Flex Mirror Notes (Longish)
zgse...@netway.com
Having recently completed my second flex-mirror telescope, I thought
I'd offer a few comments since the topic seems to be getting a bit of
play here these days. First of all, a little background for those who
might be wondering about what we mean when we talk about flex mirrors.
A BRIEF FLEX HISTORY
Flexing describes taking a spherical mirror and mechanically bending
it into the desired shape for a telescope mirror, a paraboloid. Bill
Kelley first wrote up the idea in the June 1992 issue of Sky &
Telescope. Kelley's method is often referred to as "centre pull"
since the mirror is supported at the edge and is pulled by a central
stud. With the right combination of f/ratio, aperture, and thickness,
this method can produce good parabolas. (Kelley right now is working
on developing another centre-pull method that relies upon specially
shaped mirror blanks.)
The next stage in flex development was Alan Adler's method which uses
a wide annular puller to better distribute the flex tension. Adler's
method is considerably more sophisticated than Kelley's original
technique and produces significantly better results. Often this
method is referred to as "microflexing." It is a fair characterization
to say that while Kelley's simple method approximates a parabola,
Adler's is capable of producing a perfect parabola. Adler's approach
was written up in the November 2000 issue of Sky & Telescope and is
must reading for anyone contemplating this approach.
MY RESULTS
I recently completed a telescope utilizing flexed 8-inch f/6 mirror.
I have done extensive bench testing and star testing and can say
without qualification that this is the best mirror I have ever made,
and it is probably the best I have ever tested. Foucault tests
yielded numbers like 1/40th wave or better (nonsense numbers), so
clearly what surface error there might be are beyond my ability to
measure them on the bench. Star test patterns inside and outside of
focus are identical with lots of contrast between the rings. I allow
for the possibility that under superb seeing conditions, the star test
might show some minor imperfection that has eluded me so far, but
none-the-less, I confidently characterize this mirror as being
excellent.
FLEXING PLUSES AND MINUSES
One question that comes up frequently is, is flexing a mirror easier
than using standard parabolizing methods? The answer depends on you.
If you can more easily produce a superb sphere than you can a superb
parabola, then the answer is yes. If superb parabolas come relatively
easily to you, then you will likely find the flex method has little to
offer. I will say that in general, making a flexed mirror is only
worthwhile if you seek a mirror that is very good to excellent -- it's
easy enough to make a mediocre mirror with traditional methods.
The extra work involved in making a flex mirror consists of:
1. Removing wedge from your blank. (This took me about 15 minutes,
but it did require the build a simple wedge tester like the one shown
in Adler's article.)
2. Building a flex cell. (This is described in detail in Adler's
article and I would only add that it is nowhere near as difficult to
make as some seem to imagine. Once you hold one in your hands, you
quickly realize how little there really is to it. I think making a
9-point flotation cell -- which ATMs do routinely -- is harder.)
The principle benefits of making a flex mirror are:
1. The ability to utilize the Foucault test as a null test. Since
you're generating a sphere, the Foucault test IS a null test. The
smallest errors on the surface of the glass stand out in bold relief.
Having a very sensitive test opens up the possibility of making
fantastically smooth and accurate surface -- if you can see a defect,
you can fix it. You will likely find that when making a super-smooth
sphere, your skill as a glass worker is the limiting factor, not the
tester or your ability to read the test.
2. You don't have to parabolize. There are many advantages to this,
but I will mention only one. Generally, the more parabolizing and
correcting you have to do, the rougher the surface becomes. Some
mirrors I have seen are the obvious result of overworking the glass --
the figure is good (the mirror has the right "numbers") but surface
smoothness has been sacrificed to get it.
3. You can figure the mirror in real time using the most sensitive
test available to amateurs: the star test. Although I like to do a
lot of bench testing, for my flex mirror I did my final tweaking at
the eyepiece. This consisted of nothing more than adjusting the flex
tension by turning a knob on the mirror cell.
4. For beginners, flexing probably offers the greatest chance of
producing a superb mirror.
Another question that turns up frequently pertains to mirror cooling.
Since the rear surface of a flex mirror is not available for cooling
purposes, does this slow the mirror's ability to come to ambient?
Adler has done a fair amount of testing to see if this is the case and
so far his test show that an actively cooled flex mirror cools at the
same rate as an actively cooled one in an open cell. Certainly, I
have not noticed my flex mirror taking any longer to cool down than my
other scopes, but I have only used it during the summer so far where
the temperature drop has not been greater than 15 - 20 degrees
Fahrenheit.
In closing, I hope that anyone who makes a flex mirror will offer
their comments, good or bad. The bottom line is that method does work
(and works well!) if one follows Adler's instructions and heads the
warnings and caveats outlined in his article.
Regards,
Gary Seronik
(Remove the "z" for my actual e-mail address.)
Del Johnson
in the field. What will happen during the course of a one hour prime focus
photograph? I would prefer a method where the mirror is flexed (the other
way) during fabrication and then allowed to relax into the final figure.
The final product would then be a free-standing relaxed mirror, and would
dispense with the bother of tuning the mirror flexure every time the
telescope is used.
Del Johnson
<zgse...@netway.com> wrote in message
news:3b73c7cf...@news.netway.com...
Michael Peck
>Hello S.A.A.:
>
>Having recently completed my second flex-mirror telescope, I thought
>I'd offer a few comments since the topic seems to be getting a bit of
>
>I recently completed a telescope utilizing flexed 8-inch f/6 mirror.
>I have done extensive bench testing and star testing and can say
>without qualification that this is the best mirror I have ever made,
>and it is probably the best I have ever tested. Foucault tests
>yielded numbers like 1/40th wave or better (nonsense numbers), so
>clearly what surface error there might be are beyond my ability to
>
Gary:
What do you mean by nonsense numbers? Either the test bears some
meaningful relationship to reality or it doesn't. If it doesn't you
shouldn't bother tossing out numbers that even you don't believe.
One of the things that bothers some people on the ATM list is that you
and Bill Kelley (and Alan Adler) have cited these seemingly fantastic
numbers for what's achievable with this method, and ATMs just don't
buy those numbers.
Now I think I understand what you're really trying to say - there's a
theoretical claim being made, and there's an empirical statement about
what you actually achieve. The theoretical claim is that if you take a
perfect sphere worked into a perfectly homogeneous piece of glass and
microflex it with a perfect apparatus manufactured to the recipe in
Adler's article then a 1/X wave paraboloid can be achieved. The
empirical claim is that careful star testing of a real mirror over
many nights reveals no apparent imperfections. Those are two very
different claims, and it would help immensely if you would be more
careful about distinguishing them. Of course it wouldn't do any harm
to have some flexed mirrors tested by methods generally recognized as
being quantitatively valid, but that costs money that could probably
be better spent on other things. Anyway, most ATMs are willing to
accept a careful star test as evidence of optical quality.
[A crotchet: when are you going to drop this 1/X wave nonsense? I bet
most s.a.a. readers are at least middle school graduates and probably
comfortable with decimals. And while you're at it why not be a little
more careful about what you're referring to. Is it P-V, RMS - surface
or wavefront?]
[Another crotchet: Adler's statement that "the {geometric} spot ratio
is really a better figure of merit than wavefront error" is clearly
false when you're talking about diffraction limited optics. It's also
false that "minimizing one also minimizes the other" (p. 132 of the
Nov. 2000 article). It's never too late to catch up to 20th century
optics.
BTW this error is innocuous in the diameters and focal ratios you've
actually built, but it could make a difference in more extreme
situations.]
>2. You don't have to parabolize. There are many advantages to this,
>but I will mention only one. Generally, the more parabolizing and
>correcting you have to do, the rougher the surface becomes. Some
>
Obviously, you do have to parabolize. You're just doing it by
different means. The rest of what you say is quite correct, but I
think most "advanced" ATMs figure they can make a pretty good smooth
paraboloid on, say a 15-30cm f/5-6 mirror - which seems to be the
range that's actually been built so far. ATMs will get excited when
someone demonstrates that this will work on larger and/or shorter
focal ratio mirrors. I personally lost short term interest when I
realized that the nicely spherical piece of f/3 Sitall sitting in my
closet isn't going to be flexed into a RC cass primary. I'm afraid
it's destined to be assaulted with pitch and ZrO one of these days.
A few more observations:
"Microflexing" is clearly a form of active optics as the term is
understood by the professional community. I've never seen a mention of
the connection by you, Bill Kelley, or Alan Adler - which makes me
wonder if you even realize the connection exists. If you don't, you
should. For one thing, issues that are problems for flex technology
(astigmatism, higher order spherical aberration) are solved in modern
active optics systems.
Another issue that may bother some ATMs is the lack of a complete
technical exposition of the theory behind the method. I suspect
commercial considerations may have something to do with that, and if
so it's fine with me. I'm all in favor of people making an honest buck
off their ideas. OTOH if it's because Mr. Adler thinks ATMs won't
understand the physics I think he might be surprised at how large a
percentage do understand it.
I hope I'm not sounding too skeptical or negative, because I'm not. A
lot of the skepticism that's been expressed on the ATM list I *think*
is directly the result of those 1/X wave claims, and the apparent
defensiveness of Bill Kelley when questioned closely about them.
Mike Peck
_________________
Michael Peck
email mpe...@ix.netcom.com
Wildlife photography page http://home.netcom.com/~mpeck1/index.html
Amateur telescope making http://home.netcom.com/~mpeck1/astro/astro.html
Paul Stock
Your thoughts and comments are appreciated. Nevertheless, there really is
something exciting about flexed mirrors for those of us on the *other* side
of the mirror grinding experience. I get the impression from reading your
comments and the occasional buzz on this subject on SAA and the atm list
that advanced mirror grinders feel flexed mirror concepts and claims that a
level of precision equal to that achievable by advanced atms by those less
experienced is a threat to their hard-earned status as masters of the
paraboloid.
What I really think flexed mirrors offer is hope to those of us who aspire
to have a precision optic for personal use and who might be intimidated by
the prospect of creating precision paraboloids, and who don't mind admitting
they took a shortcut in acquiring the fantastic images they enjoy at the
eyepiece when they use their flex cell telescopes. I suppose there might be
a certain smugness for some who have cheated the process and accomplished
their "poor man's paraboloid" when testing shows performance in the field is
equal to the work of the masters. But for me, it simply represents hope of
having superior optical performance should I discover I haven't the time or
aptitude to become a master.
Nevertheless, I think the gist of Seronik's article is simply to say that
"flex mirrors work great." And I don't think there's anything wrong with
that. Maybe others are claiming flexed mirrors represents a superior
technology to be universally embraced and that those who are able to achieve
optical precision the old fashioned way are on their way to being dinosaurs,
but Gary doesn't appear to be asserting this.
I for one am much more comfortable with an accurate optic and a relaxed
substrate, and I hate to tweak what ought to be static, but if there was a
greater chance of enjoying a superior image at the eyepiece via a flex
mirror rather than with a permanent paraboloid given *my* level of
experience and time to devote to the task at hand, I'd certainly consider
it. I think the Masters of the Paraboloid may rest secure. No one will be
dethroning them anytime soon. Besides, it's just plain fun to consider new
ways of doing things, especially when it works.
Paul Stock
"Michael Peck" <mpe...@ix.netcom.com> wrote in message
news:HvtzO3UcO+gZtT...@4ax.com...
Del Johnson
Nobody has to become a master mirror maker to have a good mirror. One can
always purchase a mirror. They really do not cost that much; even a $1000
expenditure isn't going to change my life.
Del Johnson
"Paul Stock" <comab...@earthlink.net> wrote in message
news:yWVc7.459$ZM2....@newsread2.prod.itd.earthlink.net...
> Paul Stock
>
>
Scott Rychnovsky
Stock <comab...@earthlink.net> wrote:
Paul,
Here is another view from a mirror maker. I have not made a flexed
mirror, but I have considered it. However, I am still not sure it is
worth the bother. These days, I find the difficult steps in making a
good parabolic mirror are to get the lap to behave well to produce a
very smooth surface, and to get a really nice, zone free "spherical"
mirror with a good edge. The final step of parabolizing does take some
time, because I do it in small steps, but I don't find it all that
difficult. The flexed mirror just lets me omit the last step, but at
the cost of a more elaborate cell design.
I have played with Alan Alder's program, and the numbers one can get
with medium to slow focal length systems are amazing. The numbers I
get with conventional figuring steps are not as high, but that is at
least in part because I have no way to accurately measure numbers as
high as his program predicts. I can still produce mirrors with
excellent star tests. (It may take more than one try!)
My main point is that making a really good smooth and zone-free sphere
to flex isn't as easy as one might think, and converting that really
good sphere to a paraboloid takes some experience but it isn't
necessarily difficult.
Scott
mjc5
must confess. I really haven't had a problem with Bill, but I got my
wings singed by a few other folk on the list.
I am very skeptical of 1/70th wave mirrors, just from testing
experience. So many things have to be just right to test a mirror to even
1/10th wave. The student can calculate just how accurate a 1/10th mirror
as an excercise. I know that some people have routinely bandied about 1/20th
wave or even greater for their results. I gotta be extremely skeptical
about that.
Remember we're the same breed of people who once saw Martian canals! We
can see anything we put our minds to. ;^)
BUT, I still say that if there is a trump card to these mirrors, it is
their potential smoothness. This can be more important than we give it
credit for. Gary speaks of overworked mirrors, and he couldn't be more
correct. I have looked through a mirror that tests out at 1/3rd wave,
yet gives very good images. I suspect this is because it is very smooth.
I'll know soon after I build mine. And if I become a convert, I'll
volunteer as the flex mirror poster child, if they want someone that
ugly. ;^)
zgse...@netway.com
"Del Johnson" <dela...@san.rr.com> wrote:
>dispense with the bother of tuning the mirror flexure every time the
>telescope is used.
As I said in my post to the other flex mirror thread:
Both Alan Adler and I have found that our flex mirrors are essentially
set-and-forget -- once adjusted to tension, no further adjustment has
been necessary. With his first cell Alan had to adjust it
periodically during the first couple of weeks, but this was using a
different sponge rubber than I used and the he now uses. Since
bringing my mirror up to tension, it has maintained its (excellent)
figure without need for further adjustment.
zgse...@netway.com
Michael Peck <mpe...@ix.netcom.com> wrote:
>
>What do you mean by nonsense numbers? Either the test bears some
>meaningful relationship to reality or it doesn't. If it doesn't you
>shouldn't bother tossing out numbers that even you don't believe.
Because I don't know with certainty the limits of the Foucault test, I
don't trust numbers like this -- they may not be the "truth." This is
why I added the second part of this statement which you deleted in
your query.
To avoid confusion further downstream when replies are piled on top of
replies, here is the complete passage:
MY RESULTS
I recently completed a telescope utilizing flexed 8-inch f/6 mirror.
I have done extensive bench testing and star testing and can say
without qualification that this is the best mirror I have ever made,
and it is probably the best I have ever tested. Foucault tests
yielded numbers like 1/40th wave or better (nonsense numbers), so
clearly what surface error there might be are beyond my ability to
measure them on the bench. Star test patterns inside and outside of
focus are identical with lots of contrast between the rings. I allow
for the possibility that under superb seeing conditions, the star test
might show some minor imperfection that has eluded me so far, but
none-the-less, I confidently characterize this mirror as being
excellent.
>
>Another issue that may bother some ATMs is the lack of a complete
>technical exposition of the theory behind the method. I suspect
>commercial considerations may have something to do with that, and if
>so it's fine with me.
No, it's that the article was a practical how-to piece on building
flex mirrors, not a detailed treatment of the theory behind it.
>off their ideas. OTOH if it's because Mr. Adler thinks ATMs won't
>understand the physics I think he might be surprised at how large a
>percentage do understand it.
Adler is happy to discuss theory with anyone who is interested.
Ralph Brown
interesting and I for one appreciate you comments.
I'd suggest however, that you NEVER say that you mounted one in a GoTo
system!
Ralph
--
Napier, a multifunction calculator for PocketPC - check it out at
www.nca-corp.com
********* PLEASE NOTE FOR REPLIES **************
My email address is: ralph AT nca DASH corp DOT com
Sorry about this, but spammers happen
**************************************************************
Dan Chaffee
<comab...@earthlink.net> wrote:
>Your thoughts and comments are appreciated. Nevertheless, there really is
>something exciting about flexed mirrors for those of us on the *other* side
>of the mirror grinding experience. I get the impression from reading your
>comments and the occasional buzz on this subject on SAA and the atm list
>that advanced mirror grinders feel flexed mirror concepts and claims that a
>level of precision equal to that achievable by advanced atms by those less
>experienced is a threat to their hard-earned status as masters of the
>paraboloid.
>
Why would it be a threat to their status anymore than it would be to
their fun of polishing a mirror to a permanent paraboloid? Do you
consider that the most interesting part of the mirror making proccess
is in figuring and that this is what "masters of the paraboloid" would
miss? Do consider that some of us make fine mirrors because we
just like doing it, not to elevate ourselves beyond those who don't or
have trouble with traditional methods.
I have no problem with those interested in flexing, yet I have no
interest in attempting it. BTW, I don't have to be a master mirror
maker to make a mirror as good as it needs to be for planetary work,
and neither do alot of other mirror makers.
D Chaffee
Michael Peck
<comab...@earthlink.net> wrote:
>Mike,
>
>Your thoughts and comments are appreciated. Nevertheless, there really is
>something exciting about flexed mirrors for those of us on the *other* side
>of the mirror grinding experience. I get the impression from reading your
>comments and the occasional buzz on this subject on SAA and the atm list
>that advanced mirror grinders feel flexed mirror concepts and claims that a
>level of precision equal to that achievable by advanced atms by those less
>experienced is a threat to their hard-earned status as masters of the
>paraboloid.
>
>
Paul:
Masters of the paraboloid? Not me! I'm strictly a duffer at glass
pushing.
I don't think amateurs feel threatened by this technology; at least I
wouldn't assume that's the motivation for any skeptical commentary you
read. Amateurs by definition are making mirrors for fun and like any
hobbyist will continue doing anachronistic and inefficient things just
for the joy of doing them. Some might even try newfangled ideas just
for fun too.
"Advanced" ATMs don't really have any deep secrets that novices need
to spend years trying to learn. If the figuring advice you see on the
ATM list sounds like hocus-pocus sometimes it's mostly because mirror
making is more art than science.
I really think it's the numbers that have triggered most of the
skepticism. ATMs just have an instinctive negative reaction to claims
of "1/X" wave wavefronts when X is much greater than 10.
Personally I have no particular problem with the numbers. The numbers
that people have jumped on are coming from a mathematical model.
Reality most likely isn't quite going to live up to the predictions of
the model, but assuming Adler and his collaborators did their homework
the model should give useful predictions of reality. Bill Kelley,
Gary, and others who've actually made these things claim that in fact
reality really does resemble the model and I completely believe them.
I'd still be happy to see an interferogram of at least one flexed
mirror though.
Paul Stock
I guess my next step is to take the two 8-inch blanks I have recently
purchased and make one permanent paraboloid and another flex mirror and
decide for myself. It is heartening to hear that most of the noise
encountered on the atm list is coming from those a couple standard
deviations or so from the most likely case scenario.
BTW, I have a quick question that is probably readily available in some
archives somewhere, but perhaps an answer here could be of some use to
others as well.
If I wanted to use a subdiameter tool, what is a good ratio to follow for
selecting the tool size? Does that ratio change as you get to larger
apertures, i.e. is a 6-inch subdiameter tool good for an 8-incher and a
12-inch tool good for a 16-incher? (75% of the mirror, or so). Am I starting
off on the wrong foot to use the full-diameter tools that came with my
mirror kits (TDE issues, etc.)?
Thanks in advance for answering what is probably a tired subject--
Paul
"Scott Rychnovsky" <n...@spam.com> wrote in message
news:100820011333015163%n...@spam.com...
Scott Rychnovsky
Paul,
It sonds like a fun project. Let us know what happens.
Regarding tool size, what are you trying to accomplish? I am using an
80% tool for grinding out a 12" mirror. For polishing, I would use 80%
or larger to avoid edge problems. For figuring, I would use 60% or
smaller, but it depends on the stroke and the current figure on the
mirror.
Scott