EdgeHD 11 expected FHWM/results

Seems pretty regular from what I can see (maybe posting unprocessed image at full scale?). 2.8" FWHM seems on the average from an average site.

Hi,

I've had a look. You seem to have a constant elongation across the field (eccentricity 0.74) which might or might not be due to mirror flop. I would need a much shorter exposure to check this, possibly unguided and in fact not even tracking (shoot the NP). As I said, just average for my location but for a dedicated site at altitude a bit on the low side. My advice to you, given the image scale. is to shoot binned 2x2 all the time. That would improve SNR quite considerably while still having a pretty high image scale.

I have C8 EdgeHD with 0.7x reducer, OAG and ASI174MM guide camera. My main camera is ASI294MM which is hardware binned when using 2x2 binning. Very ideal camera for long focal length. My pixel size is about 0,67arcsec/pixel which is close to diffraction and also seeing limits. If it's excellent seeing, I can see airy discs around bright stars.

If it's steady night I get FWHM-values lower than 2,0". But if it's bad seeing it's 3,5-4,0" or more. I don't take luminance frames if FWHM-value is more than 2,5", but it's still good opportunity to take RGB frames.

John Corban:

andrea tasselli:
Hi,

I've had a look. You seem to have a constant elongation across the field (eccentricity 0.74) which might or might not be due to mirror flop. I would need a much shorter exposure to check this, possibly unguided and in fact not even tracking (shoot the NP). As I said, just average for my location but for a dedicated site at altitude a bit on the low side. My advice to you, given the image scale. is to shoot binned 2x2 all the time. That would improve SNR quite considerably while still having a pretty high image scale.

@andrea tasselli As I understand, shooting 2x2 binned with a CMOS camera isn't hardware based and doesn't improve SNR. It's the same as resampling in software.

That is simply incorrect.  Yes, 2x2 binning in a CMOS camera isn’t done on the chip but no matter where you do it, it improves the SNR by a factor of 2.  Transmitting a 16 bit binned pixel may loose a little of that advantage but you can eliminate any question about that by binning the 16 bit pixels in post processing.  I measured this and demonstrated it in one of my AIC talks on long focal length imaging—simply to demonstrate that the math is “real.”  It is an easy measurement that you can do for yourself.

John

John Hayes:

John Corban:

andrea tasselli:
Hi,

I've had a look. You seem to have a constant elongation across the field (eccentricity 0.74) which might or might not be due to mirror flop. I would need a much shorter exposure to check this, possibly unguided and in fact not even tracking (shoot the NP). As I said, just average for my location but for a dedicated site at altitude a bit on the low side. My advice to you, given the image scale. is to shoot binned 2x2 all the time. That would improve SNR quite considerably while still having a pretty high image scale.

@andrea tasselli As I understand, shooting 2x2 binned with a CMOS camera isn't hardware based and doesn't improve SNR. It's the same as resampling in software.

That is simply incorrect.  Yes, 2x2 binning in a CMOS camera isn’t done on the chip but no matter where you do it, it improves the SNR by a factor of 2.  Transmitting a 16 bit binned pixel may loose a little of that advantage but you can eliminate any question about that by binning the 16 bit pixels in post processing.  I measured this and demonstrated it in one of my AIC talks on long focal length imaging—simply to demonstrate that the math is “real.”  It is an easy measurement that you can do for yourself.

John

Thank you, John, for bringing this new-for-me information to light. It’s exciting to know! Much appreciated!! I should search for your AIC talk. . . .
John

andrea tasselli:

John Corban:
Thank you, John, for bringing this new-for-me information to light. It’s exciting to know! Much appreciated!! I should search for your AIC talk. . . .
John

Already brought to you, by me.

Yes, indeed. I had forgotten about that. Good to know the OP got the helpful advice that was needed

I've been starting to dial in my EdgeHD 11 setup at Dark Sky New Mexico and am running into all the things folks warn about with long focal-length setups.  One thing I'm most curious about is to figure out if my setup is yielding reasonable results or if I'm way off in the config:
https://astrob.in/9ox9vj/B/

Links to my post with data, the bottom line is:

• At native focal length, after focusing with a bhatinov and then locking the primary, I am seeing FHWM's close to 10
• When I run the sub through BXT, it comes down significantly to ~5 or so
• I run the Optec Secondary Mirror Focus System and have been really happy with it so far
• I also run the TEMP-est fans on the primary mirror and used them to cool the mirror prior

Seeing wasn't perfect on the night I ran this and I did use a OAG with 174MM as my guide camera.  I will get another chance tomorrow with better seeing and will see if things improve.  Any guidance/tips/advice here - or am I already on the right track/seeing what's expected from this setup?

It looks like you are measuring in pixels here. Pixels are an absolute measure that only means something for the given configured system. Can you convert these to arcseconds? Only then would they really be comparable. An FWHM of 10 seems large to me...ten pixels across the central portion of the peak is quite a lot of pixels. But, if that measure is only say 1.8", then that might be pretty good, at least from a focused star size standpoint. Sampling 10x across the FWHM is going to be WAY oversampled, though, and that most likely will hurt your SNR, so you might be able to benefit from downsampling by a factor of two, or even three, to improve SNR without losing anything meaningful from a resolution standpoint. Again though, would need to know what these are in terms of arcseconds to know for sure.

EDIT:

It also looks like these measures are Lorentzian... Did you manually choose Lorentzian? If so, unless you know for sure that your star profiles ARE INDED lorentzian, this is probably aggressively under-estimating. I would try a Moffat 2 to Moffat 6, and see if you can find the fit that has the lowest MAD. It may be that you do indeed have a lorentzian fit, but its relatively rare. Moffat 4 is a more common fit. In any case, I would try to find the fit with the lowest deviation/error, and use those measures. Then make sure you convert from pixels to arcseconds (but report both.)