×

INDI Library v2.0.6 is Released (02 Feb 2024)

Bi-monthly release with minor bug fixes and improvements

[Fixed] Plate solving giving wrong results with ANSVR

  • Posts: 269
  • Thank you received: 53
I'm running Kstars 2.9.8 on Window 10 64 bit. Platesolving through Ansvr.
My target was NGC6822 (Barnards Galaxy) at (from KStars):
JNow: 19h 45m 58s -14° 45' 45"
J2000: 19h 44m 56s -14° 48' 23"
I get a successful solve and the mount slews. The next solve says it is within acceptable range.
But if I then take an image and solve at nova.astrometry.net it is around 15 arc-minutes off in RA.
Results from nova.astrometry.net (J2000) are:
Center (RA, Dec): (295.932, -14.862)
Center (RA, hms): 19h 43m 43.610s
Center (Dec, dms): -14° 51' 44.593"
Size: 38.1 x 28.8 arcmin
Radius: 0.398 deg
Pixel scale: 1.96 arcsec/pixel
Orientation: Up is 262 degrees E of N
I suspect the conversion of J2000 returned by nova.astrometry.net to JNOW is in the wrong direction. Perhaps as I'm in the southern hemisphere?
Had the same issue with 2.9.6 the other day which is why I upgraded.

Attaching Kstars log and image solved on nova.astrometry.net
Last edit: 5 years 6 months ago by Ken Self.
5 years 6 months ago #29494
Attachments:

Please Log in or Create an account to join the conversation.

  • Posts: 111
  • Thank you received: 40
Hi kengs,

I believe this is good question for Dustin Lang (the author) of astrometry.net --- either as an github
issue or a astrometry.net forum post.

I am not mistaken, then the astrometry.net web engine uses this generic parameter for solving:
solve-field --downsample 2 --tweak-order 2 --scale-units degwidth --scale-low 0.1 --scale-high 180.0

This is what I get, when I solve the image locally under Linux with the latest astrometry.net (master)

Field 1: solved with index index-4205-07.fits.
Field 1 solved: writing to file /tmp/Barnard.solved to indicate this.
Field: Barnard.jpg
Field center: (RA,Dec) = (295.931792, -14.862427) deg.
Field center: (RA H:M:S, Dec D:M:S) = (19:43:43.630, -14:51:44.739).
Field size: 38.1293 x 28.8327 arcminutes
Field rotation angle: up is -82.5227 degrees E of N
Field parity: pos
Creating new FITS file "Barnard_wcs.fit"...
5 years 6 months ago #29496

Please Log in or Create an account to join the conversation.

  • Posts: 269
  • Thank you received: 53
astrometry.net is doing the right thing. It returns coordinates in J2000 which Ekos converts to JNOW. Its the conversion to JNOW that appears to be wrong so the corrective slew to target is taking the target further away from the centre of the field instead of closer to it.

Added: I also solved locally and got the same result as online.
Last edit: 5 years 6 months ago by Ken Self.
5 years 6 months ago #29510

Please Log in or Create an account to join the conversation.

  • Posts: 1029
  • Thank you received: 301
That may sound stupid and I do not suggest it could be the case in your test, but I need to mention that I had very weird results when I forgot to suspend guiding while plate-solving. Guiding corrections actually offset the mount model. PHD2 has an option to disable guiding while the mount slews, the internal guider does not.

-Eric
5 years 6 months ago #29533

Please Log in or Create an account to join the conversation.

  • Posts: 1029
  • Thank you received: 301
Looking at the log, it seems to me we're dealing with an API change here:
[2018-09-12T21:59:09...][     org.kde.kstars.ekos.align] - "{\"jobs\":[62],\"processing_finished\":\"1\",\"processing_started\":\"1\",\"user\":\"0\",\"user_images\":[]}"
[2018-09-12T21:59:09...][     org.kde.kstars.ekos.align] - "{\"status\":\"success\"}"
[2018-09-12T21:59:09...][     org.kde.kstars.ekos.align] - "{\"dec\":\"-14.816192073\",\"epoch\":\"JNow\",\"orientation\":97.5146272814133,\"parity\":1,\"pixscale\":1.96562592485983,\"ra\":\"296.199887826\",\"radius\":0}"

RA/DEC seem to be returned in JNow here.
-Eric
5 years 6 months ago #29534

Please Log in or Create an account to join the conversation.

  • Posts: 269
  • Thank you received: 53
I was using PHD2 but guiding was off. I re-ran the platesolve under controlled conditions.
5 years 6 months ago #29535

Please Log in or Create an account to join the conversation.

  • Posts: 269
  • Thank you received: 53
As far as I understand it, astrometry.net returns only J2000 coordinates. I've not found any option to return JNOW coordinates.
I tested again today on another target and Ekos consistently puts the target some 15' off centre in RA. I then switched to Astrotortilla using the same engine and it correctly centred the object.

What I suspect now is that the J2000 to JNOW conversion is being applied twice. Astrometry.net returns RA of 19h43m44s as J2000 which converts to JNOW of 19h44m48s which is the value in the log (296.199887826 degrees). A further J2000 to JNOW conversion is then applied to 19h44m48s which results in 19h45m50s as shown in the next part of the log. 1m of RA corresponds to 15 arcminutes which is the error I am seeing.
5 years 6 months ago #29537

Please Log in or Create an account to join the conversation.

  • Posts: 269
  • Thank you received: 53
Looks like JNOW is an ANSVR config option which is why I couldn't find it in the astrometry.net API documentation. I'll see how to switch it off an hopefully then the solves will work.
The following user(s) said Thank You: Eric
5 years 6 months ago #29538

Please Log in or Create an account to join the conversation.

  • Posts: 269
  • Thank you received: 53
SOLVED!! (excuse the pun)
Altering the ANSVR config to return J2000 coordinates fixed the issue. I must have set it to JNOW once when messing around with SGP on Windows.
The following user(s) said Thank You: Jasem Mutlaq
5 years 6 months ago #29552

Please Log in or Create an account to join the conversation.

Time to create page: 0.499 seconds