There has been a lot of chatter recently here and on CN about using the QTCC with the 130GTX and the IMX455.  As has already been discussed, and confirmed by Roland... if you want perfect stars to the corners, this combination of optics and camera is not going to produce it.  The question for many is... just how good will it perform?

I spent a few hours the other night really dialing this setup in with both backspacing and tilt.  I was able to achieve a really decent result.  Corner stars exhibit strong astigmatism but if you crop off the outer 20% of the frame, stars look really good in my opinion. 

I thought I'd pass on what I learned about this process and provide some data so you can evaluate for yourself.  Following the standard rule of thumb for backspacing optimization doesn't work well with this combo, and I suspect its because the overall field is too small for the size of the chip at this resolution.  Generally, to determine if an adjustment in spacing is needed you can simply take three images, one with very slight infocus, one at critical focus and one with very slight outfocus.  If the corner stars are better with the infocus image, you need to add spacing and conversely if corner stars are better with outfocus, you need to reduce your spacing.

The problem that I found was that at the spacing where corner stars were best... the overall image itself suffered from field curvature.  So to make the corners "least bad" it was a compromise with quality for the rest of the image.  I found that optimizing spacing for the CORRECTED FIELD of the scope (for the resolution of the camera) yielded the best results.  So essentially, I ignored the corners and dialed in spacing to get as much of the field, as perfect as possible. 

The way that you do this is through a focus bracket analysis where you can then graph hfd (or hfr) for stars in the corners and the centers of the chip.  You then graph the results and adjust spacing until the apex of the curves for the center lines up with the apex of the curves for the corners.  This graphical representation by the way states the same thing mentioned above.  If the apex of the corners is at a closer focus position (to the left) than the center... you must ADD spacing.  If the apex of the corner curves are to the right of the center curve you must REDUCE spacing. 

To do this you can use either ASTAP or NINA Hocus Focus.  I'm going to focus on NINA simply because the process is automated, and as I am helping more people dial in their spacing and tilt... the NINA app makes it easier for me to explain and setup when I am not at their scope.  What makes this process different from the visual method is that corner analysis is better able to ignore the astigmatism due to averaging with a larger area of the chip.  So it truly lets you optimize for the field the scope is capable of... not just the corners. 

Take a look below for a visual to show what I mean.  The first image is an aberration inspector of a sub where I dialed in spacing using the visual method. At this spacing of 103.5mm the corners were least bad.  I evaluated in 0.1mm increments of spacing, and determined that this was as good as I could get it.  More or less spacing made corner stars look visually poorer.  While these corners dont look too bad, there is a significant flare towards the axis of the image that is more prominent in brighter stars.  Brighter stars are quite aberrated and if you decide to inspect the entire frame you will see what I mean.  You will also see the effect of field curvature as you move away from the corners into the frame. 

Here are the AA tiles:



Here is the graphical plot of the spacing at 103.5mm.  You can see that the corner curves are to the left (infocus) from center.  While visually, the corner stars are the least bad at this spacing it is clear from the graphical analysis that additional spacing is required. 




I decided that since I'm going to crop the corners anyway, I would optimize spacing so that I could get the absolute best corrected field that I could.  I achieved this by dialing in spacing so that the curves for the corners and center from the focus bracket analysis lined up.  This spacing turned out to be 104.4mm.   Stars improved across the frame and the effect of field curvature went away completely for about 80% of the frame.  The slight flare on the on-axis side of stars also was eliminated.  The corners have severe astigmatism, but like I mentioned... I'm cropping them anyway so I just dont care about that.  What I want is excellent stars for the area I intend to keep. 





Now, if you really want to be critical of the analysis you would see that the corner curves are ever so slightly shifted left of center on the 104.4mm spacing analysis.  So I could technically increase spacing and I might try another 0.05mm or 0.1mm to see, but at this point I am already satisfied and any additional improvement would be very incremental.

I'm pretty comfortable that this is just about as good as you will get with the 130GTX+QTCC+IMX455.  Incidentally, Roland indicated in another thread that he had designs for a 0.8x reducer for the GTX.  His analysis says that it would correct for the IMX455.  He also said, that simply cropping the current QTCC to the size of what a 0.8x reducer would be would get you nearly the same result.  My take on this is that Roland has pretty much nailed it on the head.  There is a small resolution cost to using the QTCC over a hypothetical 0.8x reducer, but it's pretty small. 

Here is a link to the subs for the spacing described above.  File names are self-explanatory. 
https://www.dropbox.com/sh/pzj7p3dyki2n55f/AADlNguuisLplPHUkf6j-XRma?dl=0

And finally, I decided to do a test integration.  I only had a few hours the other night so I collected 3.25 hours of luminance with the QTCC.  I was also testing out Russell Croman's new NoiseXterminator tool for Pixinsight.  It's really a game changing noise reduction tool that can be used on both linear and non-linear data.  My first test shows this to be an amazing piece of software.  Russell is also the author of the excellent StarXterminator Plugin as well.

The image below has had a very quick and dirty process with PI. 
-Stretch (stars and DSO separate)
-Noise Reduction (RC Astro)
-A little contrast and star reduction

It is full frame at full resolution and uncropped.  This will give you a really good idea of what the QTCC looks like to the corners with the GTX and IMX455 chip.  You will see that cropping to 80% yields excellent stars to the cropped corners.  I'm planning to use this with my stowaway to dual image.  The stowaway with it's TC and the IMX571 chip produces a fully corrected field that is the EXACT same size as the GTX+QTCC+IMX455 cropped to 80%.  So I'm very excited for this combination!!!

LDN1228 and LBN552 in between Cepheus and Draco.  Only 3.5 hours so far and looking very promising!  Can't wait to get some RGB data and add a few more hours of Lum!

Full resolution version on Astrobin here:  https://www.astrobin.com/full/z92j74/0/

Hi Chris,

Another excellent article! Thank you for sharing. I will be adjusting the backfocus of the 13035FF once I get some clear nights ;), and this is really helpful.

Best,
Wenhan

Excellent Chris! Your pursuit of a logical and consistent approach to this challenge is a service to us all. Thank you for sharing. 

Roberto

Thanks for sharing Chris, this is excellent info. 

Raghu

You are welcome everyone.  I'm glad this stuff is helpful. 

PS-  If Roland wants to make a 0.8x Reducer, I'd be all for it.  Happy to beta test too!  Remarkable equipment, all of it. I cant wait for my Mach 2.