47 replies to this topic

[DAVIDG]

 Here is the optical diagram from Celestron own literature showing the curve facing forward , opposite of the Wiki  example http://geogdata.csun...erature1972.pdf

 There are many examples showing it either way. Optical raytracing shows that it makes no difference. The odds are that the reason why the image improved was from some other factor and most likely the optical alignment and centering of the corrector and secondary since these will have a very large effect on image quality.

   Also go look up any of the articles published in Sky and Tel or any other reference on making Schmidt plates and/or building Schmidt Cass telescope and you  will not find any statements saying the the corrector must be positioned with the Schmidt curve in a certain direction. If it was important to have it placed in a defined orientation it would have been strongly pointed out.

 

              - Dave

 

 

Edited by DAVIDG, 05 September 2014 - 09:31 AM.

[hbastro]

Not to add fuel to this discussion but here are Zemax model results for the Schmidt corrector plate correction out and correction in. Out of respect this model is akin to the proprietary Celestron design but not exact... The A4 Zemax term is AD in Oslo. It is unlikely that the A6-AE term is present given the fidelity of the fixtures and metrology used in corrector plate manufacture...

Attached Thumbnails

• 
• 

Edited by hbastro, 05 September 2014 - 12:10 PM.

[N_DD]

Hi,

 

when using a diffractive element with a planar surface at infinite conjugate ratio (i. e. either the object or the image plane is at infinity) then it is more efficient to have the curved surface facing the plane which is at infinity. This is because otherwise there will be a planar waveform (the one coming from or going to infinity, where all the rays are parallel) arriving on a planar optical surface: the rays won't be diffracted and will only experience a phase change (i. e. no change in their propagation angle), which means "wasting" corrective power (all the hard work of bending the light will be done by the curved surface). Instead, if the planar wavefront arrives on a curved surface, the work will be "shared" between this and the planar surface, since the rays will experience diffraction at both surfaces (planar wavefront arriving at the curved surface and non-planar wavefront after the curved surface arriving at the planar surface).

That said, the very large curvature radii typically found in SCT correctors with respect to their thickness will result in very small differences when used reversed. As other have already pointed out, IMHO in "real life" SCT the matching of the corrector to the primary and the relative orientation of the secondary to both the corrector and the primary mirror have far larger effect on image quality!

 

Nico

[hbastro]

interesting narrative but there are no diffractive surfaces in the Celestron SCT,s... but true construction errors and chromatic errors are far worse than orientation of correction...

[Starman1]

In the repair department at Scope City, we ran into many SCTs that had the corrector plates improperly seated in the aluminum cells.

Foam spacers under the correctors were sometimes partially missing, or the spacers around the edges (Celestron used cork, and Meade used plastic)

were often misplaced so the correctors were off-center.

Additionally, we found secondaries that were loose in the correctors and, shudder, secondaries loose enough to be almost falling off

(especially some of those held in with tape).

They were brought to us because they weren't performing well, not specifically because the optics were dirty

(though the story about the C8 used as a bong was a good one--it explained the brown goo all over the inside).

Every one of them was returned with the correctors centered and seated, the secondaries centered (and secure),

and all performed excellently in the artificial star tests, giving really good star images at ~300X.

We always paid attention to lining up the registration marks on the perimeters of the correctors, but I don't think we once saw one that

had the corrector flipped.

 

My first question would be: "Flat on one side and figured on the other?  I though they were figured on both sides.  They aren't?"

[GlennLeDrew]

Don,

The flat side is certainly figured to an optically good, smooth surface, but the usual practice is to aspherize just the one face.

[hbastro]

look at posting #7 by Aurora DavidG, there is the answer...  and a nice test plate fringe set of the corrected side...

[DAVIDG]

You can read Johnson's patent how they are made here http://www.google.co...tents/US3889431 The correction is only placed on one side ofthe corrector for the telescopes.  As Glenn said and also the patent states the flat side needs to be optical smooth but not optically flat. The patent states the flat surface needs to be 1/2 fringe per inch.  

   Celestron did place the correction on both side of the corrector for their Schmidt cameras. Since the camera is much faster in F-ratio the Schmidt curve would need to be deeper which would require thicker glass or it would break under the vacuum. By putting the curve on both sides they could still use the thinner glass and vacuum technique to make it.

 

                  - Dave

[MKV]

 

"wasting" corrective power (all the hard work of bending the light will be done by the curved surface)" -- Nico

Nico, the "work" done by the Schmidt plate has to do with the optical path length, or optical path difference (OPD). In other words, some parts of the wavefront travel a longer distance then others so that, in a perfect Schmidt system, the net OPD at the focus is zero. That path length is independent of the orientation of the plate, but not of the corrector to mirror separation. If you flip the corrector around, in theory you have to adjust that separation by L(n-1), where L is the thickness of the corrector substrate, and n is its refractive index.  In reality that difference is so insignificant that it's irrelevant. It's a distinction without a difference.

 

Mladen

[N_DD]

Mladen,

 

yes, we both agree here: the Schmidt corrector modifies the wavefront OPD. However, changing a wavefront OPD cannot be achieved without actually changing it from planar to something else, therefore changing the directions the rays wil travel at. In a first approximation it is possible to assume that the rays still travel along the original direction (no refraction), but just change their phases. This holds very well for the typical curvature radii of SCT, and you would be perfectly right if the corrector would have two planar surfaces (moreover, the thickness "L" is not well defined for a surface with a curved profile!).

 

And yes, in practice I would also not expect any noticeable degradation in the optical performances of an SCT with the corrector flipped: what people do see, I think is the result of "un-matching" the optics, that were put together "by hand" at the factory.

 

hbastro,

 

yeah, stupid me!    I meant "refractive", not "diffractive"!!!

 

Nico

[MKV]

 

However, changing a wavefront OPD cannot be achieved without actually changing it from planar to something else, therefore changing the directions the rays wil travel at -- Nico

Nico, this is not rue. The fact that on-axis parallel bundles travel trough a different mediun changed the path length. The aspheric profile makes that path length shorter or longer for some sections of the ray bundle. Off-axis parallel bundles will travel different distances as well, proportional to their incidence angle.

 

Refraction plays a minor role in a Schmidt corrector (it is a thin, practically zero-power optic). That's why it doesn't matter which way the aspheric curve is pointing. No matter how you turn it around, the crux is the length of the optical path from the last surface to the focal point. 

 

Simply put, the length of the marginal ray to the paraxial ray must (in theory) be the same. The profile determines which ray bundles will travel longer, and which shorter paths to achieve that. Of course, these differences are not visible to the naked eye because the paths vary in fractions of a wavelength.

 

When you flip the corrector around, the only thing that has to be adjusted, theoretically speaking, is the separation between the corrector and the primary mirror by the amount d(n-1), where d is the plate thickness and n the refractive index. IOW, if the corrector to primary separation L = 1000 mm, the corrector plate thickness d = 5 mm, and the refractive index n = 1.5, flipping the corrector around to bring the aspheric portion to the front will require an adjustment of the separation L' = 100-(5*[1.5-1]) or 975 mm.

 

Mladen

[N_DD]

Mladen, we are saying the same thing: if the incidence angle is small (wavefront parallel to the optical surface, i. e. rays perpendicular to it), then there is practically no change in the direction the rays will travel at, and the major contribution is OPD change as you said. I was just pointing out that this is true only if the incidence angle is small over all the wavefront, which is the case for a corrector and its shallow curve (with respect to its thickness).

 

Nico

[Michael Miles]

Hey guys:

 

Great discussion - theoretical and otherwise.  Here are my takeaways:

 

1) Turning the corrector around did seem to have a noticeable effect.

2) I'll check to see which side is the flat side of the corrector

3) I'll mount it as originally mounted in the original tube and make sure it's centered

 

If I still have the problems,  I'll try turning it around.

 

More as the experiment develops,

 

Michael

[DAVIDG]

 Mike,

   The issue continues to be,  is the improvement your seeing  from the schmidt curve being on different side or is it in the process of reversing things the scope was better aligned ? Theory saids the placement of the Schmidt curve makes no real difference, theory also will show that misalignment the secondary does have  larger effect on image quality. Also there is the possibility the schmidt plate doesn't have the classic configuration and with a close to optically flat side and the 4th order curve of the other side. The variables needed to be  independently tested before the theory is tossed out and the improvement in image quality is from which side the curve is on. 

   For example I own a vintage orange C-5 and when I got it the image wasn't great. I tried to collimated it but thing were not working like they should. I discovered that the hole in the corrector plate was drill off center by about 1/8" and to compensate for that the secondary was glued to the holder off centered as well. So instead of following the typical alignment procedure which was to center the secondary holder in the hole in the corrector I had determine the mechanical offsets for the secondary holder and also to be sure that the secondary was rotated into the correct positioned in relationship to the off set hole in the corrector to get things aligned.  So when people are flipping correctors around and then using factory alignment marks as references,  maybe the real reason why the image is improving is the  mechanical alignment of the optics to compensate for errors in manufacturing vs were the schmidt curve is located.

 

                    - Dave

[MKV]

Mladen, we are saying the same thing: if the incidence angle is small (wavefront parallel to the optical surface, i. e. rays perpendicular to it), then there is practically no change in the direction the rays will travel at, and the major contribution is OPD change as you said. I was just pointing out that this is true only if the incidence angle is small over all the wavefront, which is the case for a corrector and its shallow curve (with respect to its thickness).

 

Nico

Hi Nico,I am sorry if I misunderstood you, but you seem to continue to think of waves in terms of vectors (or rays), and describe the Schmidt corrector as an optic with a significant refractive power, when it's actually a completely afocal, zero-power device.

 

For example, in #35 you write (my empahsis) "...changing a wavefront OPD cannot be achieved without actually changing it from planar to something else, therefore changing the directions the rays will travel at".

 

Plane wave bundles striking a Schmidt corrector continue to travel in the same direction, only deformed OPD-wise (see diagram below). A plane wavefront entering a Schmidt corrector at usual angles (i.e. ±3°) continues to travel in the same direction upon existing, except it is no longer plane but spatially "deformed" or "retarded", with some section of the wave "crest" being ahead or behind the rest some distance L'/unit time. If this is how you think of it, then we do indeed are saying the same thing.

 

regards,

Mladen

Attached Thumbnails

• 

Edited by MKV, 07 September 2014 - 04:08 PM.

[N_DD]

Hi Mladen,

 

thanks for the drawing: the standard way of conceiving optical "rays" is that they are perpendicular to the wavefront at every point. Therefore, the axial ray is depicted correctly in the drawing, but if you were to draw other rays, parallel to the axial one but off-center, they will be parallel to each other before the corrector, but that won't be anymore the case after they leave the back surface (in the deformed wave). However, since the corrector is almost flat (as you said it has almost zero refractive power), the deformed wavefront will be almost flat, and the rays will go on mostly parallel to each other. We both agree on the OPD, which at the end is what matter in this case: when the rays will be focused by the mirrors, they will interfere and their phase delay (OPD) will define wether the interference will be constructive or destructive.

So, the wavefront will be almost flat after the corrector, and the rays will have a OPD defined by how much they travel through the glass: I think we both agree! 

 

I hope this clarify my previous messages!

 

Thanks,

 

Nico

[Michael Miles]

 

...I discovered that the hole in the corrector plate was drill off center by about 1/8" and to compensate for that the secondary was glued to the holder off centered as well. ...

                    - Dave

 

 

Hi Dave:

 

I hadn't thought of that possibility - I'll check it the hole centering next time I take the corrector out.  I reassembled the tube with the corrector in the original position (who knows which side is actually facing out) and did a couple of test shots (through the haze, clouds and moon glow), and I still have the problem.  I verified the centering of the hole (not the outside of the corrector) to within 5-thousandths of an inch.  I really am beginning to suspect the problem is a tilt error between the Hyperstar and the camera, and will test that idea out when it stops raining in a couple of days.

 

I tried to do the interference testing, but wasn't able to see any fringes.  For the light source, I used a multi-bulb led flashlight.  The light color was modified by shining the flashlight through my O3 filter, and then I also tried using the filter glass from a green eyepiece filter.  I also tried putting translucent plastic over the flashlight to diffuse the light.  For flats, I tried the green eyepiece filter (figuring it would be fairly flat) on top of the Schmidt corrector and an aluminized Newtonian secondary underneath the corrector.  No interference fringes were seen in any of them.   Any ideas why these didn't work?

 

I've been thinking of building a Focault tester to try things out - does anyone know if that would work for the weird Schmidt shape?

 

Thanks everyone for your help - I'm actually having fun,

 

Michael 

[MKV]

Hello Nico, I think we're on the same page now. BTW, I drew that straight line with an arrowhead to show the direction of wave propagation, not as a vector (or a "ray" perpendicular to the wavefront).

 

cheers,

Mladen

[MKV]

 

"I tried to do the interference testing, but wasn't able to see any fringes.  For the light source, I used a multi-bulb led flashlight" -- Michael

Michael, at a minimum it needs to  be a gas emission light source, and it needs to be pointed at the ceiling (diffuse) . Better yet, try making it as monochrome as possible with colored celphane -- it'll increase the contrast.

 

Mladen

[GlennLeDrew]

Contact interferometric testing requires clean surfaces. The tiniest bit of dust could introduce enough wedge in the air gap to cause the fringes to crowd together into irresolvablity. And always strive to not slide the test plate about, for sleeks and scratches can result. When the fringes are separated by about a centimeter or more, even polychromatic light from ordinary fluorescent fxtures can reveal them, albeit at low contrast.

[MKV]

 

And always strive to not slide the test plate about, for sleeks and scratches can result -- GlennLeDrew

We had a discussion (to put it very mildly) not so long ago on this subject and some experienced opticians, to my surprise (contrary to what all professional literature suggests), said that sliding one glass piece over another is "standard practice" in production shops. They even had an instructional Youtube video of it, but it seems it was subsequently deleted. Makes me wonder why would professional literature advise never to do that, yet production shop workers say they use it all the time!?

 

regards,

Mladen

Edited by MKV, 08 September 2014 - 03:02 PM.

[GlennLeDrew]

After numerous such tests, with flats and concave test plates, I never liked to let the piece slide about. And my much more experienced employer stressed to remove/replace when requiring to sample elsewhere on a larger piece being tested. In a non air-scrubbed environment dust is the enemy; assume it's there, ready to sleek and scratch.

 

Furthermore, such differing and varying radii as found on the corrector's aspherized face can permit digging by the edge of a smaller test plate if allowed to slide.

 

During shop roduction, in the stages where more figuring remains, perhaps one can let standards of care relax a bit in the interest of expediency, after some experience. The casual home tester, playing around with finished optics, should exercise extreme caution at all times. Unless sleeks and scratches don't bother one (and which on objectives are really not as injurious as might be feared.)

[MKV]

 

In a non air-scrubbed environment dust is the enemy; assume it's there, ready to sleek and scratch. -- Glenn

 

Of course, which made that instructional video that much more puzzling, since it was made in a shop, basement, or a living room -- hardly an "air scrubbed" environment. The only "air scrubbing" done was to pull a paper towel between the reference and test flats out underneath them, as if forgetting that paper towels collect and shed dust and lint! Anyway, your production practice description makes more sense and is supported by the science behind it.

 

Also, I think your comment about using small flats (i.e. an eyepiece, or camera lens filter) on a Schmidt camera can indeed led to trouble on the aspheric surface side. Besides, testing large surfaces with small surfaces will make the larger surface misleading look flatter.

 

regards,

Mladen