Dear all,

I recently tried to run a simulation of an observing session, and found myself in trouble while trying to validate pointing on my own.
I appeared that the coordinates returned for the image by astrometry.net are different from the one I set to the telescope simulator, by a rather large amount. Maybe I am using the wrong solving parameter ?

I included some examples in the attachement.

The telescope simulator is set to RA: 2:13:19, DEC: 66:39:36


Here is the pointing image if someome wants to crosscheck:

I use the following command for solve_field:

<code>solve-field --guess-scale --cpulimit 15 --no-verify --no-plots --no-fits2fits --crpix-center --match none --corr none --wcs none --downsample 4 --overwrite --skip-solved --ra 33.33 --dec 66.66 --radius 1 [...].fits
</code>

The output is the following:

Reading input file 1 of 1: "[...].fits"...
Extracting sources...
Downsampling by 4...
simplexy: found 64 sources.
Solving...
Reading file ".axy"...
Only searching for solutions within 1 degrees of RA,Dec (33.33,66.66)
Field 1 did not solve (index index-4208.fits, field objects 1-10).
Field 1 did not solve (index index-4207-00.fits, field objects 1-10).
Field 1 did not solve (index index-4206-00.fits, field objects 1-10).
Field 1 did not solve (index index-4205-00.fits, field objects 1-10).
log-odds ratio 165.597 (8.28061e+71), 35 match, 0 conflict, 23 distractors, 84 index.
RA,Dec = (32.947,66.5702), pixel scale 1.19135 arcsec/pix.
Hit/miss: Hit/miss: +++++-+-++++++++-++++-+-+++-++--+-+-+-+-+-+-+--++----+---+(best)--++++
Field 1: solved with index index-4204-03.fits.
Field 1 solved: writing to file .solved to indicate this.
Field: .fits
Field center: (RA,Dec) = (32.947331, 66.570193) deg.
Field center: (RA H:M:S, Dec D:M:S) = (02:11:47.359, +66:34:12.693).
Field size: 25.4321 x 20.348 arcminutes
Field rotation angle: up is 90.0147 degrees E of N
Creating new FITS file

I was wondering if I was supposed to indicate a specific coordinate as the target for the solving informations:

--crpix-x <pix>: set the WCS reference point to the given position
--crpix-y <pix>: set the WCS reference point to the given position

Because my current virtual pointing error looks wrong.

Thank you very much for your help,
Cheers, and thank you for this wonderfull tool
PS: here is my version of indi:

indiserver --version
2019-04-09T06:07:23: startup: indiserver --version
Usage: indiserver [options] driver [driver ...]
Purpose: server for local and remote INDI drivers
INDI Library: 1.7.5
Code v1.7.4-391-g8f25e25. Protocol 1.7.

I am a bit confused what you are trying to do here....
When using sims, you can still set the telescope home position on Polaris, then slew to an object on the sky map, take an image and solve, and if the mount is not quite in the centre you can adjust then take an image and sync that position, the INdI driver does not really come into the equation...there is no need to do anything with that... or maybe I am not getting what you are trying to do at all.....sorry...

AstroNerd wrote: I am a bit confused what you are trying to do here....
When using sims, you can still set the telescope home position on Polaris, then slew to an object on the sky map, take an image and solve, and if the mount is not quite in the centre you can adjust then take an image and sync that position, the INdI driver does not really come into the equation...there is no need to do anything with that... or maybe I am not getting what you are trying to do at all.....sorry...

Thank you for your answer.
Fair enough, maybe I did not gave enough background information.
I am talking about a testbench here, meaning that I am using a laptop with absolutely no physical device that is related to astronomy, only simulators.
I am trying to slew my virtual telescop to a specific location in the sky:
RA-DEC=(33.33, 66.66) in deg,deg format or equivalently
RA: 2:13:19, DEC: 66:39:36 in hh:mm:ss / dd:mm:ss format

And when the slew is over (the virtual telescope is set to slew and track), then I take a picture with the virtual camera, that is most likely snooping for the telescope informations and retrieving the right position information.

After image acquisition, I try to check with the solve-field command if the astrometry software returns value that match up to some precision to the position I set though indi.

The coordinates that are returned by astrometry are (as you can see in the logs above):
Field center: (RA,Dec) = (32.947331, 66.570193) deg.
Field center: (RA H:M:S, Dec D:M:S) = (02:11:47.359, +66:34:12.693).

Which is not very close to what I was looking for (more than 600 arcsec drift)
I am trying to figure out what is the reason for this discrepancy

Ok I think there is something I do not understand here, and that is very simple to reproduce:
Launch simulator using ekos with telescope + ccd camera and got to the following coordinates: RA: 2:13:19, DEC: 66:39:36 in hh:mm:ss / dd:mm:ss format

Then in ekos, go to the target icon and click on acquire and solve.
Save the resulting image to fit, and use the set of astrometry parameter provided by ekos, and launch the solve_field command on your own with the same set of parameter.
When I do that, I see different results from the one reported by ekos gui ???

How is it possible ?

EDIT: might have found the reason why: indilib.org/forum/ekos/3865-plate-solvin...g-wrong-results.html
There is an implicit conversion from J2000 (Astrometry) to JNow (Indi) hapening apparently.

I had no idea this change of frame could result in such a big difference, I now need to investigate this issue further

Right, understand now
I do similar when learning the software too...
But I must admit I have never checked the solved result to the one in Kstars....maybe I should..
I would have thought that a lot would be to do with how you have set up your home position, and the mount model, ie how many sync points you have set

dolguldur wrote: Ok I think there is something I do not understand here, and that is very simple to reproduce:
Launch simulator using ekos with telescope + ccd camera and got to the following coordinates: RA: 2:13:19, DEC: 66:39:36 in hh:mm:ss / dd:mm:ss format

Then in ekos, go to the target icon and click on acquire and solve.
Save the resulting image to fit, and use the set of astrometry parameter provided by ekos, and launch the solve_field command on your own with the same set of parameter.
When I do that, I see different results from the one reported by ekos gui ???

How is it possible ?

EDIT: might have found the reason why: indilib.org/forum/ekos/3865-plate-solvin...g-wrong-results.html
There is an implicit conversion from J2000 (Astrometry) to JNow (Indi) hapening apparently.

I had no idea this change of frame could result in such a big difference, I now need to investigate this issue further

Well you go about all that a weird way indeed....

This is my way, all with no kit connected ted just sims...
Start up and unpark mount, then click right on an object on the skymap, and click track, the scope will then slew to the target and track, then in the align tab, click take image and solve, and check the box next to it that says, slew to target, then if it is not centred it will move and take another image and solve, and it will keep doing that until the target is within and acceptable range....you can also do the same but choose sync instead, this will build a model and give better accuracy.
When the images come up in the viewer, click the I button to show all stars in image and then you can click one from there and centre if you wish...