APO Reducer Backfocus change and focuser position with different cameras

Equipment Reducer Focuser
5 replies•212 views
Björn Arnold avatar
Björn Arnold avatar
Hi,

I'm just scratching my head about something and would like to hear your feedback on it. Although I am working with a specific scope, I believe my inquiry is independent of that and so I haven't placed it under the matching equipment.

I have a William Optics ZenithStar 81 doublet and the recommended reducer (Flat 6AIII, despite the name, it has a 0.8 reduction).
I'm using the scope with two cameras: 
- a monochrome CCD with a 1" (i.e. 16mm) sensor diagonal. Filters with thickness in the range of 1mm and 2mm (depending on brand)
- a APS-C sized Canon DSLR

I believe that I've dialed in the backfocus "correctly" (or sufficiently) for both cameras: on-axis stars and stars in the corner are in focus simultaneously (as far as I can tell from the Bahtinov mask) and star shapes look nice.

What confuses me now is: the focuser has a scale (mm unit) and for the smaller CCD, the focuser is at position 21mm and for the DSLR, I am close to 24mm. That's a difference of roughly 3mm. Shouldn't the focuser position be the same for both cameras if backfocus is "identical"?
Of course, the CCD operates with additional filters and probably there's more protective glass in the CCD than in the DSLR. But in my opinion it would just require additional spacers yielding the same "effective" backfocus, wouldn't it?

The only explanation that I have is that the CCD is not as accurately placed as the DSLR since the CCD is smaller and can probably be more off w.r.t. to backfocus than the larger DSLR?
I guess the approximated deviation from a flat field would be proportional to the squared distance from the optical axis?
Since the diagonals have a ratio of about 1.7, the residual error on the small 1" sensor (due to my measurement accuracy) would scale up to nearly 300% residual error on the APS-C.

From raytracing simulations my impression is that small changes in the backfocus position (if we'd just look at getting into focus on-axis) leads to much larger changes in the distance between objective lens(es) and reducer or in other words: position of the focuser. Is that correct?

Björn
Helpful Insightful Respectful Engaging
andrea tasselli avatar
andrea tasselli avatar
Björn,

Backfocus isn't identical because of the different light paths, those filters and the CMOS camera window, which to be compensated would have needed additional spacers ( I seem to understand you didn't add them, right?). As for the position of the CMOS sensor in the astronomical camera vs. that of the DSLR I don't think is the main issue here. Yes, DSLR  sensors need to be exactly placed at the right distance from the lens flange but even here there is a tolerance.

I'm not sure what you mean by "approximated deviation from a flat field would be proportional to the squared distance from the optical axis" but with a FF the field should be, well, pretty flat to the specs for the Reducer/FF.

P.S.: The ~3 mm of backfocus difference is the actual difference in optical path you have between both cameras.
Björn Arnold avatar
Björn Arnold avatar
Hi Andrea,

Regarding spacing due to filters and window: I optimised the backfocus by checking that stars are in focus on-axis and off-axis and that star shapes are good. Therefore, if I could accurately measure the distance from the last optical element of the reducer to the sensor, I would expect that CCD and DSLR are different, compensating for the thickness of the glasses (filter, camera window).

Regarding the second point: I don't know the "residual error" of the stars in the corners (by error I mean everything that deviates from the ideal, which can be shape/roundness or focus in general) as I cannot measure with an arbitrary precision. Therefore, my assumption would be if, for example, the stars on the CCD's corner are defocused by amount x, the "error" would be about 3*x in the corner of an APS-C sensor.

Björn
Helpful Insightful Respectful
andrea tasselli avatar
andrea tasselli avatar
Björn:
Hi Andrea,

Regarding spacing due to filters and window: I optimised the backfocus by checking that stars are in focus on-axis and off-axis and that star shapes are good. Therefore, if I could accurately measure the distance from the last optical element of the reducer to the sensor, I would expect that CCD and DSLR are different, compensating for the thickness of the glasses (filter, camera window).

Regarding the second point: I don't know the "residual error" of the stars in the corners (by error I mean everything that deviates from the ideal, which can be shape/roundness or focus in general) as I cannot measure with an arbitrary precision. Therefore, my assumption would be if, for example, the stars on the CCD's corner are defocused by amount x, the "error" would be about 3*x in the corner of an APS-C sensor.

Björn

Hi Björn,

Not sure I read you correctly here. In my mind the difference recorded by the focuser scale are the actual difference in optical path between the two systems. Given that nowadays all machining operations are made by CNC I'd wager the tolerances are pretty small, of the order of 0.05 mm and the position of the sensor w.r.t. the camera flange for the astro camera be within no more than twice that amount. So nearly all you read from the scale is the actual optical path difference so that you would need to move IN the focuser to compensate for that w.r.t. the DSLR (which has no glass elements ahead), right?

In an "ideal" scenario the field would be euclidean flat so zero residual error. In reality is most likely a toroidal section whose curvature is hard to predict without the precise optical prescription given the nature of the objective lens and reducer/corrector.
Helpful Insightful Respectful
Björn Arnold avatar
Björn Arnold avatar
andrea tasselli:
Not sure I read you correctly here. In my mind the difference recorded by the focuser scale are the actual difference in optical path between the two systems. Given that nowadays all machining operations are made by CNC I'd wager the tolerances are pretty small, of the order of 0.05 mm and the position of the sensor w.r.t. the camera flange for the astro camera be within no more than twice that amount. So nearly all you read from the scale is the actual optical path difference so that you would need to move IN the focuser to compensate for that w.r.t. the DSLR (which has no glass elements ahead), right?


I'm just thinking if the compensation for the additional plane parallel glass elements should matter on the focuser position?
Thought experiment:
Ideal case, no additional glass between sensor and reducer. Let's say BF=100mm. Image is perfect.
Now, let's add a 6mm filter. This will add about 6/3=2mm of optical path. So I'd add 2mm of spacing (through rings or whatever). Physical distance sensor reducer is now 102mm. Then, everyting should be fine again. I'd say I wouldn't have to touch the focuser?

I've doubt on the backfocus of the camera. Talked to StarlightXpress and I've been told 16mm +/-1mm. Reason: the thickness of the peltier elements varies that much that the BF cannot be given more precisely.
andrea tasselli avatar
andrea tasselli avatar
Björn:
andrea tasselli:
Not sure I read you correctly here. In my mind the difference recorded by the focuser scale are the actual difference in optical path between the two systems. Given that nowadays all machining operations are made by CNC I'd wager the tolerances are pretty small, of the order of 0.05 mm and the position of the sensor w.r.t. the camera flange for the astro camera be within no more than twice that amount. So nearly all you read from the scale is the actual optical path difference so that you would need to move IN the focuser to compensate for that w.r.t. the DSLR (which has no glass elements ahead), right?


I'm just thinking if the compensation for the additional plane parallel glass elements should matter on the focuser position?
Thought experiment:
Ideal case, no additional glass between sensor and reducer. Let's say BF=100mm. Image is perfect.
Now, let's add a 6mm filter. This will add about 6/3=2mm of optical path. So I'd add 2mm of spacing (through rings or whatever). Physical distance sensor reducer is now 102mm. Then, everyting should be fine again. I'd say I wouldn't have to touch the focuser?

I've doubt on the backfocus of the camera. Talked to StarlightXpress and I've been told 16mm +/-1mm. Reason: the thickness of the peltier elements varies that much that the BF cannot be given more precisely.

That's correct. You add the spacer(s) to compensate and you wouldn't need touch the focuser (ideally). So SX has changed their spacing? Used to be 17.5mm +/- 0.5...
Related discussions
Article: Sensor Tilt Adjustment
Hi everyone, I would like to share here a methodology that I developed to check tilt in our optical systems. With the arrival of the new CMOS cameras, with those really small pixels, tilt and other issues have become more noticeable. I suffered tilt ...
Explores backfocus changes affecting camera positioning and optical alignment.
Dec 28, 2020
Baader mk3 cc -m48-Nikon
Is anyone successful in the spacing using the baader mk3 cc using the m48 fitting to a T-ring using a f5 200pds so at the minute I have a Tring and 4mm spacer on the cc giving me 13mm the Nikon sensor distance is 46.5 giving me 59.5 baader give a dis...
Discusses backfocus changes with different equipment configurations and cameras.
Apr 17, 2021