92 replies to this topic

[Darren Drake]

We've done another round of DPAC testing at John Pratte's optical test shop this weekend.   Many tests were of privately owned mirrors of various quality sources and ranged from very good to abysmal. 

 

Included also were a couple of noteworthy samples from Lightholder and Zambuto.  These are obviously amoung  the highest end mirrors available and the tests show exactly why.  The first pic shows a sample of various mirrors from various sources that need not be identified. The DPAC test doubles the errors so these tests are very sensitive to any aberrations. 

 

The next shows a 12.5 inch f/5.9 Lightholder.   It is from a New Moon telescope John acquired a while back.   It is about the best mirror we have ever tested. 

 

The second is my personal 20-inch F4 Lightholder.  I feel very fortunate to own such an amazing mirror made to about the highest standard possible and in an amazing Teeters scope.  The optical test report also supports our test pic.  

 

Finally we have a 16-inch F4 Zambuto that John recently acquired.  It does show a trace of undercorrection.  John measured it to be 1% undercorrected.  This is perfectly acceptable given that cooling mirrors tend to overcorrect as they cool.  The level of smoothness in this mirror is phenomenal. 

 

These samples are clearly made to an amazing degree of accuracy and smoothness and with absolutely no zones and owners of such mirrors or scopes with these mirrors are VERY fortunate. 

Attached Thumbnails

• 
• 
• 

Edited by Darren Drake, 06 April 2025 - 08:54 PM.

[Darren Drake]

Here is the zambuto 16-inch F4 

Attached Thumbnails

• 

[Darren Drake]

Here are a few pics of the mirrors on the stand and the scopes they are from..

Attached Thumbnails

• 
• 
• 

Edited by Darren Drake, 06 April 2025 - 08:02 PM.

[Darren Drake]

Here's the 20 on the stand in a previous session and the Zambuto close up..

Attached Thumbnails

• 
• 
• 

Edited by Darren Drake, 06 April 2025 - 08:11 PM.

[MKV]

Here is the zambuto 16-inch F4 

Darren, that Zambuto mirror's wavefront is not an optical null -- your overexposure of the mirror in the article bleaches out figure irregularities. With adjusted exposure, the mirror sides appear to bulge and the central region is deep and steep.

 

[DAVIDG]

 Beside the knife edge images you should also take images on both side of the focus with a Ronchi screen. The Ronchi images show the quality of the figure easier then knife edge images. 

   As MKV has shown the 16" f/4 is not a null and has a zone. 

 

         - Dave 

[CrazyPanda]

That 12.5 lightholder looks amazing. You can tell there's a slight 1" zone on the outside, but it's very mild.

[MKV]

That 12.5 lightholder looks amazing. You can tell there's a slight 1" zone on the outside, but it's very mild.

Good call! Here's an adjusted exposure image. The TDE outer zone stands out like a sore thumb. The center is concave. Correct exposure is critical. If not done right, it can be misleading.

 

 

 

Actually, the 20" F/4 is the best so far. Very good for such a large and fast mirror but still inconclusive as to the actual size of small errors.

 

I wholeheartedly recommend to Darren and his group to invest $150 in a good Bath interferometer and get interferograms of these mirrors with 100-fold more information (all quantitative!) than these poorly exposed autocollimation foucaugrams can give them. Just test at ROC. Dale Eason's software has a numerical null.

[Dale Eason]

Good call! Here's an adjusted exposure image. The TDE outer zone stands out like a sore thumb. The center is concave. Correct exposure is critical. If not done right, it can be misleading.

 

  12_5 F5_9 TDE.jpg

 

Actually, the 20" F/4 is the best so far. Very good for such a large and fast mirror but still inconclusive as to the actual size of small errors.

 

I wholeheartedly recommend to Darren and his group to invest $150 in a good Bath interferometer and get interferograms of these mirrors with 100-fold more information (all quantitative!) than these poorly exposed autocollimation foucaugrams can give them. Just test at ROC. Dale Eason's software has a numerical null.

If only so one can calibrate our eyes to the size of the features seen in these autocollimation images.  For us laymen those shadows don't tell us if those are ok or too big.  I have seen arguments in the distant past on this exact same issue where one will look at it and say how good it is and another looking at the same image says that it is not acceptable.

 

I have enjoyed the work you have done so far but still there is that question.

Edited by Dale Eason, 07 April 2025 - 04:44 PM.

[Darren Drake]

The DPAC test is extraordinarily sensitive.  It takes no prisoners.  These miniscule "errors" mentioned here will  not be detectable in any way in a focused image.  Real world observing conditions,  even in very good seeing and with slowly changing temperatures, will be MANY TIMES more significant than these surface errors.  I'll include a pic that shows examples of surface errors from DPAC tests that are actually relevant as a reminder here.  These are from Rohr's website..

Attached Thumbnails

• 

Edited by Darren Drake, 07 April 2025 - 05:30 PM.

[Dale Eason]

Yes but those Rhor images don't help me at least.  They don't tell you when what you see makes no difference.  They just say if it looks as bad as these shadows it is bad.  How bad?  Are all of those just as Bad?  I doubt it but I don't really know that either.    What we have so far is someone saying if it looks as good as these (your images) it is good but if it looks like these images it is bad.  We also have someone saying if you heighten the contrast yours looks bad.  The old quandary of things learned by experience but hard to teach I think.

[Darren Drake]

Its true there are no strehl numbers or computer analysis involved with these pics.  Still, I'd rather have these pics over an interferometer and some corresponding numbers.  The smoothness level may not be represented by the numbers yet is very important.   

 

I've seen other mirrors that don't look anywhere near as good as these still give very good images under a real sky.  I've also never heard of a Zambuto or a Lightholder that had an unsatisfied owner.  Their reputations are at the highest level.  That alone should indicate that the mirrors shown here are very good.  It's worth studying the hundreds of optical reports on Rohr's website.   The Google translator and AI summary can now be used to read the reports that accompany the DPAC pics.  

Edited by Darren Drake, 07 April 2025 - 07:18 PM.

[Darren Drake]

Here is a link to the Rohr website...

http://r2.astro-fore...php/de/berichte

[Dale Eason]

Unfortunately Rhohr used a low resolution analysis method for his interferometry.  Not nearly as detailed as modern methods can achieve.  He has interesting data despite that.   Yes smoothness is the hardest thing to capture with an interferometer and I like to have both numbers which indicate the overall figure and images that show the smoothness. 

[BGRE]

At least interferometry doesn't inherently make zonal defects appear to be located at the incorrect zonal position.

This slope error dependent distortion has been known for over 70 years but is often blithely ignored.

Foucault testing isn't used for acceptance testing of EUV optics substrates for a good reason its neither accurate enough nor sensitive enough when such testing is done in the visible. 

[duck]

I agree with Dale entirely.  I haven't any foucault-o-grams to compare, but I think if you use a small lap to figure a mirror which needs multiple waves of correction, you'd better know what your doing.  It seems there's "ringing" in some of those fo-grams, probably pretty small.

 

I enjoy the fo-grams, though...keep posting them.

[a__l]

 Beside the knife edge images you should also take images on both side of the focus with a Ronchi screen. The Ronchi images show the quality of the figure easier then knife edge images. 

   As MKV has shown the 16" f/4 is not a null and has a zone. 

 

         - Dave 

 

Yes but those Rhor images don't help me at least.  They don't tell you when what you see makes no difference.  They just say if it looks as bad as these shadows it is bad.  How bad?  Are all of those just as Bad?  I doubt it but I don't really know that either.    What we have so far is someone saying if it looks as good as these (your images) it is good but if it looks like these images it is bad.  We also have someone saying if you heighten the contrast yours looks bad.  The old quandary of things learned by experience but hard to teach I think.

 

https://drive.google...iew?usp=sharing

 

Ronchi (green) and Ronchi simulation from interferogram (soft Dale Eason).

This is my Zambuto 14.5"

In fact, it is close to 1/4 wf, which is the level of good serial mirrors. But it is not premium, as is often written here.

Edited by a__l, 07 April 2025 - 08:58 PM.

[MKV]

The DPAC test is extraordinarily sensitive.  

Normally I wouldn't belabor the issue, but I think some things need to be cleared up, I hope you won't mind. 

 

Autocollimation for mirrors is just double sensitive. It makes errors look twice as large. Nothing extraordinary about that.  Your mirror's 1/4 wave PV  will look like a whopping 1/2 wave error in AC. This makes errors easier to spot, and correct. 

 

That said, a 16-inch F/4 in double pass (AC) will never look perfect unless it's conic is -0.999. Here is what an otherwise perfect 16-inch F/4 should look like with an interferometer, a knife-edge and a 250 LPI Ronchi screen -- if its conic were "only" -0.99, instead of -0.999!

 

 

 

The level of perfection of the figure needed to 3-signficant figure conic constant doesn't seem very realistic in ATM optics. This mirror would test as 1/18 wave PV in single pass, but in double pass it doesn't null. This is where interferometry comes in because it doesn't require a null, or an expensive large flat.

 

Interferometry provides the level of sensitivity and a plethora of quantitative information about the test optic unmatched by any other test available to ATMs. If I were investing in a large expensive mirror I would prefer that over other tests. Trouble is, interferometry sometimes, maybe even too 1often, makes our good mirrors look not so good after all.

[MKV]

Ronchi (green) and Ronchi simulation from interferogram (soft Dale Eason).This is my Zambuto 14.5" In fact, it is close to 1/4 wf, which is the level of good serial mirrors. But it is not premium, as is often written here.

Can you post the picture? The link requires "Google Drive" (?)

[MKV]

The smoothness level may not be represented by the numbers yet is very important.   

You should test your mirrors using a soot knife method (very simple to make), which is as easy as the Foucault test. But the Foucault test is not nearly as sensitive to surface roughness.

 

[a__l]

Can you post the picture? The link requires "Google Drive" (?)

ok.

Attached Thumbnails

• 

[Oregon-raybender]

I will offer few words of caution here and question

the methods

 

One, how good is the optical flat, how smooth

etc...? Is it truly flat or have a slight amount of 

power in fringes? This alone could what you

are seeing. With a center hole cut, it could have a

a minor defect caused by the milling.

 

Second, I don't trust the having the test stand on the

metal wheels. The contact on the floor is of concern.

Vibration is one. I would suggest pads, like used on

telescope leg mounts to reduce any issues.

Wood is fine, I used it on many test stations.

 

Third, air currents. The test station should be enclosed,

air movement, like having a person moving about or

standing behind the tester, will cause currents. Plus

heat and cold currents from the floor or room vents.

The test area should be covered.

 

I have worked with and designed test tunnels and

sat in dark for hours on end. I spent time chasing down 

and improving the testing results.

The Murphy laws get you all the time. 

 

I agree, using a Bath or other type of interferometer

would give a better and trusted result. Checking

other's work, must require a matching level or higher

to confirm the results. I always had the customer

buy off my test system and confirm theirs if they

want to test in house. Ours had to be better

than or equal to theirs. Don't want to play my

ZYGO is better than yours game.

 

Not being harsh, just observing what I see could

be issues and maybe improve your set-up.

 

Starry Nights

 

A test tunnel set up at Tinsley used to test large

optics, cover tunnel. A simply cloth cover will do.

 

 

 

Attached Thumbnails

• 

[Pinbout]

 Beside the knife edge images you should also take images on both side of the focus with a Ronchi screen. The Ronchi images show the quality of the figure easier then knife edge images. 

   As MKV has shown the 16" f/4 is not a null and has a zone. 

 

         - Dave 

And with ronchi images against a flat you can match them against Diffract computer images to get a very good sense of wavefront error 

 

not against a flat but matching with Diffract

https://youtu.be/3mW...FvhFFTxDu27hjpT

 

testing against a flat makes this matching twice as sensitive 

 

and the flat can be several waves flat, it just needs to be smooth.

 

its really a  great while working the figure. 
 

leave the bath for the very end if you need to put a number to the figure 

 

 

Edited by Pinbout, 08 April 2025 - 05:53 PM.

[MKV]

ok.

Has anyone worked out a tolerance envelope (accuracy calibration) for that Ronchi blinking comparator method? Although the patterns look remarkably similar to the naked eye, they are not identical. And how does that method rate when testing, say, a large and slow mirror vs large and fast one? Eyeballing is not the best method of setting up optics for testing, and judging quality of the product.

 

Again, a Bath interferometer is cheaper (currently around $150) than any large optical flat you can find on eBay, especially if the flat is certified (as it should be). A Bath IF doesn't need precision optics; commercial quality will suffice. 

Edited by MKV, 08 April 2025 - 05:57 PM.

[Dale Eason]

MKV,  perhaps you did not understand what is being blinked.  I think it is blinking against the Bath data simulated ronchi.  I think that since the ROC number is in the upper left corner like DFTFringe dispays.  Well actually he said in his post that it was the Bath and DFTFringe.

Edited by Dale Eason, 08 April 2025 - 06:15 PM.