Stop me reinventing the wheel... - Page 2

multiweb (Marc) · #21 24-10-2016, 08:02 AM

Quote:

Originally Posted by Merlin66

Hmm
I still stand by the analysis I did between the f5 conic surface and a cylinder section.....

Hi Ken, I didn't run the exact numbers but when you slice a cone you get an ellipse. When you slice a cylinder you also get an ellipse. So you can always offset a cylinder with a diameter to match the minor axis of an ellipse generated from a cone. The secondary offset actually does exactly that with the light cone. I suspect primary blanks are also cut from cylindrical bars of glass as it's far more practical?

dave brock · #22 24-10-2016, 08:29 AM

Ken it's totally fair to offset the cylindrical ellipse to match. That's what happens in the telescope and is why the offset required gets smaller as the focal ratio gets larger.

Dave

Merlin66 (Ken) · #23 24-10-2016, 08:53 AM

Guys,
I have no problems with the offset of the secondary mirrors....
The point I'm trying to make is that the shape of the ellipse cut from a conic is not the same shape as the ellipse cut from a cylinder.... they may "look" similar but in fact are not....

I think many secondary mirrors may be cast in a mould and some cut with cylindrical cutters from sheet glass.

dave brock · #24 24-10-2016, 09:16 AM

If they are different then both the reflection of the primary and the outside of the secondary would be not the same (concentric) as viewed from the focuser, particularly in large fast scopes. Your 10% would stick out like a sore thumb. If they do not look the same (circular) then there is a tilt or rotational error present in the secondary. Google image telescope collimation and see for yourself.

Dave

dave brock · #25 24-10-2016, 09:25 AM

Looking at it another way Ken, if you're happy that cutting a cone at a 45° angle gives an ellipse and not an egg shape then why shouldn't it match a cylinder cut at 45° with an appropriate offset?

Dave

Merlin66 (Ken) · #26 24-10-2016, 09:40 AM

Dave, et al,
I repeat the overall shape of the ellipses "look" similar...
But there is a variation (slight) with the superimposed edges...
Marc's CAD comes close - the cylinder section has to be off-set but when this is done, zoom in around the edge - you'll see what I mean.
I'll get the draughting board out and replicate the drawings I did 40 years ago to demonstrate.

julianh72 (Julian) · #27 24-10-2016, 10:31 AM

Merlin66 is right - the ellipses produced by a 45 degree cutting plane through a cylinder and a cone are very similar, but not quite the same.

A bit of simple geometry will show that for the cylinder, the long axis is exactly square root of two (approximately 1.414) times the short axis. E.g. if you slice a cylinder with diameter 100 mm at 45 degrees, the cut face is an ellipse which is 141.4 mm x 100 mm.

When you slice a cone with a 45 degree plane, you create an ellipse which is slightly more elongated (that is, the ratio of the long axis to the short axis is slightly greater than square root of two). The difference is very small for cones with slopes that are practical for Newtonian telescopes (> f/5 say), but real nonetheless.

For those who are playing with 3D CAD to determine the geometry, try modelling a much shorter cone (e.g. f/1 i.e. height to apex = base diameter), and you will see the effect much more clearly. (By my calculation, for an f/1 cone, the ellipse has a ratio of approximately 1.63:1)

Merlin66 (Ken) · #28 24-10-2016, 01:52 PM

Julian,
Thanks for the input....appreciated.

pmrid (Peter) · #29 24-10-2016, 05:42 PM

Hats off to the purists but for those of us with store-bought equipment, I ask the question. - how significant is the offset in real terms. It it light loss only or is there some image distortion effect as well?
Peter

Merlin66 (Ken) · #30 24-10-2016, 05:48 PM

Peter,
IMO it can affect the efficiency/ light gathering by up to 10%.
Never mentioned in the books.......

Merlin66 (Ken) · #31 25-10-2016, 10:13 AM

OK, here's an updated version of my drawing from the 70's
I've used an f1 cone for clarity (Shown in Black)....
The secondary cylindrical mirror (Shown in Red) has the same minor axis dimension as the cone ellipse.
I've offset the cylinder relative to the optical axis to align it to the top edge of the cone ellipse (if set to the same conic offset then the "error" would appear at the top and bottom).
Hope this finally explains the problem.......

julianh72 (Julian) · #32 25-10-2016, 11:59 AM

Quote:

Originally Posted by Merlin66

OK, here's an updated version of my drawing from the 70's
I've used an f1 cone for clarity (Shown in Black)....
The secondary cylindrical mirror (Shown in Red) has the same minor axis dimension as the cone ellipse.
I've offset the cylinder relative to the optical axis to align it to the top edge of the cone ellipse (if set to the same conic offset then the "error" would appear at the top and bottom).
Hope this finally explains the problem.......

Ken,

I'm impressed that you can still remember how to project conic sections and inter-penetrations using the 2D drafting board methods of our youth! (Your projection tallies with what I get from a 3D CAD analysis.)

For what its worth:

You can in fact produce an elliptical secondary with exactly the correct profile from a glass cylinder, once you have calculated the correct aspect ratio (long axis : short axis) for the focal ratio of the primary. All you need to do is cut the glass cylinder at a slightly steeper angle than 45 degrees to get a longer ellipse on the cutting plane, and then rotate it back to a true 45 degree angle for installation in the OTA. (For any given aspect ratio / eccentricity, the ellipses will be "congruent" whether they are cut from a cone or a cylinder.)

The rotation of the "blank" back to 45 degrees would lead to a slight "shadowing" on one side of the secondary, where the cylindrical side of the secondary projects beyond the mirrored surface, but the loss of illuminated light path should be a smaller than if you cut a 45 degree slice from a cylinder.

dave brock · #33 25-10-2016, 03:40 PM

Hi Ken.
Ok, I'll concede that there is a small difference in major axis length.

My first objection though was to the statement:-

"Secondary mirrors are not exactly elliptical mirrors as required by the theory and formulae"

and then:-

"Sorry... an ellipse by definition is a section from a cone...not a section through a cylinder."

A section through a cylinder does satisfy the criteria describing an ellipse.

Also I still have an issue with this:-

"Based on a 100mm diagonal I used in a 12" f5 design the total loss of "efficiency" was about 10% when a cylindrical section was used. "

and:-

"IMO it can affect the efficiency/ light gathering by up to 10%."

While your drawing shows about 10% difference, it is at F/1 not F/5.
I still maintain any light loss at F/5 (or lack of concentricity when collimating) will be virtually indetectable. Why can't I see it in my 20" F/4.5?
It's not mentioned in the books because there's no issue.
If it is an issue in your 12" I suggest it's caused by something else. I've never seen a primary mirror reflection in a well collimated secondary show as anything other than circular.

BTW, if you've used the minor axis (as is conventional) to describe your 12" F/5 diagonal, why did you need a 100mm one? Seems awfully big.

Dave

Merlin66 (Ken) · #34 25-10-2016, 04:03 PM

Dave,
I was incorrect...
I should have said the cylinder 45 degree cut doesn't give the same shaped ellipse as a conical section with the same minor axis.
I don't have the actual secondary sizes (or FOV target) for the ol' 12" available. I do remember when doing the calculations that the shape differences (over and under size) added up to about 10%.

dave brock · #35 25-10-2016, 04:04 PM

Quote:

Originally Posted by pmrid

Hats off to the purists but for those of us with store-bought equipment, I ask the question. - how significant is the offset in real terms. It it light loss only or is there some image distortion effect as well?
Peter

Hi Peter.
To anwser your question, offsetting or not offsetting the secondary does not affect the image quality, resolution wise, at all. It affects the field illumination and to a little degree digital setting circles (by not having the optical axis and tube axis aligned), although most dsc systems like for eg. Argo Navis can adjust for it.
Just be aware that when collimating, not all circles will be concentric in either case. The sillouette of the secondary in the offset case or the primary mirror reflection in the non offset case. The faster the F/ratio, the more apparent it is.

Dave

Merlin66 (Ken) · #36 25-10-2016, 04:15 PM

The Meade f3.6 Schmidt Newtonian I had was a prime example of secondary off-set!

dave brock · #37 25-10-2016, 04:21 PM

Quote:

Originally Posted by Merlin66

I found a 10% difference in effective area - some light loss due to undersize at some points and some light loss due to oversize at other points.

I'm wondering if the issue here was that you hadn't offset the cylinder ellipse to better match the cone ellipse? How did the collimation look in the actual scope? I bet the mirror reflection and secondary outline were both circular.

Dave

Merlin66 (Ken) · #38 25-10-2016, 04:41 PM

Dave,
You're REALLY testing what's left of my memory cells....
I last looked through the Canopus 320 (It was sold to Peter Nelson) back in the early 1980's.
I was primarily doing visual SN search chart review/ corrections for Gregg Thompson (published as "The Supernovae Search Charts and Handbook")

I honestly can't remember any issues with collimation.
The top end of the tube rotated on 800 (or so) 3mm balls to allow ease of use - it was a dream machine with a great Coulter (pre Dobbie) full thickness mirror.