Aligning the Sensor on a Canon camera

[RobertM]

I thought I'd start a thread on this since it is about to give me some grief.

Having modded my Canon 550D I now find that the sensor is out by more than a whisker so I now need to adjust it's orthogonality.

I plan the following attempts:

1) use a depth micrometer with 0.0005 thou resolution.
2) use the Starlight-Xpress rotation method with a laser to measure the change in angle of the reflected beam.

First problem - the shutter.

Both these methods require that the shutter is open during the process and the screws for adjustment are available.  Tried jamming the shutter open but there is no way to get even a piece of wire close to the mechanism.  So how can I have the shutter open with power off !  Putting the camera into bulb has the right effect but as soon as the power is switched off the shutter closes.  In fact even opening the camera battery door closes the shutter !

Solution:  I found that placing the camera into the mode that allows a manual clean of the sensor disables the battery door sensor   The battery can be removed and this leaves the shutter open   Now I can do the alignment with the electronics removed and sensor screws available - helelujah !

Next installment depends how brave I'm going to be with the depth micrometer

Robert

[Fay]

What a brave man!!!!!

[RobertM]

Right, step 2.

Just did a mockup of the SX camera alignment method with a laser and the beam traverses a circle of about 3" diameter.  I have no cover glass to get in the way so only one reflection to adjust

Tomorrow night I'll take the camera apart again, clean the sensor, and do the adjustment.  It can't be any worse than unusable so there's nothing to gamble (apart from the camera that is !).

If this works out then I'll check the the SX camera - why they don't fix the sensor in the right position in the first place like decent manufacturers is beyond me !

Robert

[MarkS]

A 3" circle.  That sounds a lot.  Over what distance was that?

By the way - I like your tip for keeping the shutter open.  Top tip that one ...

[RobertM]

A 3" circle.  That sounds a lot.  Over what distance was that?

About 7' in old money, that's sensor to white card.

[MarkS]

A 3" circle.  That sounds a lot.  Over what distance was that?

About 7' in old money, that's sensor to white card.

So that's 2 degrees peak to peak which indicates the sensor is 1 degree off orthogonal.

[RobertM]

A bit less than that but it's about right.  I prefer the earlier use of spacers rather than springs to hold the sensor in position, at least that way there can be no flex.  I'll check that tonight too.

Robert

[Mike]

So that's 2 degrees peak to peak which indicates the sensor is 1 degree off orthogonal.

That sounds like a lot.

[MarkS]

That sounds like a lot.

To use it on an F/2 imaging train you'll need to get the orthogonality accuracy to within 1/50 degree or better.  But then you sold off the Hyperstar didn't you

I still need to tackle my own 550D.

Everything has been on hold for far too long ...

Mark

[RobertM]

1/50th of a degree isn't doable as that equates to less than 1 mm movement !  The play in the equipment I'm using is probably at least that; then there is sag in the rotator device too.  On the other hand I don't think such accuracy is necessary especially with the Hyperstar since you have the fine tuning collimation screws for those minor tweeks

I will be happy with sharp stars on my fastest lens which is f/2.8 but I'll get it as close as I can.

Robert

[RobertM]

Well I dropped the camera on the kitchen tile floor but decided to continue anyway.  The circle is now less than 3mm and that's about as accurate as I can get.  Camera switch on was a heart in mouth moment but as soon as I put the battery in the shutter snapped shut ... Phew !  A photo of the end of the garden at f/2.8, though not much of a test, proves its not too far out so if it stays clear then I'll do a star test and post the results.

Robert

[Fay]

Phew you like to live dangerously on a knifes edge, Robert

[RobertM]

Yes Fay, but don't we all seem to be balanced between success and failure in this hobby ...

Well, the sensor is now clean and orthogonal

I did a 9 min sub on the crescent and it doesn't look too bad.  Will post the image tonight with corner crops.

Next up the SX...

Robert

[RobertM]

Well I'm going to have another go.  Although it's much better I obviously is not quite good enough.  This time I've added reinforcing to stop the rotation device from moving up and down.

Robert