Correcting image plane tilt

[david_taurus83]

For a while now I've long suspected that either my 6D has some sensor tilt or its a combination of this plus the addition of various adapters and perhaps focuser sag. There are too many variables to pinpoint exactly the problem! I first noticed the issue on my Altair Starwave 70 EDQ-R flatfield quad (same as WO Star 71) where the left hand and centre stars look good but the right hand side stars looked out of focus. Even on my 'normal' 102mm ED doublet with flattener, it also appears that one side of the image has better correction and star shapes than the other, with the only thing common with both setups is the Canon 6D and the camera rotator/tilt adapter/filter cell holder I use between both scopes for a secure threaded connection. Last night I spent a few hours taking some measurements with the quad refractor. As it has the corrector lens fixed in the OTA, simply focusing with my auto focuser and recording the steps taken should give me an idea of the correct spacing needed across the 6D sensor. Both scopes have the same R&P dual speed focuser, exactly the same as the WO and TS versions. A while ago I physically measured the steps sizes and I calculated that 450 steps with my DSD AF3 focuser is equal to 1mm. My backlash is 0.05mm so 23 steps.

 

I started in the home position with mount tracking and camera in the upright or horizontal position. To rule out any issue with the focuser causing the tilt or a collimation error I also took sets of measurements with the camera rotated 90 degrees with the sensor in the vertical position and also took measurements pointed both east and west. I used the auto focus aid in APT and picked a star on the left, centre and right hand side of the image and ran the focus routine. To rule out any inconsistancy with APT repeated this 2 or 3 times on a different star in each zone. Results as follows:

 

Scope in home position, camera horizontal.

Left side best focus : 752 (step number)

Centre best focus : 735

Right side best focus : 680

The lower the step size is to focus IN and the higher the steps is to focus OUT. So assuming we want best focus in the centre and to extend out across the sensor, the left side needed to focus out so the left is too CLOSE. The opposite is said of the right, the step size is lower and I needed to focus in so it is too FAR and by a bigger margin than the left. Overall left to right for the above the max difference is 72 steps so 0.16mm of tilt.

 

Scope in home position, camera vertical

Left side best focus : 798

Centre best focus : 755

Right side best focus : 645

Left to right 153 steps, 0.34mm tilt

 

Scope pointed NE at 45°, camera horizontal

Left side best focus : 765

Centre best focus : 705

Right side best focus : 631

Left to right 134 steps, 0.29mm tilt

 

Scope pointed NE at 45°, camera vertical

Left side best focus : 738

Centre best focus : 668

Right side best focus : 620

Left to right 118 steps, 0.26mm tilt

 

Scope pointed W at 32°, camera horizontal

Left side best focus : 715

Centre best focus : 675

Right best focus : 616

Left to right 99 steps, 0.22mm tilt

 

Scope pointed W at 32°, camera vertical

Left best focus : 775

Centre best focus : 650

Right best focus : 624

Left to right 151 steps, 0.33mm tilt

 

 

Apart from the first sample and the relationship between left and centre in the last, the results all appear to correlate. There are variables to consider, the 23 steps backlash, possible focuser sag. I noticed in NINA when autofocus is running the V curve is more of a U curve so there does appear to be a generous amount of steps in the critical focus zone either side of perfect. These factors may obviously impact the above test but all samples above show that the left side appears too close and the right side too far away in relation to the centre. An average of 121 steps and 0.27mm tilt from left to right. So the left side needs to be moved outward away from the corrector while leaving the right side as is. What I'm not sure about though is which side to of the tilt adapter do I adjust? Do I open the left side by approx 0.27mm as if you were to view from behind the camera? Or does the camera auto flip left to right due to how a lens works?! Any thoughts on my conclusions?

 

 

[david_taurus83]

Done some more tests last night. I tilted the camera on the left side as initially thought but it didnt appear to make much difference while checking focus again in APT. So I tilted the right hand side and went to about 0.6mm. After taking a few subs it looked a bit better. Not perfect star shapes but they look smaller. I only managed 22 x 120s subs on M27 but they present a nice image for such short integration. This is the full frame with no cropping at all.

 

And the inspection version:

 

Star shapes are still slightly pointy on the edge but perhaps I'm asking for too much with the full frame 6D. I usually crop out edge artifacts and the bottom mirror shadow anyway. From experience they look smaller in the subs where previously I was used to seeing good focus and tight stars in the centre with larger stars towards the edge.

 

I cropped the image down to show how this scope would perform with APSC sized chip and as expected, very good results.

 

For reference, here is an unprocessed image of M101 with slightly enlarged stars on the edges.

 

I cant decide if I've made much of an improvement?!