Discrepancy between FOV calculator and Astrometry.net

[Guest]

I've been struggling to get All Sky Plate solver to work and I've found that some of the parameters I was using were wrong.

I have a SW 200PDS newtonian which has a nominal focal length of 1000mm. The camera I'm using is a canon 450d which has a sensor of 4272x2848 pixels. I put this information into a web site called FOV Calculator by 12 Dimensional string. Someone on this forum gave me the link. The output from the calculator said pixel size 5.2um.

I uploaded a jpeg to astrometry.net and asked it to plate solve. The result came back and there were some discrepancies. The first is that it gave the focal length of the scope as 910mm instead of 1000mm. I was using a SW coma corrector which is supposed to match the scope and I was under the impression that it didn't change the effective focal length. Or is it that the 1000mm quoted for the scope is just a nominal value?

The second discrepancy is that astrometry.net gave the pixel size as 10.4um which is exactly twice what I was expecting. There was no binning and I have no idea how astrometry.net even works out the pixel size since I didn't give it any information about the sensor. The jpg file was straight from the canon camera and I did notice that in it's properties it said 72DPI which would make the image about 5 feet wide.

Obviously with the focal length wrong the field of view was wrong as well but when I put the 910mm focal length and 10.4um pixel size into ASP - bingo it worked. So I'm not sure where these discrepancies came from.

By the way the astrometry.net plate solving worked really well. All I did was upload the image and off it went.

Cheers

Steve

 

[vlaiv]

This is rather simple.

First - your scope is 1000mm of FL. There is sample to sample variation in focal length, but that is order of few mm. So your scope can be 1003mm or 998mm depending on how the mirror was figured, but overall it is ~1000mm of FL.

If you are using SW coma corrector like this one:

https://www.firstlightoptics.com/coma-correctors/skywatcher-coma-corrector.html

Then you'll notice that it is x0.9 coma corrector - which means that it reduces focal length by factor of x0.9. Actual reduction factor will depend on distance to sensor so if you vary distance to sensor by 1mm - reduction factor may change for few percent.

In any case ~ 1000 * ~ x0.9 = ~ 900mm. 910mm is "correct" result for such combination.

If Astrometry.net solved your image for 10.4µm pixel size - this simply means that uploaded jpeg straight from camera is set on lower resolution than max resolution. Here are possible settings for 450d:

You probably had it on Small / fine or normal. That is roughly the half of original resolution. In fact, when you account for two, you should get pretty good match.

Take raw image from camera and run it on astrometry.net - it should give you best indication of working resolution. Your camera has 5.2µm pixel size 22200 / 4272 = ~5.196623 (sensor has 22.2mm width).

[Guest]

Thank you. I bought the coma corrector from Rother Valley and it said magnification unchanged so I took that to mean it didn't change the focal length. I should have checked further.

The camera was set to Large/Fine. The file size is 4.3MB and shows a resolution of 4272x2848 however I notice that in the jpeg properties there is a field which says "Resolution unit 2". I'm not sure what that means. I save both the jpeg and CF2 file from the camera but I didn't realise astrometry.net would accept CF2 files so I sent the jpeg. I'm still working out how to convert from CF2 to FITS but I think the latest version of APT has a facility.

 

[vlaiv]

2 hours ago, woodblock said:

I bought the coma corrector from Rother Valley and it said magnification unchanged so I took that to mean it didn't change the focal length. I should have checked further.

Well, if it says that magnification is unchanged, then it really should not change focal length. What version of CC do you have?

2 hours ago, woodblock said:

The camera was set to Large/Fine. The file size is 4.3MB and shows a resolution of 4272x2848 however I notice that in the jpeg properties there is a field which says "Resolution unit 2". I'm not sure what that means. I save both the jpeg and CF2 file from the camera but I didn't realise astrometry.net would accept CF2 files so I sent the jpeg. I'm still working out how to convert from CF2 to FITS but I think the latest version of APT has a facility.

Again, that is strange. If it shows resolution of 4272x2848 - that should be actual resolution. Maybe try FitsWork to open raw file (https://www.fitswork.de/software/softw_en.php). I used it to convert CR2 to fits.

Once you convert to fits, then you can do your own astrometric measurement (although crude) - in ImageJ or other software that can measure pixel distance between two points. You identify two bright stars and measure pixel distance between them, then use Stellarium and angle measure tool to measure angular distance between those two stars. Divide the two and you'll have arc seconds per pixel.

Same number (or very close) should be derived from focal length and pixel size.