It is quite frequent that my plate solve attempts fail due to lack of known stars in the FOV. I'm looking for ways to improve the chance of a successful solve.
My problem stems from small ST7 chip coupled with relatively long 2190mm focal length LX200. I like the combination for other reasons so am not keen on further reducing the focal length or such solutions.
I plate solve mainly using CCDSoft coupled with TheSky. I attempt this most commonly automated using CCDCommander, which supports both Software Bisque and Maxim but not Astrometrica.
I also plate solve using Astrometrica sometimes manually.
The chance of a failure is much higher when using CCDSoft than when using Astrometrica. For example this morning the same FOV successfully solved in both, but with 8 stars in CCDSoft vs 44 stars in Astrometrica.
I think the difference in success between Astrometrica and CCDSoft is that I do not have USNO-B1 (80gb version) on disk at home for CCDSoft to utilise where as Astrometrica accesses it online through the web "on demand".
I would like to improve the chances of CCDCommander performing a successful plate solve. The following are the best ideas I have come up with:
Acquiring a copy of USNO-B1.
Purchasing TPonit such that my pointing accuracy is perhaps imrpoved to the point where I do not need to plate solve.
Purchasing a new camera that has a larger chip size but similar pixel size.
Of the above, option 2 is looking like the most realistic.
Option 1: It appears USNO-B1 is hard to come by. I have found an observatory that will put a copy on disk but you must send them a external HD etc. I expect total cost would end up about the same as TPoint or a bit more. The main down-side I see to this is TheSky would likely become very slow to use due to the enormous number of stars it would be handling, on my aging 2.6ghz machine.
Option 2: This may work but I am not sure if TPoint will be able to fully comphensate for my amateur LX200 with all it's associated inaccuracies to the extent that objects will always end up in the FOV.
Option 3: Not an option at this stage due to $$$$. ST-10XME is on the list, but year(s) away.
I'd appreciate input from anyone with experience or comments.
Thanks,
Roger.
PS - I have tried MaxPoint and while I prefer it's price to TPoint I found TheSky's usage of ASCOM unmanageable for my setup, it just didn't work smoothly enough to be reliable, so at this stage MaxPoint is not being considered, I would prefer to pay for TPoint.
Hmm, yes, interesting facility they're building, should be great. But unfortunately I don't think it's going to solve my problems that involve automation
You don't need the USNO-B1 to plate solve at the focal length and FOV you're talking about. The USNO-A2.0 catalogue will meet the criteria with no problems at all. It contains 526,280,881 stars down to Mag 20. If you can't use this, there is something wrong with your set up / software configuration. You can download it here ftp://ftp.nofs.navy.mil/usnoa - 6Gb download.
Keep in mind for successful plate solving you don't need to solve every single star in the field. Typically 20-80 stars works well. The more the better, but it can slow the process down. Finally, make sure you telescope is pointing reasonably accurate. The plate solve process looks at the RA/DEC in the FITS header to pass to the plate solve process. If this is significantly out, plate solving will fail and/or take a long time solve as it churns through the lookup catalogues.
Can someone describe the plate solve process. I also have an ST7 and LX200GPS, and am looking to get the pointing accuracy better. If for nothing else, just getting things on the of chip can sometimes take 15 minutes of visually comparing the FOV indicators in TheSky and what is actually coming up on the chip to see how far off I am.
Does the plate solve process actually move the scope and re-sync for you?
Yes. Firstly, as I indicated earlier, the output (downloaded image) must be in FITS format. Plate solving will not work with any other formats such as TIFF or RAW etc. With that out of the way...
The plate solve process is actually quite logical. The camera takes a picture of "any" area of the sky. When downloaded, the FITS format file header is updated with the telescope RA/DEC information. A camera control program (such as CCDSoft or MaximDL) then launches a plate solving engine. There are a few s/w developers that provide the engine integrated such as Bisque, while there are also some independent sorts such as PinPoint by Bob Denny. The camera control program makes reference calls to the plate solving engine which in turn begins looking up astrometry star data contained in one of many catalogues. Once a matching star field is acknowledged, the plate solving engine return back to the camera control program the precise RA/DEC (true image center), image position angle, etc which is inserted into the FITS format file header.
Now, at this point, all you've done is plate solve the image. Its up to the camera control program or supplemental program (such as CCDCommander, ACP, CCDAutoPilot) to now do something with this precise pointing information. The typical usage is to compare coordinates with the object entered and move the telescope the offset amount. This ensure the object enters is always centered on the chip. By the sounds of things, this is what you're after. You're telescope pointing needs to be reasonably accurate to begin with. If its not, the plate solving engine will have a hard time searching the catalogue for a match, thus delaying operations i.e supernova searches for example. Though there are developers allow the telescope to slew while the image is being downloaded and plates solved. This improves supernova search efficiencies. It should be noted that plate solving also has other uses such as focusing. FocusMax references the plate solved data to determine a star which is between -4 and -7 magnitude and above X degrees to perform a focus run on. This is important for camera controlled focusing (automation). IF a bright star is selected for focusing, it may saturation well depth and/or produce glare making FWHM calculations difficult. (Sorry about the deviation on the above re: focusing, alas its a valid example).
You typically need to download the catalogue, but some include it on the source CD or DVD to make it easier. Select a catalogue that works best for you. If you've got small FOV, use the USNO-A2.0 catalogue. It has the potential resolve stars even on very small FOV's. In most cases, it is possible to also use the Hubble GSC (Guide Star Catalogue) for these conditions. If you've got a wide field instrument, I'd recommend using the Hubble GSC. There is more than enough stars to plate solve on in GSC catalogue. The USNO-A2.0 is deep. I certainly can't use it with the FSQ (even when I alter the magnitude search limits in the plate solving engine) as the plate solve engine somehow tries to resolve 30,000 stars. Crazy.
You don't need the USNO-B1 to plate solve at the focal length and FOV you're talking about. The USNO-A2.0 catalogue will meet the criteria with no problems at all.
Interesting - Yet I don't see many stars in my FOV in TheSky - suggesting the databases it's using (which includes USNO-A2.0) don't contain many stars in the FOV?
Quote:
Originally Posted by jase
Keep in mind for successful plate solving you don't need to solve every single star in the field. Typically 20-80 stars works well. The more the better, but it can slow the process down.
Yeap, my problem is simply not having enough. 80! I wish My CCD images have plenty, but often there just isn't many at all visible in TheSky, presumably why it can't plate solve.
Quote:
Originally Posted by jase
Finally, make sure you telescope is pointing reasonably accurate.
Yeap, done. Pointing is not perfect, but definitely within range.
Interesting - Yet I don't see many stars in my FOV in TheSky - suggesting the databases it's using (which includes USNO-A2.0) don't contain many stars in the FOV?
Yeap, my problem is simply not having enough. 80! I wish My CCD images have plenty, but often there just isn't many at all visible in TheSky, presumably why it can't plate solve.
You sure you're using the USNO-A2.0? TheSky supports it, but its not installed by default...(its an optional add-on). Check the file locations dialogue box under data menu (see output below). Would suggest you verify your core databases and specifically the USNO-A2.0. If its got a green tick, its loaded ok. However, this doesn't mean its active. You need to go to the display explorer | Stellar options and check, that its enabled there (see other output provided).
There is likely to also be other configuration options within CCDcommander or WCS to determine the star magnitude search criteria. Seriously, if you're search for stars from mag -2 all the way down to mag 20, you'll definitely be able to plate solve the 5.3 x 7.9 arcmin FOV delivered by the 12" LX200 @ F/10 (3000mm F/L) using that ST7. You may have some configuration work / tweaking to perform, but it will work.
Thanks for your reply. I appreciate you may not be able to help me any further but anyhow, to confirm I have the correct settings I have attached a couple of matching screen shots.
I wouldn't be surprised if there is some other setting, but I cannot find anything which limits the magnitude of stars used in a plate solve, however, that is exactly the kind of setting I would expect to exist somewhere that I haven't found...
Roger, would it be possible to place a few of your subs on the internet to download, say a 2 or 5 minute one. If possible another which is much shorter, say 10 or 20 seconds. Preferably calibrated, but it shouldn't make much difference. I'd be interested in seeing if it could be plate solved manually. I'm not familiar with CCDcommander so can't advise much there.
I don't have any of the plate solve images saved so will save some when I next run it all. I have some normal images that don't plate solve, but they have very few stars in the image. My plate solves are the same FOV but 25 sec @ 2x2 rather than 30sec @ 1x1 so I get more stars.
I'm starting to wonder if it's just that I need longer plate solve exposures, perhaps 25 sec @ 2x2 isn't enough. I am getting confused now and need to try it again (won't be for a few nights, thunderstorms forcast).
I'm quite certain that there's often been countless stars but no successful plate solve ... hmm..
I think it's fair to ignore CCDCommander as all plate solves that have failed in CCDCommander have failed when done manually in CCDSoft (that I've tried). I'm quite confident that CCDCommander is simply invoking the CCDSoft operation without additional restraints.
Probably already know this, but if you're binning, the plate scale will change. You'll need to ensure that the plate solving engine is aware of this, though some calculate it for you based on the 1x1 arcsec/pixel. Others don't, thus you need to determine the arcsec/pixel 2x2 bin combo. All the settings you need for plate solving provided in CCcommander (see output below). It is possible to select the catalogue mag. search ranges here. I think you're on the right track if you can resolve the plate solves in CCDSoft/TheSky first.
Probably already know this, but if you're binning, the plate scale will change.
Yeap, got that sorted, thanks.
Quote:
Originally Posted by jase
All the settings you need for plate solving provided in CCcommander (see output below).
I'll investigate why my settings are different to yours. Perhaps it's related to what other software is installed, but I don't have the options for PinPoint (full) which is where the magnitude limits are set, probably because I don't have PinPoint installed. To my knowledge those PinPoint settings (including mag limits) have no affect on TheSky/CCDSoft plate solves.
Quote:
Originally Posted by jase
I think you're on the right track if you can resolve the plate solves in CCDSoft/TheSky first.
Yeah, quite happy to ignore CCDCommander for now, sort the problem out at it's lowest level first.
I now have the latest version of CCDCommander which includes those options, but I hve also confirmed that they do not impact the TheSky/CCDSoft plate solve process at all, so they are not the source of my problem
It still seems the overwhelming majority of people expect that my FOV is too small for a high rate of successful plate solve in many areas of the sky.
I have re-calculated my FOV and it's actually 10.85 x 7.23 arc minutes, different to what I had remembered but not by a huge amount.
I am yet to have a clear night to try longer exposures.
With such a small FOV there will be times when you fail to obtain sufficient stars to plate solve. The minimum required in pinpoint is 6 but that is a user setting and not a very reliable one. Next you need to ensure that what can be detected in your images as stars reasonably closely matches the number of stars identified int he catalogue. Have you defined a limiting magniutde for your catalogue - or is one defined for you? Pinpoint typically expects your pointing images to at least get to magnitude 18. Again it's a user setting.
Lastly, how good are your stars? Are the actually donuts? is the FWHM to large, are they over exposed, have you flat fielded and applied darks t them? The main issue other than number of stars is the software recognising the stars in your images as stars.
Focus is good - FocusMax helps with that. I can confirm anyhow that the stars are in focus. There is sometimes slight drift.
In CCDSoft I have configured the magnitudes to permit maximum range, it was set to that by default. I don't believe there is a setting for the minimum number of stars to use, as there is in PinPoint.
I'll have to take a look at one that matches and count the stars, see if they match, that'd be an interesting exercise.
Flats - no, Darks - yes. Flats could be tricky, I don't think CCDCommander supports that.
I'm still waiting for clear skies again so I can test all the good info I've learned in the last week from helpful people.
This is simply a configuration problem Roger. Both images you supplied I plate solved successfully in MaximDL & Pinpoint using the USNO-A2.0. I let it search from -2 to 20 mag stars. The first image with only a few stars visible took some time, but still returned true image center information with reasonable precision.
Below is the FITS output of the second file (failure2). I've highlighted the plate solve info in bold. It matched 17 stars. Not perfect, but still usable for what I suspect you want to achieve.
SIMPLE = T/CCDSOFT-SOFTWARE BISQUE 3
BITPIX = 16
NAXIS = 2
NAXIS1 = 382
NAXIS2 = 255
BSCALE = 1.0000000000000000
BZERO = 0.00000000000000000
BIAS = 100
FOCALLEN = +2.300000000000e+003
APTAREA = +2.258060000000e+005
APTDIA = +3.000000000000e+002
ORIGIN = 'Darling Range Observatory'
TELESCOP = 'Meade 12" LX200' / Telescope name
OBSERVER = 'Roger Groom' / Observer name
DATE-OBS = '2008-02-09T12:41:40.171' / [ISO 8601] UTC date/time of exposure start
TIME-OBS = '12:41:40' / [old format] UTC time of exposure start
SWCREATE = 'CCDSoft Version 5.00.185'
SET-TEMP = -5.000000000000e+000
COLORCCD = 0
DISPCOLR = 1
IMAGETYP = 'Light Frame '
CCDSFPT = 1
XORGSUBF = 0
YORGSUBF = 0
CCDSUBFL = 0
CCDSUBFT = 0
XBINNING = 2 / Binning level along the X-axis
CCDXBIN = 2
YBINNING = 2 / Binning level along the Y-axis
CCDYBIN = 2
EXPSTATE = 293
CCD-TEMP = -4.817803680962e+000
TEMPERAT = -4.81780368096E+000 / [Deg. C] Temperature of CCD detector chip
OBJECT = 'NGC 1185' / Target object name
OBJCTRA = '03 04 00.00' / [hms J2000] Target right ascension
OBJCTDEC = '-09 07 00.0' / [dms +N J2000] Target declination
TELTKRA = +3.115313074245e-305
TELTKDEC = +2.639312901468e-308
CENTAZ = +2.929575640478e+002
CENTALT = +4.677778573741e+001
TELHA = '02 39 22.380 '
LST = '05 42 18.676 '
AIRMASS = 1.37228533957E+000 / Airmass (multiple of zenithal airmass)
SITELAT = '-31:52:53.00 '
SITELONG = '-116:10:12.30 '
INSTRUME = 'SBIG ST-7 Dual CCD Camera' / Detector instrument name
EGAIN = +2.300000000000e+000
E-GAIN = +2.300000000000e+000
XPIXSZ = +1.800000000000e+001
YPIXSZ = +1.800000000000e+001
SBIGIMG = 5
USER_2 = 'SBIG ST-7 Dual CCD Camera'
DATAMAX = 65535
SBSTDVER = 'SBFITSEXT Version 1.0'
FILTER = 'Clear ' / Filter name
EXPTIME = 2.50000000000E+001 / [sec] Duration of exposure
EXPOSURE = 2.50000000000E+001 / [sec] Duration of exposure
CBLACK = 266
CWHITE = 330
INPUTFMT = 'FITS ' / Format of file from which image was read SWOWNER = 'Jase ' / Licensed owner of software
HISTORY File was processed by PinPoint 4.1.10 at 2008-02-09T14:58:39
DATE = '09/02/08' / [old format] UTC date of exposure start
UT = '12:41:40' / [old format] UTC time of exposure start
TIMESYS = 'UTC ' / Default time system
RADECSYS = 'FK5 ' / Equatorial coordinate system
RA = '03 04 00.00' / [hms J2000] Target right ascension
DEC = '-09 07 00.0' / [dms +N J2000] Target declination
CLRBAND = 'R ' / [J-C std] Std. color band of image or C=Color
ZMAG = 1.99983462221E+001 / Mag zero point for 1 sec exposure
EQUINOX = 2000.0 / Equatorial coordinates are J2000
EPOCH = 2000.0 / (incorrect but needed by old programs)
CTYPE1 = 'RA---TAN' / X-axis coordinate type
CRVAL1 = 4.56672494233E+001 / X-axis coordinate value
CRPIX1 = 1.91000000000E+002 / X-axis reference pixel
CDELT1 = -4.72269273228E-004 / [deg/pixel] X-axis plate scale
CROTA1 = -1.45381330567E+000 / [deg] Roll angle wrt X-axis
CTYPE2 = 'DEC--TAN' / Y-axis coordinate type
CRVAL2 = -9.13743436287E+000 / Y-axis coordinate value
CRPIX2 = 1.27500000000E+002 / Y-axis reference pixel
CDELT2 = -4.72204122167E-004 / [deg/pixel] Y-Axis Plate scale
CROTA2 = -1.45381330567E+000 / [deg] Roll angle wrt Y-axis
CD1_1 = -4.72117250540E-004 / Change in RA---TAN along X-Axis
CD1_2 = -1.19803409491E-005 / Change in RA---TAN along Y-Axis
CD2_1 = 1.19819939036E-005 / Change in DEC--TAN along X-Axis
CD2_2 = -4.72052120451E-004 / Change in DEC--TAN along Y-Axis
HISTORY WCS added by PinPoint 4.1.10 at 2008-02-09T14:58:39
HISTORY Matched 17 stars from the USNO-A2.0 Catalog
HISTORY Average residual was 0.68 arc-seconds
PLTSOLVD = T / Plate has been solved by PinPoint
Sorry, really can't help you with the CCDSoft/TheSky plate solving. I used both tools for a couple of months when I first got the CCD camera. I did have plate solving work well at the time. I've since moved to MaximDL/Pinpoint using ACP to automate the two. I have not looked back. Pinpoint 5 - http://pinpoint.dc3.com/. Its simple, quick and accurate - a considerable improvement from version 4. I still use version 4 on my laptop, but the observatory computer uses ver 5. You may want to pull down the engine and use it with CCDcommander. The added benefit with pinpoint is that you can then really automate FocusMax with its acquirestar routine. Anything that automates the mundane of image acquisition is pretty cool in my opinion. I much prefer to do less work in acquiring the data and more work in target planning and image processing.
If you're persistent, I feel certain you can get your solution working fine.
Plate Solve - Solution to failures
My CCD images have plenty, but often there just isn't many at all visible in TheSky, presumably why it can't plate solve.
(won't be for a few nights, thunderstorms forcast).
