Hi!

I'll try to explain my issue as simple as possible:
I'm doing spectroscopy. To center my object, I have a guidescope with a separate camera. This is the one doing plate solving.
Then my imaging camera sits on the spectroscope, on a separate OTA. This one does NOT plate-solve.
So I have 2 scopes, so to speak, with 2 cameras. I have set up them as 2 image trains. Only common feature is the mount.

Now my issue:
Lately, seemingly since I upgrade to this version of Kstars where image trains is a feature, I no longer get date and CRVAL keywords in the fits header of my imaging camera. That makes sense to me -if we now consider that independent from my centeringscope/guidescope. But it becomes a real problem in processing the images, since atmospheric extinction is very important for spectroscopy.
So: is there a way for me to get these important keywords in images from BOTH cameras/both image trains? Maybe I can set up the centering image train in a way so that it does not become independent from the imaging camera....(but I can not figure out how). Or can I use some script to insert the keywords in some way? This is really a serious issue for this processing.....

And I can't say I understand the logic. One could think that since they share the same mount, information about where that mount is pointed (CRVAL keywords) should be inserted. And I don't understand why the date and time is not inserted anymore. The Fits header from one of my recent main camera images is here:

SIMPLE = T / file does conform to FITS standard
BITPIX = 16 / number of bits per data pixel
NAXIS = 2 / number of data axes
NAXIS1 = 695 / length of data axis 1
NAXIS2 = 519 / length of data axis 2
EXTEND = T / FITS dataset may contain extensions
COMMENT FITS (Flexible Image Transport System) format is defined in 'Astronomy
COMMENT and Astrophysics', volume 376, page 359; bibcode: 2001A&A...376..359H
BZERO = 32768 / offset data range to that of unsigned short
BSCALE = 1 / default scaling factor
ROWORDER= 'TOP-DOWN' / Row Order
INSTRUME= 'Atik 314L' / CCD Name
TELESCOP= 'Celestron NexStar' / Telescope name
OBSERVER= 'Unknown ' / Observer name
OBJECT = 'HD019445' / Object name
EXPTIME = 3.000000E+02 / Total Exposure Time (s)
CCD-TEMP= -1.502E+01 / CCD Temperature (Celsius)
PIXSIZE1= 6.450000E+00 / Pixel Size 1 (microns)
PIXSIZE2= 6.450000E+00 / Pixel Size 2 (microns)
XBINNING= 2 / Binning factor in width
YBINNING= 2 / Binning factor in height
XPIXSZ = 1.290000E+01 / X binned pixel size in microns
YPIXSZ = 1.290000E+01 / Y binned pixel size in microns
FRAME = 'Light ' / Frame Type
IMAGETYP= 'Light Frame' / Frame Type
FOCALLEN= 0.000E+00 / Focal Length (mm)
APTDIA = 0.000E+00 / Telescope diameter (mm)
FOCUSPOS= 1041 / Focus position in steps
FOCUSTEM= 2.000E+01 / Focuser temperature in degrees C
END


Magnus

Magnus,

If I might piggy-back on your post because I have a similar problem.
I do two types of work: spectroscopy and variable stars. I have a control room; however, all the equipment is connected to a Stellar Mate in the dome.
When I do variable stars, I have a SX Xpress Lodestar as my guide camera in a filter wheel. The camera is a ZWO ASI2600. I use the Ekos guider.
For spectroscopy, my guide camera is, again, a Lodestar which is embedded in my spectroscope. Same ZWO camera. In this case I use the PHD2 program in the SM as my guide software.
I noticed that, when I did spectroscopy work, my FITS headers were sparse. So I ran some tests. Now, regardless of whether I'm doing variable stars or spectroscopy, the primary and secondary image profiles are the same except for the filter wheel which I removed for these tests.
The tests were:
1. Control h/w via SM, track with Ekos software.
2. Control h/w via SM, track using PHD2.
3. Control h/w via computer in control room, track with Ekos s/w.
4. Control h/w via computer in control room, track with PHD2.

OK -- so I see we still can't embed even txt files.

Here is the FITS header when using SM-based PHD2 to do the guiding.


SIMPLE = T / file does conform to FITS standard
BITPIX = 16
NAXIS = 2
NAXIS1 = 6248
NAXIS2 = 4176
EXTEND = T / FITS dataset may contain extensions
COMMENT FITS (Flexible Image Transport System) format is defined in 'Astronomy
COMMENT and Astrophysics', volume 376, page 359; bibcode: 2001A&A...376..359H
BZERO = 32768
BSCALE = 1
DATAMIN = 232. / Minimum value
DATAMAX = 23213. / Maximum value
MIN1 = 232. / Min Channel 1
MAX1 = 23213. / Max Channel 1
MEAN1 = 497.989868564916 / Mean Channel 1
MEDIAN1 = 498. / Median Channel 1
STDDEV1 = 11.5523194653755 / Standard Deviation Channel 1
ROWORDER= 'TOP-DOWN' / ROWORDER
INSTRUME= 'ZWO CCD ASI2600MM Pro' / INSTRUME
TELESCOP= 'Celestron CGX' / TELESCOP
OBSERVER= 'Unknown ' / OBSERVER
OBJECT = 'Unknown ' / OBJECT
EXPTIME = 1. / Total Exposure Time (s)
CCD-TEMP= 1.8 / CCD Temperature (Celsius)
PIXSIZE1= 3.76 / Pixel Size 1 (microns)
PIXSIZE2= 3.76 / Pixel Size 2 (microns)
XBINNING= 1 / Binning factor in width
YBINNING= 1 / Binning factor in height
XPIXSZ = 3.76 / X binned pixel size in microns
YPIXSZ = 3.76 / Y binned pixel size in microns
FRAME = 'Light ' / FRAME
IMAGETYP= 'Light Frame' / IMAGETYP
FOCALLEN= 0. / Focal Length (mm)
APTDIA = 0. / Telescope diameter (mm)
FOCUSPOS= 25230 / Focus position in steps
FOCUSTEM= 20.94 / Focuser temperature in degrees C
GAIN = 150. / GAIN
OFFSET = 50. / OFFSET
END

Here is the FITS header representative of the other three tests (they're basically the same):

SIMPLE = T / file does conform to FITS standard
BITPIX = 16
NAXIS = 2
NAXIS1 = 6248
NAXIS2 = 4176
EXTEND = T / FITS dataset may contain extensions
COMMENT FITS (Flexible Image Transport System) format is defined in 'Astronomy
COMMENT and Astrophysics', volume 376, page 359; bibcode: 2001A&A...376..359H
BZERO = 32768
BSCALE = 1
DATAMIN = 160. / Minimum value
DATAMAX = 22080. / Maximum value
MIN1 = 160. / Min Channel 1
MAX1 = 22080. / Max Channel 1
MEAN1 = 498.00079335684 / Mean Channel 1
MEDIAN1 = 498. / Median Channel 1
STDDEV1 = 11.3757469274552 / Standard Deviation Channel 1
ROWORDER= 'TOP-DOWN' / ROWORDER
INSTRUME= 'ZWO CCD ASI2600MM Pro' / INSTRUME
TELESCOP= 'Celestron CGX' / TELESCOP
OBSERVER= 'Unknown ' / OBSERVER
OBJECT = 'Unknown ' / OBJECT
EXPTIME = 1. / Total Exposure Time (s)
CCD-TEMP= 0.9 / CCD Temperature (Celsius)
PIXSIZE1= 3.76 / Pixel Size 1 (microns)
PIXSIZE2= 3.76 / Pixel Size 2 (microns)
XBINNING= 1 / Binning factor in width
YBINNING= 1 / Binning factor in height
XPIXSZ = 3.76 / X binned pixel size in microns
YPIXSZ = 3.76 / Y binned pixel size in microns
FRAME = 'Light ' / FRAME
IMAGETYP= 'Light Frame' / IMAGETYP
FOCALLEN= 1702. / Focal Length (mm)
APTDIA = 304. / Telescope diameter (mm)
FOCUSPOS= 25230 / Focus position in steps
FOCUSTEM= 20.75 / Focuser temperature in degrees C
SCALE = 0.4556438 / arcsecs per pixel
SITELAT = 32.4875 / Latitude of the imaging site in degrees
SITELONG= -107.9844 / Longitude of the imaging site in degrees
AIRMASS = 2.047123 / AIRMASS
OBJCTAZ = 1.068237 / Azimuth of center of image in Degrees
OBJCTALT= 29.1732 / Altitude of center of image in Degrees
OBJCTRA = '17 32 52.37' / Object J2000 RA in Hours
OBJCTDEC= '86 35 10.27' / Object J2000 DEC in Degrees
RA = 263.2182 / Object J2000 RA in Degrees
DEC = 86.58619 / Object J2000 DEC in Degrees
PIERSIDE= 'WEST ' / West, looking East
EQUINOX = 2000 / EQUINOX
DATE-OBS= '2023-04-02T00:59:50.222' / 2023-04-02T00:59:50.222
COMMENT = ' ' / Generated by INDI
GAIN = 150. / GAIN
OFFSET = 50. / OFFSET
END


This second example has 13 more rows than the example where SM PHD2 is doing the guiding. Things like observer Lat/Lon, RA/Dec, the equivalent focal length of my telescope with is computed correctly given my 0.7 focal reducer.

So here's my question: what writes the FITS header? If I doing my usual spec work using the SM PHD2 guider, is the header written on the Stellar Mate, or on the computer in the control room. The FITS header in this instance is not satisfactory. In the case of SM PHD2 guiding, there is NOT the comment "Generated by INDI" near the end of the header while the other three instances do have that notation.

Robert, Magnus,

I'm pretty unfamiliar with the fit header code, but I could look into specific problems if you could recreate the problem using Ekos with the Ekos/INDI simulators, and gave me detailed instructions on how to replicate that.

FWIW, one of the fields you both are saying is missing is DATE-OBS. It is written in the INDI library here:
github.com/indilib/indi/blob/master/libs...se/indiccd.cpp#L2217
which is likely called at the end of an exposure when the captured file is uploaded from the camera.
It creates a data blob which is passed into Ekos and eventually to this method, processBlob()
invent.kde.org/education/kstars/-/blob/m.../indicamera.cpp#L661
which calls writeImageFile() in the same file.

Hy

Hy,
I can certainly do that. With a blazing full moon, what else is there!
I'd like to hear from Magnus as to how he's doing his guiding. His FITS header looks the same as mine when I do guiding via an instance of PHD2 running on Stellar Mate. Again, since there is no Comment "Generated by INDI" at the bottom of the header, is something else creating the header besides INDI?

Thanks,
Bob

Well, this should keep the s/w developers busy.
The difference between an adequate FITS header and one not so good depends on one combination:
Whether a real telescope is being guided via a remote instance of PHD2 on Stellar Mate.

If a sim telescope is guided by PHD2 on SM, regardless of whether anything else in the chain (e.g., cameras) is real or sim, the FITS header is good.
If a real telescope is guided by PHD2 on SM, regardless of whether anything else in the chain (e.g., cameras) is real or sim, the FITS header is UNSAT.
If a real telescope is guided by PHD2 on a local computer, the FITS header is good. (See my post from yesterday).

So, in order to have an unsatisfactory FITS header, one needs:
1. Real telescope
2. PHD2 guiding
3. Remote SM host

Having stared at these three items, something I need to do is
1. Login to SM
2. Create Ekos chains like I have on the computer in the control room
3. See if I get the same results.

Have to run. I will post the relevant FITS headers shortly.

FITS Headers

Telescope SIM using PHD2 guidance on SM

SIMPLE = T / file does conform to FITS standard
BITPIX = 16
NAXIS = 2
NAXIS1 = 3120
NAXIS2 = 2088
EXTEND = T / FITS dataset may contain extensions
COMMENT FITS (Flexible Image Transport System) format is defined in 'Astronomy
COMMENT and Astrophysics', volume 376, page 359; bibcode: 2001A&A...376..359H
BZERO = 32768
BSCALE = 1
DATAMIN = 423. / Minimum value
DATAMAX = 2995. / Maximum value
MIN1 = 423. / Min Channel 1
MAX1 = 2995. / Max Channel 1
MEAN1 = 497.490033580081 / Mean Channel 1
MEDIAN1 = 497. / Median Channel 1
STDDEV1 = 5.51636762122051 / Standard Deviation Channel 1
ROWORDER= 'TOP-DOWN' / ROWORDER
INSTRUME= 'ZWO CCD ASI2600MM Pro' / INSTRUME
TELESCOP= 'Telescope Simulator' / TELESCOP
OBSERVER= 'Unknown ' / OBSERVER
OBJECT = 'Unknown ' / OBJECT
EXPTIME = 1. / Total Exposure Time (s)
CCD-TEMP= -0.9 / CCD Temperature (Celsius)
PIXSIZE1= 3.76 / Pixel Size 1 (microns)
PIXSIZE2= 3.76 / Pixel Size 2 (microns)
XBINNING= 2 / Binning factor in width
YBINNING= 2 / Binning factor in height
XPIXSZ = 7.52 / X binned pixel size in microns
YPIXSZ = 7.52 / Y binned pixel size in microns
FRAME = 'Light ' / FRAME
IMAGETYP= 'Light Frame' / IMAGETYP
FOCALLEN= 900. / Focal Length (mm)
APTDIA = 120. / Telescope diameter (mm)
FOCUSPOS= 50000 / Focus position in steps
FOCUSTEM= 0. / Focuser temperature in degrees C
SCALE = 1.723751 / arcsecs per pixel
SITELAT = 32.48778 / Latitude of the imaging site in degrees
SITELONG= -107.9847 / Longitude of the imaging site in degrees
AIRMASS = 1.858717 / AIRMASS
OBJCTAZ = 359.9927 / Azimuth of center of image in Degrees
OBJCTALT= 32.48814 / Altitude of center of image in Degrees
OBJCTRA = '23 55 01.12' / Object J2000 RA in Hours
OBJCTDEC= '89 52 09.61' / Object J2000 DEC in Degrees
RA = 358.7546 / Object J2000 RA in Degrees
DEC = 89.86934 / Object J2000 DEC in Degrees
EQUINOX = 2000 / EQUINOX
DATE-OBS= '2023-04-03T19:54:08.976' / 2023-04-03T19:54:08.976
COMMENT = ' ' / Generated by INDI
GAIN = 150. / GAIN
OFFSET = 50. / OFFSET
END

Real telescope (Celestron CGX) using PHD2 guidance on SM. Notice the lack of "Generated by INDI" comment

SIMPLE = T / file does conform to FITS standard
BITPIX = 16
NAXIS = 2
NAXIS1 = 3120
NAXIS2 = 2088
EXTEND = T / FITS dataset may contain extensions
COMMENT FITS (Flexible Image Transport System) format is defined in 'Astronomy
COMMENT and Astrophysics', volume 376, page 359; bibcode: 2001A&A...376..359H
BZERO = 32768
BSCALE = 1
DATAMIN = 419. / Minimum value
DATAMAX = 5566. / Maximum value
MIN1 = 419. / Min Channel 1
MAX1 = 5566. / Max Channel 1
MEAN1 = 497.582197214373 / Mean Channel 1
MEDIAN1 = 497. / Median Channel 1
STDDEV1 = 7.04317746280766 / Standard Deviation Channel 1
ROWORDER= 'TOP-DOWN' / ROWORDER
INSTRUME= 'ZWO CCD ASI2600MM Pro' / INSTRUME
TELESCOP= 'Celestron CGX' / TELESCOP
OBSERVER= 'Unknown ' / OBSERVER
OBJECT = 'Unknown ' / OBJECT
EXPTIME = 1. / Total Exposure Time (s)
CCD-TEMP= 5.4 / CCD Temperature (Celsius)
PIXSIZE1= 3.76 / Pixel Size 1 (microns)
PIXSIZE2= 3.76 / Pixel Size 2 (microns)
XBINNING= 2 / Binning factor in width
YBINNING= 2 / Binning factor in height
XPIXSZ = 7.52 / X binned pixel size in microns
YPIXSZ = 7.52 / Y binned pixel size in microns
FRAME = 'Light ' / FRAME
IMAGETYP= 'Light Frame' / IMAGETYP
FOCALLEN= 0. / Focal Length (mm)
APTDIA = 0. / Telescope diameter (mm)
FOCUSPOS= 50000 / Focus position in steps
FOCUSTEM= 0. / Focuser temperature in degrees C
GAIN = 150. / GAIN
OFFSET = 50. / OFFSET
END

Hi!

Sorry to be a bit late now.

I haven't had this issue for a while because I changed my setup. I had this when using a side-by-side with 2 Stellarmates - now I have two separate mounts and one SM by each.
However: the topic developed some in direct conversation with Jasem. This is how I finally understood it: I was using the two SMs in chained mode, so that the mount was directly connected to one, and chained to the second via the first (a remote device in the profile setup). What I understood was that the DATE_OBS information was not shared in the chained mode, so the camera connected to the second SM never received that information. Hence a truncated FITS-header.

And for spectroscopy, I use a separate (guide)scope for positioning and plate solving, not the guidecam in the spectroscope. But I guide with that one, using a reflective slit plate and PHD2. But that means that the poistioning information is associated with the positioning camera, not the guidecam - and thus possibly not with the imaging camera (so I have 3 cameras for spectrosocopy).

If I get it right, you are not chaining but connecting a client (like PHD2) directly to the indi server, from a different computer than the one the mount is connected to (right?). It is not exactly the same setup but there is some similarities.

Does this make sense?

Magnus

Hy,

Here's what I have. To say that results have been inconsistent is an understatement. Is there a cache somewhere? Anyway, after doing and redoing these tests at least a dozen times, I now have consistent, definitive results.

The problem seems to be a combination of Stellar Mate and a real telescope.

I have a host computer in a control room and all the equipment is attached to SM in a dome. All the s/w is up-to-date. If I connect via Ekos using my Celestron CGX, my FITS headers are the truncated version (no RA/DEC, Lat/Lon, or "Generated by INDI") in the header regardless of guiding method.
If I substitute a Telescope Sim, then all the values we would expect appear in the FITS header. This is true whether I'm using the internal Ekos guide or PHD2.

If a VNC into SM, and now run everything via SM (in effect taking the computer in the control room out of the loop), the results are the same. Using the CGX results in truncated FITS headers; if I substitute a Tel Sim, then the headers are robust.

I have FITS files from Jan 2023 when I was doing variable star work (not using PHD2) and they are of even better quality than the "good FITS" header I'm seeing when using a Tel Sim. These headers contain values for CRVAL1 and CRVAL2 that Magnus was referencing. Maybe these test results are being influenced in that I'm not doing any real tracking; I just start Ekos and take a 1 sec frame. Tomorrow night this howling wind in s/w New Mexico should subside and I can take some actual data, i.e., track on a star, take some data, and look at the headers.

Here's the header from a file using Ekos internal guiding from back in Jan 2023. The FITS headers when using PHD2 were, as they are now, truncated.

SIMPLE = T / file does conform to FITS standard
BITPIX = 16
NAXIS = 2
NAXIS1 = 3120
NAXIS2 = 2088
EXTEND = T / FITS dataset may contain extensions
COMMENT FITS (Flexible Image Transport System) format is defined in 'Astronomy
COMMENT and Astrophysics', volume 376, page 359; bibcode: 2001A&A...376..359H
BZERO = 32768
BSCALE = 1
ROWORDER= 'TOP-DOWN' / Row Order
INSTRUME= 'ZWO CCD ASI2600MM Pro' / CCD Name
TELESCOP= 'Celestron CGX' / Telescope name
OBSERVER= 'Unknown ' / Observer name
OBJECT = 'M 67 ' / Object name
EXPTIME = 3.000000E+01 / Total Exposure Time (s)
CCD-TEMP= -1.980E+01 / CCD Temperature (Celsius)
PIXSIZE1= 3.760000E+00 / Pixel Size 1 (microns)
PIXSIZE2= 3.760000E+00 / Pixel Size 2 (microns)
XBINNING= 2 / Binning factor in width
YBINNING= 2 / Binning factor in height
XPIXSZ = 7.520000E+00 / X binned pixel size in microns
YPIXSZ = 7.520000E+00 / Y binned pixel size in microns
FRAME = 'Light ' / Frame Type
IMAGETYP= 'Light Frame' / Frame Type
FILTER = 'Red ' / Filter
FOCALLEN= 1.702E+03 / Focal Length (mm)
APTDIA = 3.040E+02 / Telescope diameter (mm)
FOCUSPOS= 25600 / Focus position in steps
FOCUSTEM= -3.438E+00 / Focuser temperature in degrees C
SCALE = 9.112876E-01 / arcsecs per pixel
SITELAT = 3.248750E+01 / Latitude of the imaging site in degrees
SITELONG= -1.079844E+02 / Longitude of the imaging site in degrees
AIRMASS = 1.494492E+00 / Airmass
OBJCTAZ = 2.556431E+02 / Azimuth of center of image in Degrees
OBJCTALT= 4.195705E+01 / Altitude of center of image in Degrees
OBJCTRA = ' 8 52 28.96' / Object J2000 RA in Hours
OBJCTDEC= '11 45 01.96' / Object J2000 DEC in Degrees
RA = 1.331207E+02 / Object J2000 RA in Degrees
DEC = 1.175054E+01 / Object J2000 DEC in Degrees
EQUINOX = 2000 / Equinox
CRVAL1 = 1.3312066677E+02 / CRVAL1
CRVAL2 = 1.1750544265E+01 / CRVAL1
RADECSYS= 'FK5 ' / RADECSYS
CTYPE1 = 'RA---TAN' / CTYPE1
CTYPE2 = 'DEC--TAN' / CTYPE2
CRPIX1 = 1.5600000000E+03 / CRPIX1
CRPIX2 = 1.0440000000E+03 / CRPIX2
SECPIX1 = 9.1128759167E-01 / SECPIX1
SECPIX2 = 9.1128759167E-01 / SECPIX2
CDELT1 = 2.5313544213E-04 / CDELT1
CDELT2 = 2.5313544213E-04 / CDELT2
CROTA1 = 4.9317010923E+01 / CROTA1
CROTA2 = 4.9317010923E+01 / CROTA2
DATE-OBS= '2000-01-21T11:12:52.298' / UTC start date of observation
COMMENT Generated by INDI
GAIN = 300. / Gain
OFFSET = 50. / Offset
END

Robert,

I don't use StellarMate, and am not connected with that product.
So, this is likely best looked at by Jasem, but I'm curious...

Can you take StellarMate out of the equation? Simply run KStars on your telescope computer via VNC (so KStars/Ekos on the same machine as your devices), run with no guiding, and take a single short photo and see if you get the headers you want. Make sure you successfully align first, so the system definitely knows the sky position (e.g. CRVAL1 has to do with the center position of the image). Make sure your optical trains are filled in properly so it knows your focal length and camera. FWIW I just checked and my images all have CRVAL1 and CRVAL2.

Hy

Hy,

Forecast is good for me being able to do tests tomorrow. Yeah -- near full moon. But all I need to do is point at something and take some data.
I can run six tests where, for each scenario, I track via Ekos internal tracker and PHD2.

1. Control everything from the computer in the control room.
2. My usual configuration, which is to use Stellar Mate to run the equipment while the computer in the control room connects to SM.
3. VNC into SM and run everything from SM.

Should keep me busy for a while.

CS,
Bob

Those tests you describe may help you debug, and perhaps help Jasem, but don't do them for me.
All I ask is that you run a simple/basic capture where the system knows the WCS info, and see if that makes it to the headers. E.g.
- Over VNC,
- Start KStars (e.g. /usr/bin/kstars)
- Start Ekos, connect to Indi,
- Don't guide
- Focus if you need to,
- Slew somewhere,
- Align successfully,
- Capture an image, and
- See if the headers are ok.

If the headers are OK, then see what you can add (one thing at a time) that causes the headers to go missing.

Sorry it took me so long to respond to this, but this is a classic SNOOP issue. The camera is not getting the mount's RA/DE and this explains why most the headers that depends on such information is missing. It is Ekos responsibility to set the snooped device property. Check the screenshot below, you see that the CCD Simulator is set to listen to the EQMod Mount driver. I suspect in your case, the camera is not getting set to the CGX? Can you check?