Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Is a masked SCT like an APO for viewing Mars?

This topic has been archived. This means that you cannot reply to this topic.
117 replies to this topic

#76 JerryWise

JerryWise

    Cosmos

  • *****
  • Posts: 9,692
  • Joined: 26 Dec 2003

Posted 16 October 2005 - 08:21 PM

Hello Rich. Excellent reply. Just what I have in mind. I figured I better do it. Here is my plan.

I have measured the 9.25 and there is over 80 mm available from the CO to the edge of the corrector. I have sacrificed my last Hartman Mask blank for 8" to 10" OTAs (a metal Martha Stuart stove burner cover, Kmart, 2 for 4 dollars) to this project. I have carefully measured and drilled an 80mm hole off center.

As the 9.25 did the best in the above post, I am using that OTA. It will compete with the Orion ED80 instead of the TAK. I think the Tak performed better due to the extra aperture and how this would factor in on the Paper Curve mentioned earlier. While the ED80 is not a true APO it should be close enough for this. I have both mounted on the Losmandy dual plate and cooling since 5:00pm today.

I am about to walk over now to run some test. In a test this afternoon (just after completing the mask) on a distant target with similar magnifications the 9.25 did indeed appear to have as good an image. Let's see how it fairs on a cluster or two, the moon and mars.

Will post the long awaited answer soon.

#77 snorkler

snorkler

    Aperture Aficionado

  • *****
  • Posts: 11,579
  • Joined: 11 Oct 2004

Posted 16 October 2005 - 08:57 PM

Jerry,

We're all looking forward to your results. BTW, thanks for going to all the work to do your 4 scope comparison earlier.

I wonder whether it's unfair to be testing your worst refractor, the ED80, aginst your best SCT, the C9.25. Still, it's a start and it will answer the question of whether a masked SCT can give APO views. But if the ED80 fails to measure up, the APO owners will surely want to see the Tak 102 go up against the masked C14.

#78 Rich N

Rich N

    Fly Me to the Moon

  • *****
  • Posts: 5,624
  • Joined: 22 Sep 2004

Posted 16 October 2005 - 09:41 PM

"But if the ED80 fails to measure up, the APO owners will surely want to see the Tak 102 go up against the masked C14."
-------------------

Jerry cherry picks the CAT and not the refractor. Maybe Jerry could find someone in his area with a Tak FS-78. However, since Jerry has the FS-102, why not mask the FS-102 down to 80mm and compare it to his C9.25 with an 80mm mask?

Rich

#79 JerryWise

JerryWise

    Cosmos

  • *****
  • Posts: 9,692
  • Joined: 26 Dec 2003

Posted 17 October 2005 - 06:00 AM

Well, it didin't turn out the way it should have but seems to have turned out the way the APO folks have said it would. The ED80 beat out the stopped down 9.25. In marginal conditions (Mars lower in the sky with more atmospheric aberrations) the ED80 actually performed as well as the non-stopped down 9.25.

I did a few images and also wrote the test up in more detail. This work is in another post on this forum. (Thank goodness I didn't use the Tak.) Whew.

#80 snorkler

snorkler

    Aperture Aficionado

  • *****
  • Posts: 11,579
  • Joined: 11 Oct 2004

Posted 17 October 2005 - 07:50 AM

The theories didn't prepare me for these results. The oft-cited reason for APO superiority, the unobstructed light path, doesn't tell me how a mass-produced doublet could equal or outperform a mass-produced mirror assembly with roughly 27 times the light gathering ability. I'm dumbfounded.

#81 jrcrilly

jrcrilly

    Refractor wienie no more

  • *****
  • Administrators
  • Posts: 35,563
  • Joined: 30 Apr 2003

Posted 17 October 2005 - 08:08 AM

The oft-cited reason for APO superiority, the unobstructed light path, doesn't tell me how a mass-produced doublet could equal or outperform a mass-produced mirror assembly with roughly 27 times the light gathering ability.


When the C9.25 is masked, the light-gathering ability of the two telescopes is the same. I find his report that the ED80 did better visually at the same magnification interesting and surprising.

The imaging results are too hard to interpret because of the large difference in image scale and F ratio. It'd be interesting to see the full-aperture M27 shot compared with a masked M27 shot taken with an increased exposure time to compensate for the F ratio change, then an ED80 shot taken with a 4X Barlow to get it into the same F ratio and image scale.

#82 JerryWise

JerryWise

    Cosmos

  • *****
  • Posts: 9,692
  • Joined: 26 Dec 2003

Posted 17 October 2005 - 08:41 AM

I think this goes back to the point of prediction using simulated charts. The charts don't account for focal length, exposures and other variables. The focal length has a tremendous influence in the results. A SN-10 or any other reflector masked down may have a completly different result.

As for the visual comparison on Mars, the conditions were much worse than the test I ran (posted in this thread) the night before. The full aperture 9.25 made the Orion ED80 look very bad. One night later in much worse seeing the 9.25 view declined significantly and the Orion ED80 declined but not as much. On the first night the views of Mars in the 9.25 were stunning. The second test when stopped down the views in both OTAs were much poorer than the previous night. Somehow, I expect seeing may have an influence on visual and imaging results. :grin:

I think the question (not being flippant here) was general. I take it to read "Is a masked SCT stopped down to a similar aperture as an APO going to have the same views of Mars". No qualifiers as to focal length, viewing conditions, EPs, etc. So therefore the test was done with two sample OTAs under identical conditions to answer this question. If qualifiers are added the question changes.

I think it might be good to add I don't have a preference as to how the test comes out. I will say I expected the Celestron 9.25 to beat the Orion ED80. Obviously I have an APO, EDs, SCTs and the SN-10 on the way so my preferences are broad. On this night under these conditions with these particular instruments these were the results. Visually and photographically. Based on results from the previous night I am confident the results could well be different on another night with different seeing.

#83 snorkler

snorkler

    Aperture Aficionado

  • *****
  • Posts: 11,579
  • Joined: 11 Oct 2004

Posted 17 October 2005 - 09:27 AM

I must have been tired when I did my earlier calculations. The unmasked C9.25 has about 9 times the light-gathering ability of the 80 ED, not 27 times. Maybe the ED80's performance isn't as shocking as it first appeared to my muddled mind.

#84 JerryWise

JerryWise

    Cosmos

  • *****
  • Posts: 9,692
  • Joined: 26 Dec 2003

Posted 17 October 2005 - 10:06 AM

In the visual performance of the OTAs, I would like to ask a couple questions of the group.

In both sets of seeing conditions Mars was very bright in the 9.25 XLT OTA. It was much dimmer, of course in the 80ED.

Could the visual effect of a very bright highly detailed object under more scintillation not "wash out" or "white out" sooner than a dimmer view of the same object?

Could reflection (rather than transmission) possibly minutely aberrate the image under these circumstances?


#85 sixela

sixela

    Hubble

  • *****
  • Posts: 15,261
  • Joined: 23 Dec 2004

Posted 17 October 2005 - 06:16 PM

Let's say that miscollimation is 1/20 wave RMS ,

It seems low, doesn't it, until you realise you're talking about an RMS value. That's a *really* large RMS error.

Which leads us to...

The square root of the sum of all the RMS
errors squared comes to 0.076, or 1/13.1 wave RMS combined
error.

Which is huge. Do you have any indication that this is actually typical (well, it certainly is for a scope that isn't cooled down - I would bet my life that the RMS error from thermals dwarves that caused by miscollimation of my scope for the better part of two hours)?

Of course, you do make a valid point - one which no one denied; I think many Newt users all know how much importance thermals and collimation play, and how much easier it is for an APO to perform well.

But we *were* discussing a situation in which their impact was much less than 1/13 lambda RMS at the wavefront.

#86 wh46gs

wh46gs

    Mariner 2

  • -----
  • Posts: 270
  • Joined: 01 Oct 2005

Posted 17 October 2005 - 06:26 PM

The ED80 beat out the stopped down 9.25



That shouldn't be much of a surprise. Scatter, thermal errors and miscollimation are still greater in the SCT.
Also, there is no reason to assume near-uniformity of
optical surfaces in a mass-produced SCT. There are local
deviations on all three elements - corrector, primary and secondary - that become much greater source of error if nearly coinciding with the area selected by masking the aperture. For instance, a local deviation of ~1/10 the area of unmasked aperture, with 0.1RMS wave error over it, will contribute only 0.01 wave RMS to the aperture as a whole.
But if it happen to cover most of the area selected by
a mask, it will contribute nearly the entire error to the
masked aperture. This can happen on either of the three
elements, two of them, all three or none. Considering expected degree of smoothness of an average SCT optics (less than great), this source of error is more likely than not to be significant, at least on average.

Vlad

#87 sixela

sixela

    Hubble

  • *****
  • Posts: 15,261
  • Joined: 23 Dec 2004

Posted 17 October 2005 - 06:31 PM

Well, it didin't turn out the way it should have but seems to have turned out the way the APO folks have said it would. The ED80 beat out the stopped down 9.25.


I would have been surprised to see a different result - after all, if you stop down a C9.25, what factor *would* make it better than a good APO?

The thermals are nastier, the quality of the optical figure is worse, it's got more mirrors so less system transmission, a collimation that's probably not as good, and even if everything were perfect on that C9.25, spot diagrams for an off-axis mask aren't as symmetrical as those for a real unobstructed scope, so why *would* the stopped down C9.25 win?

As for bad seeing: there are conditions in which the convective cells in the atmosphere are roughly 80mm wide; and on those nights, it's fairly possible for an 80mm scope to show as much as you'll ever see in any scope (unless in those brief moments of good seeing).

#88 sixela

sixela

    Hubble

  • *****
  • Posts: 15,261
  • Joined: 23 Dec 2004

Posted 17 October 2005 - 06:33 PM

The theories didn't prepare me for these results. The oft-cited reason for APO superiority, the unobstructed light path, doesn't tell me how a mass-produced doublet could equal or outperform a mass-produced mirror assembly with roughly 27 times the light gathering ability. I'm dumbfounded.


Well, he *did* stop down the C9.25 to 80mm, so that mirror assembly with the mask has roughly 1 times the light gathering ability of an ED80, and actually less given the imperfect reflectivity of the mirrors (this number, BTW, is irrelevant in a discussion about resolution).

#89 wilash

wilash

    Fairy Godmother

  • *****
  • Posts: 5,746
  • Joined: 30 Sep 2003

Posted 17 October 2005 - 06:43 PM

In that experiment, based on the data, the ED80 was resolving 53 line/mm at the retina and the stopped down 9.25 was resolving 75 lines/mm based on an f/34 optical train for the 80 and an f/24 optical train for the SCT. At their best magnification for images, the ED80 was at 158x with an exit pupil of 0.5mm and the SCT, 118x and 0.7mm exit pupil. If the SCT had similar quality optics, there should have been no significant difference between the two. BTW, unmasked, the SCT should have been projecting an image with a resolving power of 212 line/mm on the retina. The SCT was obviously not able to approach the calculated values.

There may be one reason why the SCT did poorly. The optical train is off-axis. This is never the best situation for any optical system. I wonder if a visual difference can be seen in a refractor as an aperture mask is moved from the center to the edge of the aperture. But that probably does not explain everything. But then we are still working with a sample of one.

I don't think the result is that surprising. Photographic lens manufacturers rarely use Catadioptrics. Not just because of the problems of placing an aperture diaphram in them, but also they are always out performed by a lens system, granted, not as simple as the ED80. The larger the system gets, the more difficult it is to make. Considering the size, a 9.25 SCT is very cheap. In optics, you get what you pay for.

This is why I think manufacturers should put thier products on an optical bench and generate real-life MTF plots. Here again, not just a simple on axis plot, but off-axis as well showing radial and tangental results. This would give us the data to show how the different models actually perform. I doubt they will do it as calculated performance is always better then actual performance, especially since they only publish the angular resolution base on aperture size.

#90 wilash

wilash

    Fairy Godmother

  • *****
  • Posts: 5,746
  • Joined: 30 Sep 2003

Posted 17 October 2005 - 06:45 PM

BTW, the average resolving power of the eye is 90 lines/mm.

#91 JerryWise

JerryWise

    Cosmos

  • *****
  • Posts: 9,692
  • Joined: 26 Dec 2003

Posted 17 October 2005 - 09:09 PM


Whew. Now we are getting somewhere. Great post Wilash and Sixela.

Dale, I think we are about ready to say "a Masked down SCT is not like an APO for viewing Mars". (Don't bank on it just yet.)

#92 Guest_**DONOTDELETE**_*

Guest_**DONOTDELETE**_*
  • -----
  • topic starter

Posted 18 October 2005 - 08:23 PM

Yes, it seems that way. I've been following the discussion with great interest.
Dale

#93 snorkler

snorkler

    Aperture Aficionado

  • *****
  • Posts: 11,579
  • Joined: 11 Oct 2004

Posted 18 October 2005 - 08:37 PM

Agreed. This has been one of the best threads I've read on CN. My thanks to everyone who has contributed their knowledge and experience to the discussion. My next scope may be a 4" APO refractor!

#94 wh46gs

wh46gs

    Mariner 2

  • -----
  • Posts: 270
  • Joined: 01 Oct 2005

Posted 19 October 2005 - 08:59 AM

It seems low, doesn't it, until you realise you're talking about an RMS value. That's a *really* large RMS error.



Not really. It is comparable to 1/6 wave p-v of spherical
aberration. For a 16" f/4.5 Newtonian it is caused by 1.3
arcmin misalignment. It is probably better than average for
this kind of instrument in actual use. Smaller apertures
will fare better, larger worse - I was aiming for a rough average.

Which is huge. Do you have any indication that this is actually typical (well, it certainly is for a scope that isn't cooled down - I would bet my life that the RMS error from thermals dwarves that caused by miscollimation of my scope for the better part of two hours)?



It is what it is. Thermals are
certainly much greater, on average, than 1/40 RMS (which is
rather negligible alone - 1/12 wave of spherical aberration
level). I wasn't trying to cut out exact numerical values, rather to illustrate figuring out the final outcome (image quality). 1/40 wave RMS is also unrealistically
low for deformations caused by mirror cell and gravity,
and likely so for the diagonal (which is more sensitive
to local surface deviations than the primary).

In a well executed telescope, most of these errors are
negligible alone. Yet, each does throw little bit of energy
out of the Airy disc, and the combined effect can become
significant. And, there is quite a few less than well
executed telescopes. I don't know what is typical - it sure
does vary significantly - but it
is there. It can be somewhat better than the example shown,
and it can be much worse. If we are talking ~20"
and larger Newtonians, even really well executed telescopes
with great optics will be bellow 1/2 wave s.a. level
most of the time. And 10" (any) aperture will suffer from more than 1/10 wave RMS seeing error alone in as good as
1 arcsec seeing. A 6" apo would have it barely diffraction
limited (high 8 on the Pickering's scale). Most people
seem to have unrealistically optimistic views on typical
operating quality of telescopes in general.

Vlad

#95 wh46gs

wh46gs

    Mariner 2

  • -----
  • Posts: 270
  • Joined: 01 Oct 2005

Posted 19 October 2005 - 09:22 AM

In that experiment, based on the data, the ED80 was resolving 53 line/mm at the retina and the stopped down 9.25 was resolving 75 lines/mm based on an f/34 optical train for the 80 and an f/24 optical train for the SCT.



Hmh...wonder where do these numbers come from?
Eye can only see what is resolved in the focal plane
of the objective. Conventional MTF gives linear cutoff
frequency as 1/LambdaF, which for 0.00055mm Lambda comes
to 1818/F lines per mm (F being the system F#). That
gives 242 lpmm to an f/7.5 system (80mm ED), and 61 lpmm
for an f/30 (approx. C9.25 masked down to 80mm). Sure,
angular resolution is the same.

That is for bright, contrasty lines. For low-contrast lines,
number of lines resolvable will be only about half as many.
Any other factor lowering image contrast will have it further reduced.

Vlad

#96 sixela

sixela

    Hubble

  • *****
  • Posts: 15,261
  • Joined: 23 Dec 2004

Posted 19 October 2005 - 02:47 PM

And 10" (any) aperture will suffer from more than 1/10 wave RMS seeing error alone in as good as
1 arcsec seeing.

1 arcsecond seeing is 1 arcsecond *for long exposure times*. You can, and do, see details for brief periods that beat that seeing, and very easily (and fortunately for photography, planets are bright enough not to require minutes of exposure time).

As for your assessment of the errors of a newtonian *system* (including the mirror cell, angled up, pointing at the diagonal and everything), having seen such a *system* from the knife edge of a Foucault test setup close to the centre of curvature of the entire system (yes, that's a strange thing - a truss dob without the trusses, and the Foucault tester somewhere decidedly odd), I can only say that your expectations about the system wavefront errors are hard to reconcile with what I've *seen*.

Before I fixed an error in the way the secondary was glued, there was an error at the wavefront which was glaringly obvious and *not* swamped by other factors that was smaller than that you indicate would be typical s.a. for the system, and I can assure you that it's now gone, and hasn't been replaced by any other error of a magnitude that is comparable - not in a Foucault test, and not in a star test.

Not that any of your figures are really *widely* off the mark - but exaggerate five contributing factors slightly, and the end result will still be off the mark...

With a star test, it's rather easy to see in good seeing that the dominating error of my scope is equivalent to about lambda/6 to lamba/8 peak to Valley at the wavefront spherical aberration once it's cooled down - and it *does* cool down to the point it's obvious thermals don't contribute enough to make that other error no longer the dominating error.

Not that I have many that many nights in which I can evaluate this - but there are enough nights like that to have made me obviously run out of power at around 500x on Mars, prompting me to buy a shorter TMB supermono...

#97 sixela

sixela

    Hubble

  • *****
  • Posts: 15,261
  • Joined: 23 Dec 2004

Posted 19 October 2005 - 03:00 PM

For a 16" f/4.5 Newtonian it is caused by 1.3
arcmin misalignment. It is probably better than average for
this kind of instrument in actual use.

I don't know what collimation tools you use, but 1.3 arcmin over 1780mm is about 1.34mm, and that's an error that's *glaringly* obvious - with an autocollimator that magnifies that error, it simply would be impossible to miss this even if you tried very hard.

#98 wh46gs

wh46gs

    Mariner 2

  • -----
  • Posts: 270
  • Joined: 01 Oct 2005

Posted 19 October 2005 - 06:45 PM

Poster: sixela
Subject: Re: Is a masked SCT like an APO for viewing Mars?


Quote:
--------------------------------------------------------------------------------

And 10" (any) aperture will suffer from more than 1/10 wave RMS seeing error alone in as good as
1 arcsec seeing.



--------------------------------------------------------------------------------


1 arcsecond seeing is 1 arcsecond *for long exposure times*. You can, and do, see details for brief periods that beat that seeing,



Of course, when it comes to seeing error, we are talking
averages: 1 arcsec seeing implies certain value for
the atmospheric coherence factor that does determine average
*visual* error as well, not only long-exposure error. Actual
error will constantly vary around that value, to the better
and worse, but significant deviations are exponentially less
likely to occur.

As for your assessment of the errors of a newtonian *system* (including the mirror cell, angled up, pointing at the diagonal and everything), having seen such a *system* from the knife edge of a Foucault test setup close to the centre of curvature of the entire system (yes, that's a strange thing - a truss dob without the trusses, and the Foucault tester somewhere decidedly odd), I can only say that your expectations about the system wavefront errors are hard to reconcile with what I've *seen*.



Are you claiming that you can detect asymmetric wavefront
errors of 1/40 to 1/50 wave RMS with Foucault test? Hats down! Same goes for the star test...

Not that any of your figures are really *widely* off the mark - but exaggerate five contributing factors slightly, and the end result will still be off the mark...



I already said that numbers I used are fairly arbitrary.
How close to typical (if there is such thing) they are
depends mostly on the aperture. Just curious, what is your
basis - and concept - in which these are "exaggerated".
If you don't mind to use some extra reference, take a look
of Suiter's comments on this subject (p267). I understand
that it can be hard to swallow for those who take "high
operational quality" of their telescopes as a matter of
personal pride.

With a star test, it's rather easy to see in good seeing that the dominating error of my scope is equivalent to about lambda/6 to lamba/8 peak to Valley at the wavefront spherical aberration once it's cooled down - and it *does* cool down to the point it's obvious thermals don't contribute enough to make that other error no longer the dominating error



Star test won't show cumulative effect of a number of small
errors simply because they don't change perceptibly shape
of the patterns: you can't measure by the eye the changed
proportion of energy between the disc and the rings, and
they don't affect inside- and out-of focus patterns either.
Still, the energy *is* out, and contrast is lowered.

Vlad

#99 wh46gs

wh46gs

    Mariner 2

  • -----
  • Posts: 270
  • Joined: 01 Oct 2005

Posted 19 October 2005 - 06:49 PM

1 arcsecond seeing is 1 arcsecond *for long exposure times*. You can, and do, see details for brief periods that beat that seeing



Of course, when it comes to seeing error, we are talking
averages: 1 arcsec seeing implies certain value for
the atmospheric coherence factor that does determine average
*visual* error as well, not only long-exposure error. Actual
error will constantly vary around that value, to the better
and worse, but significant deviations are exponentially less
likely to occur.

As for your assessment of the errors of a newtonian *system* (including the mirror cell, angled up, pointing at the diagonal and everything), having seen such a *system* from the knife edge of a Foucault test setup close to the centre of curvature of the entire system (yes, that's a strange thing - a truss dob without the trusses, and the Foucault tester somewhere decidedly odd), I can only say that your expectations about the system wavefront errors are hard to reconcile with what I've *seen*.



Are you claiming that you can detect asymmetric wavefront
errors of 1/40 to 1/50 wave RMS with Foucault test? Hats down! Same goes for the star test...

Not that any of your figures are really *widely* off the mark - but exaggerate five contributing factors slightly, and the end result will still be off the mark...



I already said that numbers I used are fairly arbitrary.
How close to typical (if there is such thing) they are
depends mostly on the aperture. Just curious, what is your
basis - and concept - in which these are "exaggerated".
If you don't mind to use some extra reference, take a look
of Suiter's comments on this subject (p267). I understand
that it can be hard to swallow for those who take "high
operational quality" of their telescopes as a matter of
personal pride.

With a star test, it's rather easy to see in good seeing that the dominating error of my scope is equivalent to about lambda/6 to lamba/8 peak to Valley at the wavefront spherical aberration once it's cooled down - and it *does* cool down to the point it's obvious thermals don't contribute enough to make that other error no longer the dominating error



Star test won't show cumulative effect of a number of small
errors simply because they don't change perceptibly shape
of the patterns: you can't measure by the eye the changed
proportion of energy between the disc and the rings, and
they don't affect inside- and out-of focus patterns either.
Still, the energy *is* out, and contrast is lowered.

Vlad

#100 wh46gs

wh46gs

    Mariner 2

  • -----
  • Posts: 270
  • Joined: 01 Oct 2005

Posted 19 October 2005 - 06:59 PM

I don't know what collimation tools you use, but 1.3 arcmin over 1780mm is about 1.34mm, and that's an error that's *glaringly* obvious - with an autocollimator that magnifies that error, it simply would be impossible to miss this even if you tried very hard.



1.3 arcmin @ 1780mm f.l. comes to 0.67mm. It is not only
what you can perceive as centered, it is also how really
centered it is when it appears dead on. And, of course,
do you really think that the structure flow
itself can't re-induce this kind of error *after* you are done with collimating?

Vlad


CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics