690648340, TDec:45.277767970, RRa:20.500561443, RDec:45.346090353, NRa:20.891985603, NDec:45.280666979
...
2019/09/20 22:10:23 (UT 2019/09/20 21:10:23) Op Exp 1/1 : Bin(1x1), Gain(400), Duration(5") CCD T(-15°C)
...
2019/09/20 22:10:24 (UT 2019/09/20 21:10:24) DBG File saved: D:\Documents and Settings\All Users\Documents\APT\CameraCCD_1\Focusing_CCD.fit
...
2019/09/20 22:10:56 (UT 2019/09/20 21:10:56) OpErr PointCraft: Solving failed. PlateSolve2 returned '-1:Maximum search limit exceeded'
2019/09/20 22:10:56 (UT 2019/09/20 21:10:56) OpErr PointCraft Error: Can't continue.
2019/09/20 22:10:56 (UT 2019/09/20 21:10:56) DBG D:\Documents and Settings\All Users\Documents\APT\CameraCCD_1\TemporaryStorage\AfterFlip_22_10_56.jpg
2019/09/20 22:10:56 (UT 2019/09/20 21:10:56) OpErr Can't re-center the exact position after flip!
2019/09/20 22:10:56 (UT 2019/09/20 21:10:56) OpErr Can't complete an Automated Meridian Flip! Stopping all actions...

han.k
Posts: 289
Joined: 23 Nov 2017, 15:04
Location: The Netherlands
Contact:

Re: Platesolve works then doesn't?

Post by han.k » 26 Sep 2019, 13:50

As soon APT supplies raw or unstretched JPG with a low low compression factor images it will get better.

There are three tools indicating what's going on:

1) CCD inspector in viewer under tools menu. This will show which stars are recognized and which not.
2) In tab alignment there is a check mark "show extended solver log" This will show all steps but also mark in the viewer the detected stars RED and the database stars used YELLOW. This is the most useful indication.
3) In tab alignment, "Show tetrahedrons".

Regards, Han

jst200
Posts: 27
Joined: 20 Mar 2018, 12:43

Re: Platesolve works then doesn't?

Post by jst200 » 26 Sep 2019, 14:02

Hi Han,

Strange. I'm still not getting notifications of your posts/replies :?:

Anyway, Thanks for the tools pointer. :D

As of v3.75.4 ASTAP should be receiving .fits files for solving. Hence my question; could a .fit file fail to solve, but a .jpg file created from the .fit file solve OK? That's what appears to be happening here.

Cheers, Jim

han.k
Posts: 289
Joined: 23 Nov 2017, 15:04
Location: The Netherlands
Contact:

Re: Platesolve works then doesn't?

Post by han.k » 26 Sep 2019, 15:28

I have to test this latest version of APT. What could happen is that ASTAP reads from the FITS header different parameters (size, position) then from the PlateSolve2 command line. The FITS header data has priority of parameter data.

I will download the latest APT.

Regards,Han

han.k
Posts: 289
Joined: 23 Nov 2017, 15:04
Location: The Netherlands
Contact:

Re: Platesolve works then doesn't?

Post by han.k » 26 Sep 2019, 16:12

I did a test. First it didn't solve, because I had set the the values in APT's "Object calculator wrong". The manual entered image dimensions in Pointcraft are ignored. After that it solves fine.

You could test it by loading ImagetoSolve.FIT and check the reported image dimensions

So for image dimensions it takes the dimension from the FITS file.
For the estimated position it takes the position from Pointcraft. Not from the FITS header.

Posted an remark at the APT forum. http://aptforum.com/phpbb/viewtopic.php?f=24&t=2816

Post Reply