You are not logged in. [Login] Entrance · Main Index · Search · New user · Who's Online FAQ · Calendar

Equipment Discussions >> ATM, Optics and DIY Forum

Pages: 1 | 2 | 3 | 4 | 5 | (show all)
MKV
Carpal Tunnel

Reged: 01/20/11

Re: Baker Reflector Corrector help [Re: Ajohn]
#5673377 - 02/11/13 12:35 AM

Quote:

This is as near as I can get to what Baker seems to have intended - coma a slight problem at a semi angle of 3 degrees covering an 8in dia circle. Going on the MTF curve it looks like it even nearly meets Rayliegh's 1/4 wave limit at 3 degrees. Astigmatism seems to be the main problem there.20in mirror and 17in aperture.

John, what are you doing??? The spots look good consideirng the angle, but in order to cover a 6 degree field your corrector will be 14 inches in diamater and the clear aperture is 17 inches! That's an 82% central obstruction! That leaves 32% of the clear aperture or equivalent to 5.6 inches inch telescope in light gathering power. You have a 20inch f/3 mirror and end up with the light of a 5.6 inch equivalent aperture, working at an effective f/14 or 15, and an 82% CO. Are actually thinking of making one of these?

 Post Extras:
Ajohn
sage

Reged: 12/03/07

Re: Baker Reflector Corrector help [Re: MKV]
#5673697 - 02/11/13 09:39 AM

His basic suggestion for any design is to follow the 17in corrector 20in F5 mirror example with a 50% obstruction = 8.5in in this case and the system comes out at F4.5. The 50% obstruction is a constant for any design as he sees it. He shows a spot diagram over 3 degrees but the scale is too small to compare really. The spots for an F4.5 parabola are shown above it. He points out that there is vignetting at both the mirror and the corrector. The recommended plate size for this one is 4x5in = 6.4in diagonal. All I have done is reproduce his spot diagram.

In a way the spot diagram isn't as it should be for photography. As he mentions the lens spacing has to be adjusted for the most even circles of confusion across the field.

John
-
PS Yes if I can make it work in colour with a 220mm mirror. Field angle TBD. I need some info on Olso's schmidt constants in relationship to actual glass deformation and their relationship to the formulae the vac pan spread sheet uses. That one isn't provided in EDU.
-

 Post Extras:
Ed Jones
Carpal Tunnel

Reged: 04/06/04

Loc: Sin-sin-atti
Re: Baker Reflector Corrector help [Re: Ajohn]
#5674611 - 02/11/13 06:51 PM

Did you consider a Honders/Busack? I'm working on one myself; all spherical optics using ordinary crown glass.

 Post Extras:
MKV
Carpal Tunnel

Reged: 01/20/11

Re: Baker Reflector Corrector help [Re: Ed Jones]

I second that Ed, a Busack or a Honders-Riccardi f/3.7 or thereabouts is ideal.

John, here's a Honders-Riccardi 220 mm f/3.7 configuration. The corrector is a simple biconvex BK7 lens, and the primary is a Mangin mirror. The field corrector is an almost a pcx lens. All surfaces are spherical.

It covers a fully illuminated 44 mm FOV or 3.1 degrees.

I would rather do that than mess with aspheric plates and achromatic doublets. Baker's design was intended for monochromatic photography and it can be corrected extremely well for that purpose, but not for polychromatic applications.

I have no shop experience with aspheric plates, but my understanding is that they are not easy. Thank God there are viable alternatives.

As for your questions about the actual glass deformation for various zones as related to deformation coefficients, OSLO provides that information. The Schmidt vacuum method is well covered in literature and tells you exactly how much of a sag you will need for a desired vertex radius plate.

Edited by MKV (02/12/13 12:09 AM)

 Post Extras:
Ajohn
sage

Reged: 12/03/07

Re: Baker Reflector Corrector help [Re: MKV]

The whole idea is a design based round a conventional parabolic mirror in my case but that one may be of interest at some point. I suspect that it wouldn't be as good at F3 as the Baker anyway and no need for a high quality chunk of BK7 for the mirror. Interesting though as I think Bk7 normal mirrors can be bought. I can also get the image out of the side of the tube. Not entirely happy with the way I can do that at the moment though.

On correctors MKV the vacpan info that is available to me is based round a corrector curve of Z=Ar^2+Br^4 where Z seems to be the actual numeric deformation of the glass. Oslo EDU doesn't offer that is uses a 4,6,8, and 10th power series and so far I haven't seen any indication of what this means physically on the glass. On none Schmidt scopes people get round the problem by using factors of a Schmidt plate's power. I'm not surprised some have problems. They would also have problems if they tried to put the correction on both sides as per the design you posted. Baker's F4.5 systems plate deviates from a sphere by about 5 waves. Trying to split that and centre it on both sides would be a joke. It usually is on all schmidts. There is also the possibility that Baker's plate deviates from a Schmidt plate. Basically it seems to form part of the colour correction. In blue light actually a very obvious choice of Schott glasses hardly makes any differences at all. Chromatically I'm 90% sure it can still be better than the spot diagram.

I still await an answer relating to the meaning of the olso edu coefficients.

As this is the Baker corrector thread I may as well post the spot diagram again. Pointing out that it still isn't as it probably can be. That's part of the other changes I want to make mentioned in the other thread.

John
-

 Post Extras:
MKV
Carpal Tunnel

Reged: 01/20/11

Re: Baker Reflector Corrector help [Re: Ajohn]
#5675509 - 02/12/13 10:09 AM

Quote:

the vacpan info that is available to me is based round a corrector curve of Z=Ar^2+Br^4 where Z seems to be the actual numeric deformation of the glass

Okay, John, the reason I am saying you probably need to read more is because the expression above is the formula for the depth of the aspheric plate at a specific aperture zone radius, r. The coefficients B, C, D, etc...are the 4th, 6th, 8th etc. deformation coefficients. The term A is not a deformation coefficient. This is all very, very basic, and if you don't know that the coefficients you have are the terms in the expression you posted above then you probably need to learn more about the subject.

 Post Extras:
MKV
Carpal Tunnel

Reged: 01/20/11

Re: Baker Reflector Corrector help [Re: Ajohn]

Quote:

I suspect that it wouldn't be as good at F3 as the Baker

And what is your suspicion based on? Have you tried designing an f/3 of Honders-Riccardi type and compared it to the Baker?

I have, and here is a non-optimized 220 mm diameter f/3.01. The PSF shows the more realistic image of a point source at the very edge of the 1.9 degree semifield (22 mm off axis). All this with nothing but spherical surfaces and not a single achormat in the configuration, never mind aspheric plates! Oh, btw, the spectral range is from 405 to 706 nm. Try that with the Baker!

So, I am curious, how did you come to the suspicion that the Baker would be better?

Edited by MKV (02/12/13 10:55 AM)

 Post Extras:
Mike I. Jones
Post Laureate

Reged: 07/02/06

Loc: Fort Worth TX
Re: Baker Reflector Corrector help [Re: MKV]
#5675586 - 02/12/13 10:55 AM

My guess is that he's trying to make use of the 220mm f/3 paraboloid he already has. I agree - dialytes like Honders, Busacks, etc. systems blow away almost anything else for field and spectral coverage in ATM-sized apertures. But John's wanting to work with what he has in hand.

Correct, John?
Mike

 Post Extras:
MKV
Carpal Tunnel

Reged: 01/20/11

Re: Baker Reflector Corrector help [Re: Mike I. Jones]
#5675602 - 02/12/13 11:03 AM

Quote:

My guess is that he's trying to make use of the 220mm f/3 paraboloid he already has. I agree - dialytes like Honders, Busacks, etc. systems blow away almost anything else for field and spectral coverage in ATM-sized apertures. But John's wanting to work with what he has in hand.

That seems like a good enough reason, Mike, but he seems to suggest that he's going with the Baker because he suspects the Honders-Riccardi "wouldn't be as good at F3 as the Baker", which is a totally different reason you give.

Edited by MKV (02/12/13 12:41 PM)

 Post Extras:
Ajohn
sage

Reged: 12/03/07

Re: Baker Reflector Corrector help [Re: MKV]
#5675907 - 02/12/13 01:52 PM

Thanks for answering for me Mike. It may finish up just being in some form of F15 scope but I like the idea of interchangeable ends and as that is a near life long desire want to look at the options. It all needs to go in the car easily too.

I looked at the spots and assumed it would be worse at F3.

Bakers for me has a big advantage - camera can go on the side of the tube for one as the image is further away from the lens than any other design I'm aware of. Not entirely happy with how that could be done at the moment. The other advantage is the parabola but at the moment a solution could use a spherical mirror providing a quality F15 scope is also viable. As an aside I have spent some time looking at that with a spherical mirror but advantages are dubious so far. There looks like there may be a spherical 2ndry solution and that would be an advantage but not a straight DK.

On it's performance it currently isn't any where near optimised for colour. The spots show near red optimisation which is obviously a bad idea. However I will never ever sort that aspect out without relating the usual schmidt aspheric profile formulae to the one oslo uses as currently correcting the plate for sa is very definitely likely to mess up the colour correction else where. As it stands it seems to be capable of sub arc sec photography. Not sure about IR and UV but until the colour correction is sorted out there isn't much point in looking. I wont be taking the filters of the cameras ccd either.

For a laugh I just let 3 of oslo's optimise routines try and sort SA out on something else from a flat surface. The results were awful. Hardly surprising really but if I look more closely I may find that they can be forced to work to closer limits on SA. Doing the correction manually as well showed just how messed up colour correction can get via oslo's constants.

I've asked the same question about the asphere constants several times and no answer. Interesting as it seems to be an insurmountable problem using edu.

John
-

 Post Extras:
MKV
Carpal Tunnel

Reged: 01/20/11

Re: Baker Reflector Corrector help [Re: Ajohn]
#5676064 - 02/12/13 03:19 PM

Quote:

For a laugh I just let 3 of oslo's optimise routines try and sort SA out on something else from a flat surface. The results were awful. Hardly surprising really but if I look more closely I may find that they can be forced to work to closer limits on SA. Doing the correction manually as well showed just how messed up colour correction can get via oslo's constants.

I've asked the same question about the asphere constants several times and no answer. Interesting as it seems to be an insurmountable problem using edu.

Yeah John, OSLO.edu is just awful, flawed, even laughable through and through. If you know so much, why do you ask so many questions? Apparently you figured it all out and decided to fail OSLO.edu all on your own.

 Post Extras:
Ajohn
sage

Reged: 12/03/07

Re: Baker Reflector Corrector help [Re: MKV]
#5676095 - 02/12/13 03:35 PM

I am beginning to think you are a rather offensive person MKV. Perhaps you just can't answer the question if that is the case there really isn't any need to be rude. On the other hand on the web Mike for instance and me in other areas are as helpful as we can be who ever it is who asks for help.

One thing for sure you haven't even read the part that you quoted.

John
-

 Post Extras:
Mike I. Jones
Post Laureate

Reged: 07/02/06

Loc: Fort Worth TX
Re: Baker Reflector Corrector help [Re: Ajohn]
#5676204 - 02/12/13 04:27 PM

Let me see if I can explain it.

The "depth" of a figure-of-revolution surface like your aspheric plate varies only with radius.

The depth "z" of a curve can be expressed as

z = (cr²)/[1+SQR(1-(1+K)c²r²)] + Ar^4 + Br^6 + Cr^8 + Dr^10

where c = 1/(vertex radius) and K is the conic constant:

K > 0 oblate spheroid
K = 0 sphere
-1 < K < 0 prolate ellipsoid (like Dall Kirkham primaries)
K = -1 paraboloid
K < -1 hyperboloid (like Ritchey-Chretien or Rosin primaries)

All the A,B,C and D terms above do is change the local curve depth relative to the conic surface.

Enough? More?
Mike

 Post Extras:
MKV
Carpal Tunnel

Reged: 01/20/11

Re: Baker Reflector Corrector help [Re: Ajohn]
#5676446 - 02/12/13 07:09 PM

Quote:

I am beginning to think you are a rather offensive person MKV. Perhaps you just can't answer the question if that is the case there really isn't any need to be rude. On the other hand on the web Mike for instance and me in other areas are as helpful as we can be who ever it is who asks for help. One thing for sure you haven't even read the part that you quoted.

I am sorry you feel that way, John, especially in view of the fact that I gave you several alternate designs and suggested at least three well known books that may help you with the concepts, short of doing your own homework. Also, despite the fact that I gave you specific answers to your question in post #5675509 about the expressions, you still claim, only a couple of posts later that you "asked the same question about the asphere constants several times and no answer." Like I said previously, you can lead a horse to the water but you can make him drink it.

I just find it somewhat over the top for you to trash OSLO.edu as all but useless, and flawed, and laughable, when it is obvious that you don't know some basic concepts, refuse to accept friendly advice how to learn them, and conclude that OSLO doesn't have something or can't do something when in fact it does and it can. It that makes me "offensive", well, excuse me, Sir! I should have known better not to waste my time.

 Post Extras:
Ajohn
sage

Reged: 12/03/07

Re: Baker Reflector Corrector help [Re: Mike I. Jones]
#5676459 - 02/12/13 07:24 PM

Thanks Mike. I suspected it was Z as per what seems to be the more usual formulae but had no way of telling. There never is any info on something as basic as that. It might have been a slope at a zone for all I know.

From what I gather the conic constant for a schmidt plate would be as per a parabola in that formulae?

The usual formulae kicking about for schmidt plates is the other one I mentioned earlier but I have seen it with additional terms. Z=Ar^2+Br^4.

If I understand that one correctly the A and B terms are calculated on the basis of the schmidt that the plate is intended to be used on so that the neutral zone will be at 70%. That leaves me with the problem of either establishing a base curve to use optimisation with that ensures the neutral zone comes out at 70% or using the sliders in a way that does or even checking afterwards which would be tedious. One thing I have notices is that the 1st adjustment of SA can always be done with the first power term on the sliders but that always has to be backed off later. What I'm really asking now is how the oslo string of powers and constants can be handled to come out with the zone at 70%. I have been thinking about MKV's comment about a few rays determine what power of schmidt plate. I might manage that but it isn't something I do every day so may well get it wrong. Bound to really. Then I need to relate the one formulae to the other some how so that it can be entered into oslo edu.

Sorry MKV I missed an earlier post some how. I made a bad choice of words. I realised that the Z in that one was effectively the profile. Thinking about vacpans too so deformation came to mind.

I can't read Mike's post while typing this so may find that the Ar^2 term is the gubbins at the start and the vacpan method only uses 2 terms. To get down to small errors I have always needed more than 2 terms which makes the 2 term aspect more difficult to mentally relate to the one with more terms. I have to ask because there would be no point in guessing so if that is the case I need to know.

John
-

 Post Extras:
MKV
Carpal Tunnel

Reged: 01/20/11

Re: Baker Reflector Corrector help [Re: Ajohn]
#5676474 - 02/12/13 07:41 PM

No hard feelings John. I am writing a reply to you and Mike that may help clear up some things. Ok?

 Post Extras:
Mike I. Jones
Post Laureate

Reged: 07/02/06

Loc: Fort Worth TX
Re: Baker Reflector Corrector help [Re: Ajohn]
#5676499 - 02/12/13 07:54 PM

The conic constant and r² coefficient (if accessible) should ALWAYS be set to zero and frozen during optimization of an even asphere (even powers only) like yours. The base radius and high-order terms are all that are needed to describe and optimize the surface. If you make K variable, the optimizer mushes it in with the higher orders and K often goes wild, like K=+8.25e+89 or K=-1.24e-125 or some stupid value like that. I have found to NEVER let the conic constant vary for a high-order asphere. K is also unstable when the base radius is very long, as in Schmidt plates, and can explode into meaningless numbers as well. Do all the aspherization design with the base radius and even powers of r only.

To confuse matters worse, I've seen people allow the r² aspheric coefficient to vary as well. This has the crazy effect of changing the base radius without it showing up in the lens spreadsheet, since the base radius is already a function of r². A nice feature of OSLO-EDU is that it doesn't even allow you access to the r² coefficient.

A good goal is to achieve the desired performance with the minimum number of high-order terms. Aspherization is sort of like the MacLaurin expansion of sine(theta). Low-order terms have the most influence on plate aberrations, and higher-order terms are often just unnecessarily gilding the lily. If you can pull off a suitable plate design with base radius, r^4 and at most r^6, you're way better off. Raising "r" to the eighth or tenth power introduces severe numerical sensitivity and instability that overshadow any real measureable benefit they bring to image quality. Only extremely fast systems like f/0.9 Schmidts need those higher order terms.

Mike

 Post Extras:
MKV
Carpal Tunnel

Reged: 01/20/11

Re: Baker Reflector Corrector help [Re: Mike I. Jones]

Mike, just so there is no confusion between the expression John mentioned earlier as z = AY^2 + BY^4... and yours

z = (cr²)/[1+SQR(1-(1+K)c²r²)] + Ar^4 + Br^6 + Cr^8 + Dr^10

as both use ABC terms but in a different context.

Yours is the "complete" formula, with the (cr²)/[1+SQR(1-(1+K)c²r²)] being the uiversal sgaitta equation, and A, B, C, D, etc. being the deformation coefficients of the 4th, 6th, 8th, etc. power by which a surface deviates from a conic.

John's expression is simplified as z = Ar^2 + Br4 + Cr^6, which John mentioned ealrier, where Ar^2 is NOT the same "A" as in your expression.

Rather, in this simplified version, the A is the coefficient of the vertex radius (Rv), such that Rv = ½A. The A can be found algebraically by r²*D²/[8*(n-1)*R³), while B = -1/[8*(n-1)*R³]. Here, the R is the radius of curvature of the mirror (or the equivalent radius of curvature of the mirror), r = Yz/Ym (zonal radius divided by the aperture radius), D is the aperture diameter (2Ym), and n is the refractive index of the glass used.

The expression ½A will therefore give an Rv depending where the "neutral zone" is located. The easiest way to find this Rv is using spherical aberration and from that value determine the value of A, and then the Rv. This is particularly helpful in compound telescopes which are OPD-equivalent to some single mirror. The OPD-equivalent mirorr can be found by Schwarzchild equations (coefficient G), or by raytracing, knowing that (1/G)^1/3 is the equivalent (or elative) OPD radius if curvature.

To the best of my knowledge OSLO.edu doesn't provide a way to determine the A coefficient or the Rv automatically, rod eos it ask for the "neutral zone" height. It requires a manual entry and therefore it requires that the operator know how to calculate the Rv.

Automatic optimization in OSLO can also determine the vertex radius but it doesn't allow you to pick the neutral zone. Depending on how the operands are set, this program will look for the best solution to match the operands.

So, this is not exactly simple and straight forward (but let's not forget that OSLO is free!) unless one knows what is needed and how to get it. But it's there. All you really need is to determine where your neutral zone will be, find A and obtain the vertex radius. OSLO will proceed from there and give you all the pertinent information about the plate profile. For example below is a profile of a Schmidt corrector plate of a 200 mm f/3 Schmidt camera in Text Window One (TW1).

As far as I know Schmidt' vacuum formula is somewhere on the net. This is from memory, so don't hold me on it, but I think it's as follows:

Rv = [16*(n-1)*(f/D)³]*D, where D = diameter of corrector, n the index of glass, and f the focal length of the mirror. You can calculate the dagitta of that Rv by using (cr²)/[1+SQR(1-(1+K)c²r²)], as Mike gave above, or simply approximate it with the usual (Y² + z²)/(2*z).

Edited by MKV (02/13/13 11:39 AM)

 Post Extras:
Ajohn
sage

Reged: 12/03/07

Re: Baker Reflector Corrector help [Re: MKV]

Thanks that helps a lot but as expected Oslo wont do it for me.

I've spent a couple of hours looking for info in the help files for the sag table and for the location of the coefficients internally. No luck on either but found the sag table in the menus quickly. Defaults to surface 1 without asking but the command and syntax can be found from there by clicking on the blue output. Mentioned in case some one else is reading this that lacks the very basics.

No information at all on the table content though, Z (Sag) is obvious but there are also NVL,NVK and NVM and I have no idea what these are and have to hope some one else does in physical terms on the glass. I take it that the lowest values of NVL indicate the neutral zone? If it does though does this leave the power of the underlying sphere?

My current original Baker just converted to mm gives this. No good as I suspected.

John
-

 Post Extras:
Mike I. Jones
Post Laureate

Reged: 07/02/06

Loc: Fort Worth TX
Re: Baker Reflector Corrector help [Re: Ajohn]
#5677538 - 02/13/13 11:52 AM

I know next to zero about OSLO-EDU so I can't be of much help. I imagine (hope) the full-up OSLO can handle these kinds of aberration corrections in stride.

With almost 30 years experience in CODE-V (that's where the A,B,C,D nomenclature comes from) and 20+ years with ZEMAX, picking the "pinch point" in the merit function for Schmidt systems is kid stuff.

I did a little personal study years (decades?) ago to see how accurate the Schmidt plate writeup in ATM-3 was. In it, the author states that pinching the colors at the 70% zone will give a plate profile with the least material removal. However, to minimize color, the pinch point is more like at the 86% zone. I found that to be true in a general sense, although it did vary depending on the spectral band, focal ratio and field angles used. The 70% zone was consistenty not the best zone to achromatize to.

In ZEMAX's merit function, this pinch point is easily simulated by REAY (real ray Y-height) operands. I pick a zone on the plate at which all the colors should have zero Y-height at a common focus. I can then quickly change the zone, re-optimize and watch the residual color behavior. Easy Peezy.

I have no idea how to do the same thing in the little student freebie version of OSLO.
Mike

 Post Extras:
Pages: 1 | 2 | 3 | 4 | 5 | (show all)

Extra information
16 registered and 30 anonymous users are browsing this forum.

Moderator:  ausastronomer, richard7, Starman81

Forum Permissions
You cannot start new topics