chip alignment issue?

[markse68]

Hi all,

I've been fighting severe vignetting so I double checked my collimation with the nifty free collimation circles app "Mire de Collimation" (http://sweiller.free.fr/collimation.html) and i think I can see what the problem is- it's not collimation but rather sensor chip alignment. I think my collimation looks ok.

This is with my asi120 not my main camera but it seems the asi chip is quite off centre with the lens mount (the Asi viewer has circles too that should be accurately centred to the image). I can turn the camera and nothing changes- it always sees the secondary way off centre. And it's definitely not focuser tilt or anything like that.

I think my imaging camera must also be off. Here is a light and a flat form last session.

Trouble is I don't think there's anything I can do about this? Is it a common problem? I can calibrate most of the vignetting out but the top and bottom left corners aren't getting any light at all I think and don't want to calibrate out (my CC is known to vignette anyway and the filters are quite tight on the image circle)

Mark

[MillHey Nebula]

If nothing changes when you rotate the camera, then the sensor chip is not off centre.  Are you sure that there is not tilt somewhere?  It is difficult to tell from the screen shots but I am not convinced that you have uniform focus across the image

[markse68]

I meant nothing changes in the framing- the image rotates but remains offset relative to the frame the same way which I think must mean the chip is offset?  Of course this doesn't prove my imaging camera also has an offset but it's making me wonder if it has. I'd have thought the manufacturers would try to get the sensors pretty accurately centred though.

sorry you'll have to download this avi as I don't have anything to convert it to web friendly version:  2021-05-30-1444_9-CapObj.AVI

Mark

Edited May 30, 2021 by markse68

[alacant]

23 hours ago, markse68 said:

my collimation looks ok

Hi

JTOL.

My advice would be to forget the concentricts, centre spot the main mirror (and if it is already, check it is central) and use a Cheshire sight tube, preferably one with cross hairs, to collimate. Having firstly positioned the secondary mirror, go no further until the cross hairs hit the spot reflection dead centre, after which you can then use the other tools to help tilt the focuser. 

I've never come across a misaligned camera sensor. Having two misaligned....? 

Above all, be sure to read the collimation myths.

Cheers and HTH

Edited May 30, 2021 by alacant

[markse68]

57 minutes ago, alacant said:

I've never come across a misaligned camera sensor. Having two misaligned....? 

In the image above from the ZWO capture sw you can see the inside of the focuser barrel and it’s quite clearly not centred to the frame. Nothing to do with collimation or tilt. I am using the extreme wide angle lens that came with the camera and i guess that could have a tilted optical axis but the rear element is so close to the sensor i don’t think that could explain the misalignment? No i think the ASI120 has a clearly misaligned sensor. 

The trouble with my main camera is it’s a much larger 4/3 sensor and with the CC and filters being somewhat restrictive, any misalignment of that sensor is an issue. But i’ve just thought about it and it has a  bolted on T mount plate and that is attached to a bolted on filter wheel cover so i might have a bit of wiggle room with screw hole tolerances to at least improve it a bit. I’ll have a play.

My mirror is centre spotted and i’ve used combination of the circles with camera to align secondary to focuser tube and laser/cheshire to align primary and i think my collimation is pretty ok.

Mark

[markse68]

I did a bit more experimenting with this yesterday and I think I improved it a tiny bit shifting the various parts of the camera on their screw tolerances but I have found the real issue.

My 1.25" Astronomik filters are clipping the light cone! This is very unfortunate and shouldn't be- the camera specs say they should be fine but they aren't. But they shouldn't- I found out that if I turn the filter wheel a small degree anti-clockwise from where the automatic alignment settles, then I get a good clear frame. But there's no way to adjust this as the alignment is done with an optical sensor that is soldered to the camera motherboard which is fixed snuggly in the camera body and registers laser etched markings on the underside of the wheel. The only way would be to move the sensor but that is in a sealed chamber and I really don't want to touch that.

filter as positioned:

rotated a bit:

So I could either buy some 31mm filters which I really don't want to do, or I am hatching a plan to dismount my filters and 3d print some rings to make them up to 31mm giving a larger aperture and hopefully curing the problem.

The Asi Fits viewer was a bit of a red herring- it really accentuated the light coverage gradient and I found that by turning the camera 90 degrees I could make the hot spot move all over the frame- the focuser is clearly flexing from the weight of the camera but that should all be calibratable I think

Mark

 

 

Edited June 1, 2021 by markse68

[markse68]

just a quick update, i tried 3d printing some filter holders but it was asking too much if our printers accuracy so i had another thought- I unsoldered the optical sensor and resoldered it further up on its pads so that it would cause the wheel to rotate a bit more anti-clockwise and it seems to have cured the vignetting  

[globular]

Brave and we’ll executed solution. 

[markse68]

Cheers Glob, it was that or fork out for a set of bigger filters which i’d prefer to avoid! 

Funnily enough so far all my targets have been globs 😉 It’s nice to see all the colours in them (and they’re easier for a beginner!) 

Mark

[david_taurus83]

What camera is that? Surely there's a way to calibrate the filter wheel position? I'm sure this can be done on stand alone wheels.

[markse68]

15 minutes ago, david_taurus83 said:

What camera is that? Surely there's a way to calibrate the filter wheel position? I'm sure this can be done on stand alone wheels.

Hi David, it’s a QSI583. I searched for such a thing but came up blank- I guess as it’s a factory fitted thing the alignment should be built in. My guess is that as mine had according to seller been recently serviced with replenished argon that the sensor may have been slightly out of perfect alignment on reassembly? That plus the Astronomik filters i have having possibly a smaller aperture than others- smaller than Baaders anyway, caused the issue i was having.

Mark

[globular]

4 hours ago, markse68 said:

Funnily enough so far all my targets have been globs 😉 It’s nice to see all the colours in them

When you get a glob picture you're happy with maybe you could send it my way for use as my avatar? 😇
I'm not an imager and the stock one I'm using is... well.... I'm sure you can do better 😉

[scotty38]

6 minutes ago, globular said:

When you get a glob picture you're happy with maybe you could send it my way for use as my avatar? 😇
I'm not an imager and the stock one I'm using is... well.... I'm sure you can do better 😉

If you're ok using someone else's picture have you looked on Astrobin as I'm sure you'll find hundreds of worthy contenders there?

[markse68]

Turned out that even with the modification I made, some of the filters still hard vignetted unless I screwed the camera body together shifting it over as far as I could against the screw clearance- not ideal and a bit too close for comfort.

Then a friend at work bought a resin 3D printer cheap so I asked him if he'd print me a prototype of my holder. Turns out the resin print was very accurate and fit perfectly

The simplest thing would have been to buy a set of 31mm unmounted filters but these are perfectly good filters! Baader filters would have been fine- they have a much larger clear aperture of 26.5mm vs the 24.4mm of the Astronomiks! These adapters I've made have a 25.7mm aperture and drop the glass 1mm closer to the sensor than either the Astronomiks or Baaders

I should probably paint them black but this'll do for now. I think I really fixed the problem now.

While I was messing around with this I had a thought- I could use the stretched vignetted image of the flats panel to tweak my collimation probably more accurately than any other method (I think?) It represents the light cone from the secondary and assuming the sensor is perfectly centred to the focuser, the vignetted circle should be perfectly centred in the frame. I may have over tweaked it but it's a lot better than it was. I tweaked the secondary until the circle was more centred and then re-adjusted the primary- my cheshire says its bang on too.

Mark

 

Edited July 9, 2021 by markse68