Jump to content


Photo

Anyone have the CO measurements for Meade SCTs?

  • Please log in to reply
59 replies to this topic

#26 PowellAstro

PowellAstro

    Apollo

  • *****
  • Posts: 1094
  • Joined: 14 Oct 2009
  • Loc: Tennessee

Posted 12 August 2014 - 07:19 PM

You are not telling anything new about testing. I have written data reduction and wavefront profile software for the industry and I understand how to test optics. The 14 performs as I stated and you are welcome to send me a PM if you yourself would like to see/test the scope in person, just let me know.



#27 bratislav

bratislav

    Viking 1

  • *****
  • Posts: 518
  • Joined: 07 Sep 2006

Posted 12 August 2014 - 08:47 PM

I don't need to see anything, besides I'm probably 20,000km away from the scope in question. Just show us a 0.99 Strehl refractor design with drastically different Fressnel in & out patterns. The only way I could do something remotely resembling that (patterns are still more alike than different) was by adding sharp & narrow zone (very high order, but very narrow aspheric residuals), that's the only way to keep Strehl at .99.

I doubt very much that AstroPhysics would let one like that out. Their manufacturing process leads naturally to very smooth surfaces. Any residual errors will be of low order.  



#28 PowellAstro

PowellAstro

    Apollo

  • *****
  • Posts: 1094
  • Joined: 14 Oct 2009
  • Loc: Tennessee

Posted 12 August 2014 - 09:25 PM

As I said on one side the rings are clean and sharp and the other side there are no rings to be found. Brightness of the defocused pattern was the same as was the size, its just one side you could see the rings and the other you could not. The instrument had the sharpest view of mars I have seen. He was told by AP when he purchased the unit it was .99 and then again when he sent it back to AP after many years because he noticed oil around the edge of the lens.  The images in the scope are very, very sharp but this has been the results when star testing the unit on many occasions.


  • Asbytec likes this

#29 Dakota1

Dakota1

    Viking 1

  • *****
  • Posts: 620
  • Joined: 18 Oct 2010
  • Loc: West of the Mississippi

Posted 12 August 2014 - 09:43 PM

20,000km away must be at the south pole ? Brain freeze



#30 hudson_yak

hudson_yak

    Surveyor 1

  • -----
  • Posts: 1580
  • Joined: 15 Nov 2007
  • Loc: Hyde Park, NY, USA

Posted 12 August 2014 - 09:56 PM

 

Where the CO% concern is for an object in the center of the field, focused at infinity, and with the assumption that the scope is not being used in a configuration that reduces its aperture in the center of the field (long visual accessory path, R/C, etc.) the simple relative-diameter measurement method seems valid to me.

 

Mike

 

Sure, and world is flat :lol:

If you can't visualize the problem have a read at http://www.telescope.../two-mirror.htm and imagine primary baffle getting longer (the thing marked with 'L' on the diagram). That is more or less how every SCT is designed today, as they strive to have baffle wide enough to have maximal unvignetted field of view (SCTs are used as astrographs too) but still minimize the sky flooding. The end result is front of primary baffle starts to cut into the converging cone, effectively increasing the obstruction.  

But if you can't visualize (or don't want to), it's OK. Differences are usually "only" a few %, so if Celestron/Meade tells you the obstruction is 33% (but in fact it is 38%) it matters very little. MTF has already suffered quite a bit as you crossed the 30% mark ...

 

 

Ok, thanks. I finally got which cone you were talking about here.

 

Using the laser-through-the-eyepiece projection method for determining true aperture, I found the CO appeared to gain between 1 and 2mm, the shadow was close to 78mm so it's gone from 37 to a little over 38%. I can't tell if the cause is the OD of the inner end of the primary baffle but will assume it is.

 

Another observation, the same sort of components added to the visual train that can decrease the true aperture due to the ID of the inner end of the primary baffle encroaching also decrease the OD effect of enlarging the CO. After putting the R/C in and refocusing to infinity (thus moving the mirror forward) I found the CO shadow back to around 76mm.

 

Mike



#31 bratislav

bratislav

    Viking 1

  • *****
  • Posts: 518
  • Joined: 07 Sep 2006

Posted 12 August 2014 - 10:58 PM


 


 

Ok, thanks. I finally got which cone you were talking about here.

 

Using the laser-through-the-eyepiece projection method for determining true aperture, I found the CO appeared to gain between 1 and 2mm, the shadow was close to 78mm so it's gone from 37 to a little over 38%. I can't tell if the cause is the OD of the inner end of the primary baffle but will assume it is.

 

Another observation, the same sort of components added to the visual train that can decrease the true aperture due to the ID of the inner end of the primary baffle encroaching also decrease the OD effect of enlarging the CO. After putting the R/C in and refocusing to infinity (thus moving the mirror forward) I found the CO shadow back to around 76mm.

 

Mike

 

 

 

Yes, as you move primary forward, you will get to clear the primary baffle at some point. Unfortunately, as you have found out, by then the secondary baffle starts to cut into the beam from the outside, reducing the aperture and increasing the CO again. 

I'm not sure what RC's you are talking about, but if it is a current Chinese/Taiwanese model, they tend to have very small circle of unvignetted area (10" model is only about 10mm), which helps keeping the CO relatively sane. (they are still quite large)

It is hard to have large unvignetted field and small CO in Cassegrain designs. I have simply accepted that if a telescope is to be used solely for astrophotography, CO upwards of 50% is just a fact of life. 



#32 bratislav

bratislav

    Viking 1

  • *****
  • Posts: 518
  • Joined: 07 Sep 2006

Posted 12 August 2014 - 11:25 PM

As I said on one side the rings are clean and sharp and the other side there are no rings to be found. Brightness of the defocused pattern was the same as was the size, its just one side you could see the rings and the other you could not. The instrument had the sharpest view of mars I have seen. He was told by AP when he purchased the unit it was .99 and then again when he sent it back to AP after many years because he noticed oil around the edge of the lens.  The images in the scope are very, very sharp but this has been the results when star testing the unit on many occasions.

 

You keep skipping the vital piece of data : is this with a narrowband green filter? If yes, then someone is twisting the truth big time. There is no way an APO with Strehl of 0.99 will form Fressnel patterns like you described ("one side you could see the rings and the other you could not".). I have modelled just about every possible combination of defects in Zemax and it simply does not exist. Not if aberrations are kept to level of 0.99 Strehl. 

 

If you are using white light, then yes, I can see how it is possible.  But then we come to the original question - why would anyone star test a scope known  for spherochromatism in polychromatic light?

 

I would further suggest using Roddier to measure a true value of Strehl for that scope. In green light, of course! Then you don't have to trust anyone, and you will actually have hard data to show as a bonus.

I test all my scopes this way (as well as Foucault,  or interferometry, if possible). For example, here's my own 4" APO with Strehl of "only" 0.98, in green light. Yes, Fresnel patterns are somewhat different - center gets "hot" outside of focus (exaggerated by software here, visually they are much more alike), but rings are quite visible and even on both sides.

Attached Files

  • Attached File  102.JPG   76.7KB   0 downloads

Edited by bratislav, 12 August 2014 - 11:28 PM.


#33 freestar8n

freestar8n

    Soyuz

  • *****
  • Posts: 3923
  • Joined: 12 Oct 2007

Posted 13 August 2014 - 02:35 AM

If the primary baffle cuts into the cone - it would reduce the entrance pupil size - but based on measurements of celestron sct's I don't see any evidence this happens unless the backfocus is made very large.  The situation for a typical celestron sct is depicted in this thread:  http://www.cloudynig...le#entry6090545  .

 

If you have a situation where the limiting aperture is in one part of a complex optical system, and a different aperture blocks the central region in another part - then the CO measured in the entrance pupil can be different from the CO measured in the exit pupil.  Also, the impact of the central obstruction on the image would be different since it breaks the standard assumption that you have a single pupil at one point in image or object space that is producing the image.  So you need to clarify what you mean by "CO" in the first place.

 

An example of this is a maksutov, where there may be a wide conical baffle flaring out behind the secondary spot.  It will have a large physical size - but it is reduced by the meniscus - as seen from object space.  But from image space it is large compared to the meniscus and the exit pupil does not see the reduction that object space sees.

 

People often talk about different ways to measure the entrance pupil size exactly, but getting accuracy to a few percent - particularly when the aperture stop may be internal - has a lot of difficulty - and the ISO recommended procedure is to use a traveling microscope on a linear stage looking in.  This would work both for the entrance pupil diameter and for the central obstruction - as seen from object space.  But for something like the mak - it may be different from the image space result.

 

As for visual star testing or Roddier - people have star tested sct's for decades and described the Airy pattern - but they almost never mention trefoil astigmatism, which is something I see quite often with mass produced cat optics - both with MetaGuide using a real star, and in lab tests such as Rohr's.  This is why I prefer recorded data rather than visual assessments of performance.  If the star appearance had been faithfully observed - trefoil would have been commented on up front.

 

On the other hand, my experience with Roddier has not been consistent - though maybe it has been improved.  It isn't trivial to deduce the wavefront accurately from out of focus star images, even though there is a published method for it that worked with a space telescope and no atmosphere.

 

In summary - for the CO of sct's, where the schmidt corrector has very little power, I am not aware of any that have an internal stop or baffle that limits the entrance pupil size - except for the possibility of a slightly undersized mirror.  The size of the central obstruction can be measured accurately by taking it out and measuring the baffle with a micrometer.  This would not work for a maksutov because the meniscus makes it look smaller than it is physically.

 

Frank

 

[Edit]  OK - I now realize my figure/animation in the other thread doesn't show the possible clipping of the CO by the baffle, so I will need to look at that to check.  But if the baffle does clip it - then it is definitely a case where the pupils formed by the aperture stop and the CO are at different locations in image and object space - so the measured CO will be different.  Also - the predicted impact of the CO on the MTF and optical performance will not be the same as if both were due to the same physical stop and location.  In that case, neither "CO" will behave as expected by normal imaging theory.


Edited by freestar8n, 13 August 2014 - 03:24 AM.


#34 bratislav

bratislav

    Viking 1

  • *****
  • Posts: 518
  • Joined: 07 Sep 2006

Posted 13 August 2014 - 03:33 AM

Frank, there is a mass of evidence that SCTs obstruct internally (by front of the primary baffle). Have a look at foucaultgrams and interferograms posted on Rohr's site - they invariably ALL show larger obstruction than Celestron (or Meade) publishes.

 

By secondary alone, my own C11 CO is 34%. Yet, when I defocus my C11 I measure 37.6%. And this is at both closer as well as further defocus point around the prescribed BFL (4" behind the backplate). The same 37.6% CO is seen in Foucault images (double pass against a flat or in collimated beam against a Newtonian with smaller secondary).

How do you explain that? 

 

Also pay close attention to weasel-words used by Celestron to describe CO :
"Secondary Mirror Obstruction 3.75 in (95 mm)
Secondary Mirror Obstruction by Diameter 34%"

 

No mention of Central Obstruction - they only say "Secondary Mirror Obstruction". 

 

In any case, it is trivially easy to measure CO by imaging the far out of focus image of a star. I encourage everyone to do it - any camera will suffice. It really is quite easy and fast. And precise. 

 

BTW, Roddier reports trefoil (and higher) astigmatism just fine. Check Zernike 9 and 10 printouts here. They are definitely not zero when I measured my C11.

 

PS I don't think wavefront cares where CO is introduced. After all, in normal Newtonian entrance pupil is defined by primary, while obstruction is sometimes meters away at secondary. Does light care? I don't think so.

Attached Files


Edited by bratislav, 13 August 2014 - 04:16 AM.


#35 freestar8n

freestar8n

    Soyuz

  • *****
  • Posts: 3923
  • Joined: 12 Oct 2007

Posted 13 August 2014 - 06:41 AM

As I said - it may be possible the baffle is enhancing the "CO" - but I will check by ray trace.  Measuring any of this stuff from the front or the back will be problematic - so I don't place much weight on various flashlight and other methods that have been used.

 

The Roddier method is particularly sensitive to the behavior of the wavefront near the boundaries - and having the obstruction at a different location in image space will alter the wavefront behavior around that obstruction.  Like most of this stuff including MTF, it assumes a single pupil out there with a wavefront emanating from it - but that isn't the case if the pupil emits a wavefront for some distance, and it is then clipped by an obstruction at a different location, and then radiates toward the image.

 

There is likely to be a small reduction in the entrance pupil due to clipping at the edge of the primary - but that would also be hard to measure, and it wouldn't impact CO much.

 

So - I will check from my own numbers if the baffle clips anything in celestron sct's.  If it does it is probably a very minor effect - but it is something certainly not accounted for in typical mtf descriptions of how sct's work.

 

The main way I would measure it is by the ISO method - from object space.  No one has done that and reported numbers.  A traveling microscope would directly reveal any clipping of the beam by something inside the scope and not only measure the clipping diameter - but identify the object doing the restriction.

 

Frank



#36 hudson_yak

hudson_yak

    Surveyor 1

  • -----
  • Posts: 1580
  • Joined: 15 Nov 2007
  • Loc: Hyde Park, NY, USA

Posted 13 August 2014 - 06:04 PM

 

 

 

I'm not sure what RC's you are talking about, 

 

Sorry, I was referring to the 6.3 Reducer/Corrector, a common photo/visual accessory for Meade and Celestron SCTs. Used visually, it gets you wider fields with smaller eyepieces than would otherwise be needed. However, it is easy to get into configurations where aperture is reduced significantly due to the forward primary movement required to attain focus. Mine is of early vintage before Meade shortened the focal length of their version.

 

Although Jim never got many of the numbers he was looking for, I've enjoyed this thread. Was already quite familiar with aperture-reduction issues with SCTs but this CO-enhancement aspect is something I had not read about before. So thanks for that.

 

Mike


Edited by hudson_yak, 13 August 2014 - 06:05 PM.


#37 fred1871

fred1871

    Viking 1

  • -----
  • Posts: 871
  • Joined: 22 Mar 2009
  • Loc: Australia

Posted 13 August 2014 - 06:55 PM

deleted by fred1871


Edited by fred1871, 14 August 2014 - 09:20 AM.


#38 DesertRat

DesertRat

    Fly Me to the Moon

  • *****
  • Posts: 5193
  • Joined: 18 Jun 2006
  • Loc: Valley of the Sun

Posted 13 August 2014 - 07:13 PM

Hi Frank,

 

In the classic Roddier algorithm, pupil edge effects are minimized by apodization in Fourier space. In a narrow band surrounding the pupil, radial slopes are set to zero.  In any event, the trick if there is a trick (aside from capturing photometrically accurate images), is to have sufficient defocus to give a good sampling of the pupil, without too much defocus lowering the sensitivity of the test.

 

The most attractive aspect of Roddier is that it represents in situ performance.  In any test proof comes from multiple trials and in the demonstration of repeatability.  Since it is not completely independent from seeing, it does require some effort.  It has been used successfully in earth based observatories, where orientation lookup tables are used to compensate for mirror mount problems and other maladies of large mirrors.

 

The Roddier test can be performed by an amateur astronomer motivated to diagnose trends in performance.  These include things like cooling and backfocus dependency of telescopes.  It is not a test for the casual astronomer.  However help is available at low cost.

 

The operating characteristics of Roddier as implemented in the version used by amateurs has plenty of idiosyncrasies.  It is very picky in the ordering of operations, some options are not fully implemented, etc.  In short its not a commercial quality program by any means.  But it is very useful and its free.

 

Triangular stars?  Yes I've seen them.  I have a collection of near perfect airy patterns along with trefoil and astigmatic corrupted patterns, all taken with the same scope at different times.  I have not Roddier tested a scope under those conditions however, but normally trefoil should be no more than 8-12 nm rms.  More than that the diffraction pattern will be noticeably effected, and even the defocused image will appear triangular.  I have also seen them in a good newtonian with a carefully mounted mirror under no strain.  But it is not always present.

 

Glenn



#39 freestar8n

freestar8n

    Soyuz

  • *****
  • Posts: 3923
  • Joined: 12 Oct 2007

Posted 14 August 2014 - 03:52 AM

Hi Glenn-

 

I have limited experience with Roddier but mainly I was concerned by the lack of repeatability of my results with sct.  From my reading of the method, it was sensitive to how the boundary is handled - and if there is any funny business with the CO I think that would complicate things.

 

Most of what I'm saying about the CO wouldn't matter very much - but when people are talking about a few percent here and there - the details start to matter.  Even for a simple example like the Newtonian mentioned above, the CO seen from object space will be different from the CO seen from image space - because the former is set by the lateral dimensions seen by the incoming parallel beam, and the latter is set by how big the two separate objects: primary mirror and secondary - appear from the focal plane.  Since they are at different apparent distances, their angular sizes will be dependent on that distance.

 

If the CO is being clipped by the sct primary baffle, that would place it at a complicated location relative to the entrance aperture (the sct corrector) because it lies between the secondary and the primary.  This again would make the CO different from object space vs. image space.

 

If the CO is different from expected for a single pupil, and if the CO changes with backfocus - then I think that would appear as spherical aberration to Roddier and it would have problems matching the intra and extra focal images.

 

I will try to tell if the baffle is clipping the CO and if it isn't then most of this is moot.  But if it is - it does complicate the normal diffraction behavior of a system like this - since it can't be modeled as a single pupil radiating a wavefront to a focus point.  It radiates some distance in image space, then gets clipped again, then radiates some more - and when it is clipped there are spatial frequencies lost that would otherwise have made it.

 

Frank



#40 DesertRat

DesertRat

    Fly Me to the Moon

  • *****
  • Posts: 5193
  • Joined: 18 Jun 2006
  • Loc: Valley of the Sun

Posted 14 August 2014 - 11:01 AM

Hi Frank,

 

The CO has only a minor role in Roddier.    In the presence of spherical errors the obstruction can have different apparent sizes in the inner and outer images.  In addition the apparent size can be effected by the masking value used for the zero point.  However the important point is that Roddier does not base its calculations on CO sizes but on the gradations of luminosity across the pupil.

 

Obviously if the masking value is chosen poorly it can effect results.  I recommend the common practice in any measurement process of collecting a number of sets and trying to be consistent with any parameters needed.  The emphasis should be on precision not 2% accuracy or even 5% for that matter.

 

Luminosity gradients work well for spherical and coma.  For astigmatism and its relatives the calculation is based on non circularity.  Here Roddier does tend to underestimate primary terms in the presence of large errors as it struggles to balance the wavefront for minimum rms.

 

These problems exist in any optical testing.  In IF testing the number of fringes and sampling points can effect results.  In addition the methodology and its practicioner can have a significant role. 

 

I don't know who is obsessing with getting accurate <2% error reading for obstruction sizes.  Sounds pointless.

 

Finally, empirical formulas do exist for calculating things like spherical and astigmatism from apparent CO sizes and pupil shapes.  They are basically extensions of Hartmann's formula and can be found in Wilson as well as some of his papers and some ATM publications.  I have developed some of these formulas myself for people who don't want to bother with anything more complex than a calculator.  They are estimates only for large errors and accurate at a 10-20% level.

 

Glenn


Edited by DesertRat, 14 August 2014 - 11:02 AM.


#41 freestar8n

freestar8n

    Soyuz

  • *****
  • Posts: 3923
  • Joined: 12 Oct 2007

Posted 14 August 2014 - 01:40 PM

Well using the measurements I have for a celestron 8" sct I find that the baffle does not clip the central obstruction, so the normal measurement procedure should be ok.  There could still be a slight reduction of the outer diameter if the primary isn't big enough - but it would be a small effect.

 

For the measurements I have, the min allowed sec. obstruction for clipping to happen at the tip of the baffle is at 31%.  So if the sec. obs. is 35% it will clear the baffle.

 

Also, if the primary is moved forward and backfocus is allowed to increase - the baffle is cleared even more.  So increased backfocus at some point will start to clip the outer diameter of the pupil, but the central area will be fine.

 

Here is a drawing of an 8" f/10 sct using the measurements I have, and the blue line shows the path of the ray that just clears the 35% secondary obstruction at the baffle.

 

There is really no reason to make the baffle so long that it would clip the central region - so it would be weird if that were the case.

 

Frank

Attached Files



#42 bratislav

bratislav

    Viking 1

  • *****
  • Posts: 518
  • Joined: 07 Sep 2006

Posted 14 August 2014 - 08:59 PM

Frank,

 

something doesn't seem right. Rohr's Foucalt images of C8 show CO closer to 36% (not 31% based on secondary assembly size):

 

http://rohr.aiax.de/C8_PM-05.jpg

http://rohr.aiax.de/C8S16126_03.jpg

http://rohr.aiax.de/C08-Theis_04.jpg

http://rohr.aiax.de/@C8Kling03.jpg

 

Meade 8" is even worse, at over 40%.

 

http://rohr.aiax.de/M_ARC_203-03.jpg

 

I don't have a C8 nor M8 handy here, so I can't really confirm it either way.

But that is why I advocate measuring CO using far defocused image of a star or Foucault. Everyone has a camera of sorts - very few, if any will have microscope on a precision linear stage.

SCTs, as all Cassegrains have widely conflicting requirements of having as wide as possible unvignetted field of view, and a complete freedom from sky flooding. But as Glenn points out, damage has already been done once you cross ~30% mark so couple percent here of there doesn't matter that much. Still it is useful to know, and if we want to measure it why not.

 

As far as Roddier testing, I have experienced very good to excellent repeatability. Here's 3 measurements of my 4", two done on the same night, third more than a year later, in excellent seeing. Strehl measured 0.975, 0.976 and 0.980. That is only 0.5% error!

Of course, this is somewhat special case - it is very stable, won't flex (it's built like a brick, 4" focuser oversized tube, etc.) and doesn't get affected by seeing much. My larger scopes aren't that repeatable of course, I usually get about 2 to 3% discrepancy, but this is still good enough to give me full confidence in Roddier. Especially as I find it to align well to Foucault measurements (at least spherical components - Foucault is nearly blind to astigmatism, another reason to use Roddier!).

Attached Files


Edited by bratislav, 15 August 2014 - 01:29 AM.


#43 bratislav

bratislav

    Viking 1

  • *****
  • Posts: 518
  • Joined: 07 Sep 2006

Posted 14 August 2014 - 09:12 PM

I think I have found a problem with your diagram, Frank. In CAD programs, you can have tubes of infinitely thin wall; in reality they have to be thick, as they have to be rigid and strong - they carry primary up and down with supposedly no flexure.

Quite thick in some cases (in C11 baffle OD is nearly 10mm larger than ID). 

Can you publish the dimensions? Is this a C8 measured, or merely guessed? 


Edited by bratislav, 14 August 2014 - 10:42 PM.


#44 Jeff B

Jeff B

    Surveyor 1

  • *****
  • Posts: 1988
  • Joined: 30 Dec 2006

Posted 14 August 2014 - 09:28 PM

I'm looking for the CO sizes for the 6", 8" 10" and 12" Meade SCT OTAs.

 

Lemme know if you have 'em.

 

Thanks,

 

Jim

 

Jim:

 

Now just look what you started.  You of all people should know better than to ask a simple straight forward question.  :shameonyou:



#45 DesertRat

DesertRat

    Fly Me to the Moon

  • *****
  • Posts: 5193
  • Joined: 18 Jun 2006
  • Loc: Valley of the Sun

Posted 14 August 2014 - 10:35 PM

This is optics.  A simple straight forward question does not exist in this field!  :grin:

 

Not sure how much accuracy you guys are looking for in these pupil numbers.

 

R.N. Wilson recommends at least 5 measurements be taken.  Angular aberrations and any decentering can influence these measures and additional accuracy follows from taking multiple slices across the image at different orientations.

 

He adds that a turned edge will also result in incorrect values.  This comes from blurred and sharp edges on either side of focus depending on the edge deformation (up or down).  Even outside the caustic zone this will be the case. For most SCT's this is generally not a problem but should be kept in mind. 

 

One could also measure using some threshhold criteria for an edge since here we are looking for a ratio.  Note the error of a ratio is larger than the individual errors in the measures so

 

dObs/Obs  =~ sqrt((dPrim/Prim)^2 + (dSec/Sec)^2)

 

where dPrim and dSec are the estimated errors of the Prim and Sec.

 

Warning:  the image should never be sharpened before these numbers are taken.

 

Glenn

 

Ref:
Wilson, R.N., 1980, ESO (La Silla) Internal Memo “Procedures and formulae for the adjustment of telescopes and analysis of their performance”, 18 June 1980

 

PS - I hope Jim is happy also.  :lol:


Edited by DesertRat, 14 August 2014 - 10:35 PM.


#46 freestar8n

freestar8n

    Soyuz

  • *****
  • Posts: 3923
  • Joined: 12 Oct 2007

Posted 15 August 2014 - 01:58 AM

I think I have found a problem with your diagram, Frank. In CAD programs, you can have tubes of infinitely thin wall; in reality they have to be thick, as they have to be rigid and strong - they carry primary up and down with supposedly no flexure.

 

The drawing is not meant to be exact and it is just using first order properties of the system - and doesn't even have the corrector involved.  The thickness of the wall doesn't play a role because the measurements are internal.  I think the measurements are from Ken Hutchinson's vignetting study from a few years ago.  I have an Edge8 and can't measure the baffle length easily.

 

The two things I know for certain are that the recommended way to measure the entrance pupil accurately is by the ISO standard with a travelling microscope looking in from object space.  I also know that I don't know the geometry used for the Foucault measurement - and if a large defocus was involved then clipping may have been induced that wasn't happening at focus.  From the diagram I show - the ray does come close to the baffle - as it should - and if anything about the testing geometry caused it to clip - it would be an artifact of the test.

 

Finally - although there is an incentive to over-sell the aperture of a scope - I see no reason at all they would go to the effort to make the baffle much longer than it needs to be and therefore increase the CO.  From the dimensions I show - it is pretty much exactly as I would want it to be - with maximum glare prevention but no clipping.  It is possible the true aperture is smaller than the front corrector diameter - but that would be hard to measure.

 

So - my conclusion is there is no clipping of the CO by the baffle, and there are many possible reasons that particular views of the pupil from image space would look different.  One thing to check is that if the baffle really is clipping, then if you focus the scope at infinity with the exact backfocal distance - and then use tracing paper to study the exit pupil image - you should see the outer edge of the pupil form a sharp image at a very different location from the CO - because they would be at different locations in image space.

 

As for whether this stuff is important or interesting or not - for anyone either asking for precise CO numbers on sct's, or clicking on a thread related to that topic - possible clipping and enlargement of the CO by internal systems would certainly be on topic and interesting to them.

 

Frank



#47 Asbytec

Asbytec

    Guy in a furry hat

  • *****
  • Posts: 8209
  • Joined: 08 Aug 2007
  • Loc: La Union, PI

Posted 15 August 2014 - 03:34 AM

Then why don't Celestron, Meade and Intes-Micro use this method to calculate the CO?  Even with the numbers you report, the views through these scopes with smooth, high Strehl optics are excellent.

 

Bill

 

Good question. Do they simply give the specs for the illuminated secondary surface so we can discuss illuminated FOV instead of obstruction affects? Dunno.



#48 bratislav

bratislav

    Viking 1

  • *****
  • Posts: 518
  • Joined: 07 Sep 2006

Posted 15 August 2014 - 04:01 AM

 


 

The drawing is not meant to be exact and it is just using first order properties of the system - and doesn't even have the corrector involved.  The thickness of the wall doesn't play a role because the measurements are internal.

 

 

 

Oh yes, baffle tube thickness plays a vital role. See how thicker baffle (added in red) now clips the rays and increases CO. Is that so difficult to understand?

 

> The two things I know for certain are that the recommended way to measure the entrance pupil accurately is by the ISO standard with a travelling microscope looking in from object space. 

> I also know that I don't know the geometry used for the Foucault measurement - and if a large defocus was involved then clipping may have been induced that wasn't happening at focus.

 

Foucault by default is done at very focus - no defocus involved, let alone a large one. I thought everyone knew that.

 

I'll leave this now, it is clear we don't seem to be talking about the same things. For me, central obstruction is plainly visible from focal plane - by definition. Foucault, Ronchi and interferometric images (as those from Rohr's site) are reality, not simulation, ray trace printouts or CAD drawings. Foucault can also be done on a real star, no large flats are needed, although they make things easier. As I spend time making optics, not just using (or modelling), all these methods are second nature to me and their geometry (double pass, collimation using second scope, real star) are perfectly clear. But not all people can perform Foucault, let alone interferometry, and I start to understand why it can be hard to visualize. It became way off topic and yes, Jeff B - it is time to stop beating the dead horse.

Attached Files


Edited by bratislav, 15 August 2014 - 04:05 AM.


#49 kkokkolis

kkokkolis

    Vanguard

  • *****
  • Moderators
  • Posts: 2137
  • Joined: 23 Sep 2009
  • Loc: Piraeus, Greece

Posted 15 August 2014 - 04:09 AM

For me, the real question is this: do you have a good time with this telescope under the stars?  :flowerred:



#50 freestar8n

freestar8n

    Soyuz

  • *****
  • Posts: 3923
  • Joined: 12 Oct 2007

Posted 15 August 2014 - 04:40 AM

 

 

Oh yes, baffle tube thickness plays a vital role. See how thicker baffle (added in red) now clips the rays and increases CO. Is that so difficult to understand?

 

 

Umm - I know how Foucault testing is normally done - but I don't know how exactly it was set up by Rohr and what primary location/backfocus distance was used.  You see - an sct doesn't have a fixed configuration like a refractor or a Newtonian - and the backfocus can be changed by moving the primary.

 

I am aware of how thickness might play a role - but it still makes no sense to make the baffle so long that it clips.  Furthermore, the tip of the baffle is almost always tapered - so that it nicely matches the cone from the CO and avoids clipping.

 

So I agree there is something odd in all this - but one of the most uncertain things is the way the Foucault image was exactly captured.  And once again - I have never seen the entrance pupil measured from object space by the ISO method.

 

Either way - you bring up a good point that the baffle *might* play a role in the true CO - but I don't see enough evidence that it does.  I just see independent views that are proxies for the real thing, prone to artifact, and lacking details on how exactly they were captured.

 

Here is a marked up image from the EdgeHD white paper - showing a cross section of an EdgeHD8.  I drew extensions of the primary focus - and then drew back from there to intercept the tip of the baffle.  Notice that it exactly meets on the primary where the shadow of the secondary baffle falls.  This is the optimal layout to maximize shielding while avoiding the clipping of the CO.

 

You may think the baffle is actually much longer in an sct for some reason and the drawing is hiding that fact - but all I see is evidence it was sensibly designed and there is no problem at all.  Your conclusion from the Foucault image is likely missing something about how it was captured.

 

Frank

Attached Files


Edited by freestar8n, 15 August 2014 - 04:40 AM.







Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics