Pointing precision of nova.astrometry.net

[spal...@gmail.com]

Hey,

is there any approximate value for the pointing precision of images uploaded at nova.astrometry.net? For example my LX200 has a pointing precision of +- one arc-minute.

Thanks in advance
Erik

[Dustin Lang]

Hi,

The reference catalogs (USNO-B1, 2MASS) mostly have star positions good to 1 arcsecond or better, but how well we can calibrate your images is another matter!  It depends a lot on how well we can centroid the stars in your images, and how strong the distortion is.  No really simple answer...

cheers,
--dustin

[spal...@gmail.com]

Thanks anyway! Is it possible to connect a calulated point spread function with a certain quality to the centroid of the star? So, if I say, my psf is a good looking peak without aberrations, my pointing precision is at least that of the catalogue. But as you said, hard to say, i'd probably stick to your answer ^^

Regards
Erik

[Dustin Lang]

Hi,

The centroiding precision should be proportional to the flux signal-to-noise divided by the PSF width -- narrower PSFs or brighter stars can be centroided more accurately, assuming no saturation.

There will also be some error due to our matching.  In theory, matching four stars should reduce the error, but only on the position of the center of the matched quadrangle.  Any rotation or scale errors in the matching will result in a WCS solution that is good in the middle of the matched set of stars but drifts away with distance from the center.  Our "tweak" algorithm tries to find additional stars that match and uses those to nail down the solution, but this doesn't always work...

In short, "it's hard to say". :)

--dstn