Click here if you are having trouble logging into the forums
Privacy Policy |
Please read our Terms
of Service | Signup and
Troubleshooting FAQ | Problems? PM a Red or a Green Gu.... uh, User
Mark Harry
Post Laureate
Reged: 09/05/05
Posts: 3124
Loc: Northeast
|
|
"Mark, it's no fun agreeing with you. "
******************************
Ok, I'll try to disagree next time! 
--------------------
So many projects, so little time!
|
Dick Parker
sage
Reged: 08/17/07
Posts: 247
|
|
thought for the thread, no one in particular.
FWIW - As I understand it, "Strehl" is a measure of light in the Airy disk, v.s. light that should be in the Airy disk. In the strictest sense, to get a real "Strehl" you would have to measure the point spread function. In other words, measure on a real mirror, the light in the disk, measure the light in the rings, and ratio that. I do not know of any sensible way for us to do that.
I think what you see when you see a "Strehl" number in a mirrors rating is a mathematical extrapolation based on a few measurements. From Jim Burrow's website, I see that Strehl can be APPROXIMATED by the relationship SR = 1 - 4*Pi^2*RMS^2. No need to get bogged down in the math. Point is that what we see as "Strehl" is an approximation based on RMS, and that approximation is "close" as long as RMS is 1/14 wave or less.
Bottom line - smoothness is not captured in the "Strehl" as you see it in the ratings. This is because the ratings are based on measurements of a few sample points.
I am also of the camp that believes that smoothness is very important.
Dick Parker
|
sixela
Postmaster
Reged: 12/23/04
Posts: 10853
Loc: Boechout, Belgium
|
|
Quote:
thought for the thread, no one in particular.
FWIW - As I understand it, "Strehl" is a measure of light in the Airy disk, v.s. light that should be in the Airy disk. In the strictest sense, to get a real "Strehl" you would have to measure the point spread function.
The finer point is that many different point spread functions yield a given numerical value for the Strehl ratio, and not all of them affect different types of observation activities the same way...
--------------------
400mm f/4.46 self made Dobsonian on Tom Osypowski equatorial platform
Orion Starblast (114mm f/4 reflector, Alt/Az)
|
sixela
Postmaster
Reged: 12/23/04
Posts: 10853
Loc: Boechout, Belgium
|
|
Quote:
thought for the thread, no one in particular.
FWIW - As I understand it, "Strehl" is a measure of light in the Airy disk, v.s. light that should be in the Airy disk. In the strictest sense, to get a real "Strehl" you would have to measure the point spread function.
The finer point is that many different point spread functions yield a given numerical value for the Strehl ratio, and not all of them affect different types of observation activities the same way...
Quote:
Bottom line - smoothness is not captured in the "Strehl" as you see it in the ratings. This is because the ratings are based on measurements of a few sample points.
Actually, smoothness is usually actually reflected in the ratings, at least if you have enough sample points. The problem isn't merely that you can't trust a derived Strehl number; even if you could magically derive the true Strehl ratio, the number gives you a *floor* for performance but there's no way to differentiate between best- and worst-case scneratios giving that exact number.
The problem is that even small percentages of light can cause problems if they're converted to veiling glare, because sometimes there are objects close to dim objects that have an integrated magnitude that is *a lot* more than that of the object, so even a couple of percents of stray light from the bright objects converted to veiling glare will affect contrast.
By contrast, another system with identical Strehl might actually diffract the same couple of percents of the light from irrelevant objects almost exclusively into the first diffraction ring for the Airy pattern of those same objects, and that will not be as nefarious to observing something dim in the neighbourhood as veiling glare.
--------------------
400mm f/4.46 self made Dobsonian on Tom Osypowski equatorial platform
Orion Starblast (114mm f/4 reflector, Alt/Az)
|
mark cowan
Vendor (Obsidian Optics)
Reged: 06/03/05
Posts: 2159
Loc: salem, OR
|
|
The way I see it, the qualities that make a mirror excellent are fairly easy to define, if hard to achieve repeatedly. But the measurements that would distinguish a mirror that has all of those qualities from one that doesn't - and performs less well as a result - can't be condensed down to a single figure of merit, or even several: eg, overall smoothness can be captured in RMS but the numbers derived for that from various tests are only as accurate as the sampling density; interferometry can assign a Strehl on an optic that says "it's this good or better" but the limit of the test/tester determines what level of error can be determined; some tests look at a few axes of the optic over a limited range of samples (handful) and extrapolate to the entire surface based on quality of figure of revolution and smoothness of figure - the same test on a poor mirror with bad figure of revolution is nigh useless; star testing will prove the mettle of any optic (plus the secondary) if done by somebody with experience, but unless you go with a Roddier analysis it provides no numbers (although I'd argue that "null" is a number as in "no error visible" ).
Now we (or some at least) know that it's possible to make excellent mirrors using any of these tests to validate fabrication. And it's equally possible to make not-so-good mirrors if you just pick one test and rely on it without verifying that the assumptions (figure of revolution, etc) are being met. That's why I always provide four measures in addition to whatever numbers came off the test stand - polish, smoothness, diffraction edge and figure of revolution. But they aren't numbers, they're just evaluations. And I give an estimated PVW/Strehl ratio which is generally never what the bench test broke down to, taking into account the error levels inherent in the bench test. All of which are ignored on those rare occasions when a mirror goes up for resale. 
Best,
Mark
|
Mark Harry
Post Laureate
Reged: 09/05/05
Posts: 3124
Loc: Northeast
|
|
"the assumptions (figure of revolution, etc) are being met. "
*********************
-THIS- relies on the skill, and experience of the manufacturer handling the glass. I think this is where certain manufacturers fall down and let the turkeys out the door. (gobble-gobble!)
M.
--------------------
So many projects, so little time!
|
wargentin
member
Reged: 05/04/07
Posts: 29
Loc: France
|
|
Maybe some of you will be interested in this video. This is a 16" GSO, tested in autocollimation :
http://www.youtube.com/watch?v=5x42QF2KVsM
This video illustrates very well what a rough surface is. Even if the numbers were acceptable (I am not aware of the actual strehl), obviously this mirror would not be a good one.
|
Darren Drake
Pooh-Bah
Reged: 10/09/02
Posts: 1293
Loc: Illinois
|
|
Is that supposed to be nulled out? If so that is a really bad mirror for a number of reasons.
--------------------
Astronomy educator/Sidewalk astronomer
Owner of Astronomy Delight franchise
18 inch f4.42 Dob on eq platform w ST120 f/5 finder
12 inch Zhumell Dob modified by Brian Reed on eq platform
8 inch f/8 eq planetkiller
William Optics red 10th Anniversary 80mm FD
PST
24lb eyepiece box
Cernan Space Center astronomer
Member of Northwest Suburban Astronomers
|
wargentin
member
Reged: 05/04/07
Posts: 29
Loc: France
|
|
My understanding of the autocollimation is that the spherical aberration is due to the mirror, and not to the testing method.
|
davidpitre
Carpal Tunnel
Reged: 05/10/05
Posts: 1824
Loc: Central Texas
|
|
What I still can't really understand are the discrepancies between the advertised and actual test numbers from some of the most popular "premium" mirror makers. For instance, Dave's club tests a 1/2.8 wave Galaxy and the Jack Kreamer article cites 2 Galaxy mirrors testing at 1/3 wave . Yet if you look at the Galaxy web site they show scores of tests done with an interferometer all of which show strehls from about .93 to .98 . Obviously it's not just Galaxy, there seems to be huge discrepancies between the what some of these makers are claiming and what they are sending out.
If clubs can so easily test the mirrors, and they are so far from their advertised specs, I have a hard time not coming to the conclusion of wide-spread dishonesty. I'd sure like to be convinced otherwise.
--------------------
David
|
sixela
Postmaster
Reged: 12/23/04
Posts: 10853
Loc: Boechout, Belgium
|
|
Quote:
This video illustrates very well what a rough surface is. Even if the numbers were acceptable (I am not aware of the actual strehl), obviously this mirror would not be a good one.
A lot of handwaving, little facts. This is actually farily typical for a GSO - they've got that radial pattern from their parabolisation procedure. The Strehl ratio (mainly because of that and sometimes some undercorrection) is usually between 0.89 and 0.93-0.95. This one's definitely at the bad end of the normal range, so my guess would be (from Wolfgang Rohr's tests of similar mirrors) that the Strehl ratio would be around 0.86-0.89.
They're actually not "obviously not good" - I've star tested enough of them to know it.
They're obviously not Zambutos and the roughness will contribute a tad to veiling glare, but we're still talking only some percents of the total light from the objects that go outside of the central portion of the diffraction pattern. A few years ago, anyone would have killed for this level of quality (or lack thereof) in a mass produced scope.
The complete mirror set can be purchased for €800 around these parts, to put things in perspective (and prices in Europe are higher than in the US).
--------------------
400mm f/4.46 self made Dobsonian on Tom Osypowski equatorial platform
Orion Starblast (114mm f/4 reflector, Alt/Az)
|
sixela
Postmaster
Reged: 12/23/04
Posts: 10853
Loc: Boechout, Belgium
|
|
Quote:
Is that supposed to be nulled out? If so that is a really bad mirror for a number of reasons.
Yes, it's supposed to be nulled out, and no, it's not "really bad" - unless you paid a lot more for that mirror. It's one of the worse ones if you see the range Wolfgang Rohr, some other Germans and one French astronomy magazine have tested, though.
Here are two other ones to look at:
http://www.astro-foren.de/showthread.php?p=37905
There are also a couple of pictures on a null tester.
--------------------
400mm f/4.46 self made Dobsonian on Tom Osypowski equatorial platform
Orion Starblast (114mm f/4 reflector, Alt/Az)
|
sixela
Postmaster
Reged: 12/23/04
Posts: 10853
Loc: Boechout, Belgium
|
|
Quote:
My understanding of the autocollimation is that the spherical aberration is due to the mirror, and not to the testing method.
Depends on the null tester you're using and how good the tester is at scrupulously following a correct protocol, of course...
--------------------
400mm f/4.46 self made Dobsonian on Tom Osypowski equatorial platform
Orion Starblast (114mm f/4 reflector, Alt/Az)
|
|
4 registered and 12 anonymous users are browsing this forum.
Moderator: Don W, werewolf6977, Jason B
Print Thread
|
Forum Permissions
You cannot start new topics
You cannot reply to topics
HTML is disabled
UBBCode is enabled
|
Thread views: 3208
|
|
|
|
|
|
|
Previous
Index
Next
Threaded
[Re: 
Edit
Reply
Quote
