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.

Photo

Making a first mirror: 10" f6

  • Please log in to reply
166 replies to this topic

#126 Chris Adamczyk

Chris Adamczyk

    Sputnik

  • -----
  • topic starter
  • Posts: 43
  • Joined: 18 Nov 2019
  • Loc: Southern California

Posted 25 January 2020 - 04:13 PM

I redid the KE test early this morning and got some results still on the same diameter that I have more confidence in and checked to ROC at 119" previously I had 119.125". If I understand the graphing based on percentages of the 0.025" target, zones 2 and 4 would be high (guessing). I also ran the numbers through Figure XP, the M-L plot tornado shows zone 3 & 4 as high. 

 

I'll try another couple of diameters as time allows

Chris

Attached Thumbnails

  • 20-0125 A-2.jpg
  • Chris 10inch f 6 (2)_M-L_Plot.gif
  • Chris 10inch f 6 (2)_Surface.gif


#127 Stathis_Firstlight

Stathis_Firstlight

    Stathis Firstlight

  • -----
  • Vendors
  • Posts: 29
  • Joined: 28 Dec 2019
  • Loc: Germany

Posted 25 January 2020 - 04:38 PM

Hello Chris,

this is a nice FXP analysis. If this data are correct, you are slightly overcorected, but realy only little.

The CC should be around -1.13. You can press the "best fit conic constant", to find out.

I would try TOT and concentrate to polish at the approx. 70-85% zone. Not more than 3 minutes.

 

But how it comes, that the data before showed approx. 50% correction? Dos you parabolize in between?

 

P.s.

You folks at the other side of the ocean still use inches instead of milimeters?


Edited by Stathis_Firstlight, 25 January 2020 - 04:46 PM.

  • LarsMalmgren and Pinbout like this

#128 dogbiscuit

dogbiscuit

    Apollo

  • *****
  • Posts: 1,212
  • Joined: 18 Oct 2018
  • Loc: Barksdale

Posted 25 January 2020 - 04:54 PM


P.s.

You folks at the other side of the ocean still use inches instead of milimeters?

We have hybridized here.  We use the milli-inch.  lol.gif
 


  • gregj888, LarsMalmgren, Pinbout and 1 other like this

#129 Pinbout

Pinbout

    ISS

  • *****
  • Posts: 24,742
  • Joined: 22 Feb 2010

Posted 25 January 2020 - 06:17 PM

We have hybridized here.  We use the milli-inch.  lol.gif
 

It depends on the subject... some of us need use feet as a unit of measurement lol.gif lol.gif lol.gif


Edited by Pinbout, 25 January 2020 - 06:20 PM.

  • Stathis_Firstlight likes this

#130 Pinbout

Pinbout

    ISS

  • *****
  • Posts: 24,742
  • Joined: 22 Feb 2010

Posted 25 January 2020 - 06:25 PM

Those twin peaks are classic overcorrection

84863D9E-2C65-4CE1-9742-601DEF329B00.jpeg

 

a single hump in the middle is undercorrection

 

as stated above I’d do ToT short narrow ovals to lower correction and smooth out surface

 

as stated 3x’s around and test, your that close.

 

 



#131 Chris Adamczyk

Chris Adamczyk

    Sputnik

  • -----
  • topic starter
  • Posts: 43
  • Joined: 18 Nov 2019
  • Loc: Southern California

Posted 25 January 2020 - 07:08 PM

Can you post a diagram of what the stroke looks like, I'm afraid of messing up.

 

I grew up in New Zealand, at about 8yrs old we went metric, I'm a Land Surveyor and worked in metric until I got to the USA 34 years ago and relearn't the imperial system, Surveys over here us a ft divide into 100 intervals, people are analog based and somehow to me the ft fits in. The real problem is there in and international ft and a US ft, the difference is very very small but when you do geodetic conversions that involve very large values, things get ugly.

 

The 2nd set of data was my 2nd or 3rd attempt with the Foucault test, I was at it at 5am this morning, I have turned the mirror 120 degrees and will test again to make sure the measurements are valid.

 

Chris 



#132 dogbiscuit

dogbiscuit

    Apollo

  • *****
  • Posts: 1,212
  • Joined: 18 Oct 2018
  • Loc: Barksdale

Posted 25 January 2020 - 07:31 PM

These test results are a better match to what I see in the Ronchigram.

I thought it looked near full correction, maybe a bit over.

It is over a little.

 

10f6secondtest.jpg

 

numeric differences here are just rounding differences of no real significance.

But for graphing, all 3 transitions between zones are downslopes.  Over corrected between each zone.

105% is still pretty  good but the 137% between zones 2&3 hurts a bit.

If all three transitions were 105% the mirror would be very good.

10f6zonetozone.jpg

 

I would do

one turn 1/3 W MOT 

and

one turn 1/3 W TOT

That might be enough.

What the others said could work too.

 

 

There is much I can't be sure of because the Ronchi focus is bad.

If you don't get better images I'll just guess the Gugolz is making a smooth surface.

It should if you do a good job maintaining the lap and pressing good before work sessions.

Far as I can tell the edge is good.

 

 

 

 

 



#133 BGRE

BGRE

    Mercury-Atlas

  • *****
  • Posts: 2,993
  • Joined: 21 Mar 2016
  • Loc: New Zealand

Posted 25 January 2020 - 07:53 PM

Can you post a diagram of what the stroke looks like, I'm afraid of messing up.

 

I grew up in New Zealand, at about 8yrs old we went metric, I'm a Land Surveyor and worked in metric until I got to the USA 34 years ago and relearn't the imperial system, Surveys over here us a ft divide into 100 intervals, people are analog based and somehow to me the ft fits in. The real problem is there in and international ft and a US ft, the difference is very very small but when you do geodetic conversions that involve very large values, things get ugly.

 

The 2nd set of data was my 2nd or 3rd attempt with the Foucault test, I was at it at 5am this morning, I have turned the mirror 120 degrees and will test again to make sure the measurements are valid.

 

Chris 

That only applies to surveys for consistency with old survey data and is known as the US survey foot and derives from an old (1866) definition of 1m = 39.37 inches. Prior to adoption of a unified definition (~ 1966??) the US, Canadian and British inches all differed slightly in length.

That value isn't used in other fields where by definition 1 inch = 25.4 mm.



#134 Pinbout

Pinbout

    ISS

  • *****
  • Posts: 24,742
  • Joined: 22 Feb 2010

Posted 25 January 2020 - 08:00 PM

https://www.youtube....salR23dE&t=154s

 

something like this figure 8 but ToT and maybe maybe a little shorter... keep the strokes flowing, that's why you do ovals...

 

starts and stops create zones... so ovals. you could do figure 8's also just to smooth^2   grin.gif



#135 dogbiscuit

dogbiscuit

    Apollo

  • *****
  • Posts: 1,212
  • Joined: 18 Oct 2018
  • Loc: Barksdale

Posted 26 January 2020 - 12:58 AM

Made a mask overlay and placed it over the simulated Foucault images for 10" f:6 perfect parabola.

10f6maskoverlay.jpg 10f6maskoverlayROC.jpg

 

 

 

Zone 1                                                                                       Zone 2

10f6maskoverlayZ1.jpg 10f6maskoverlayZ2.jpg

 

 

 

Zone 3                                                                                       Zone 4

10f6maskoverlayZ3.jpg 10f6maskoverlayZ4.jpg

 

Looks good to me.  Send it out for coating.



#136 dogbiscuit

dogbiscuit

    Apollo

  • *****
  • Posts: 1,212
  • Joined: 18 Oct 2018
  • Loc: Barksdale

Posted 26 January 2020 - 12:59 AM

Thanks Pinbout for those Foucault simulations.


  • Pinbout likes this

#137 ccaissie

ccaissie

    Surveyor 1

  • *****
  • Posts: 1,945
  • Joined: 13 Sep 2010
  • Loc: Whitefield, Maine

Posted 26 January 2020 - 08:58 AM

and so would have I... but Jim's program is correct.

 

and Chris should place the ke there and measure the RoC again to verify...

 

making better optics is verifying with more than 1 test.

bueno.   one surface, many tests.



#138 dogbiscuit

dogbiscuit

    Apollo

  • *****
  • Posts: 1,212
  • Joined: 18 Oct 2018
  • Loc: Barksdale

Posted 26 January 2020 - 09:14 AM

Going only on your Foucault test numbers and what looks like probably a good edge in the not so good Ronchi image:

 

 

In post #132  Foucault Test Analysis program shows for those test readings wavefront errors and Strehl are in the marginally acceptable range.  Not that you should quit now, but if you couldn't get it any better it's good enough. a best fit conic is -1.21 or you could say 121% correction.

 

It shouldn't be too difficult to cut in half the current deviation from a parabola.  About 110% correction or .025" +/- .0025" knife movement between zones gets errors in the good mirror, pretty well inside the acceptable range, strehl near .96.

 

You might do better than that.  Maybe not. 

 

Now for a comment of figuring. For now lets look at the sawtooth profile made from percentage of correction.  Looking at the graphs, at the boundary between zones you want those kinks, slope changes, straightened out.

Upward changes of slope are under correction, less than 100% of target correction between zones.

Downward slope changes are over correction, more than 100% of target correction between zones.

No slope change is 100% correction between zones.

 

 

For a shape like yours where all the slope changes are downward, over correction.

As a general rule, as you stroke whatever pattern you have chosen, there is a central area of the mirror that is always in contact with the lap. The size of this area depends on the maximum overhang of the stroke. If maximum ovehang is 1/3 diameter, then the central 2/3 diameter is always in contact with the lap. For a 2/3 maximum overhang the central 1/3 of the mirror is always in contact with the lap.  Whatever the size, that central area always in contact with the lap tends go toward spherical. 

 

That's what you want to do to your over corrected mirror.  It's overcorrected in all zones so short strokes have minimum overhang, large central area of the mirror always in contact with the lap.  Overhang can be chosen depending on the location of an overcorrected zone. 

The 1/3 or less overhang is not enough to generate a significant increase of pressure at the edge of the lap, so even the outer area of the mirror that overhangs usually reduces correction also.

 

I'll tell you about increasing correction between undercorrected zones later.  Maybe you will have some before you finish.  Or maybe not. 

You are close enough that any single short figuring session could get you in the keeper range. 

 

 

 



#139 Chris Adamczyk

Chris Adamczyk

    Sputnik

  • -----
  • topic starter
  • Posts: 43
  • Joined: 18 Nov 2019
  • Loc: Southern California

Posted 26 January 2020 - 04:38 PM

I took the attached readings this morning and think I'm picking the data I want to see, I have one set of data that is too good to believe, The attached data was collected on a diameter rotated 120 degrees, I think I'll rotate another 120 degrees and collect an additional set of data.Chris 10inch F6(3)_Surface.gif Chris 10inch F6(3)_Trans_Err.gif Chris 10inch F6(4).gif 20-0126_Trans_Err.gif  



#140 Pinbout

Pinbout

    ISS

  • *****
  • Posts: 24,742
  • Joined: 22 Feb 2010

Posted 26 January 2020 - 04:41 PM

Yes that’s a more under corrected shape waytogo.gif

 

go star test it before you bag it and tag it to get coated. 
 

more tests = better confidence 



#141 dogbiscuit

dogbiscuit

    Apollo

  • *****
  • Posts: 1,212
  • Joined: 18 Oct 2018
  • Loc: Barksdale

Posted 26 January 2020 - 07:12 PM

How stable are the test conditions?  Temperature and air movement?

Plate glass and unstable temperature might change the shape of the glass enough to make a difference.

 

If you can't control the temperature in your testing space, usually there are two times during a day that are most stable, Late afternoon near sundown and early morning just before sunrise.  There is a time frame there where the temperature is neither rising or falling, sun is not heating the walls of the building your are testing in.  

 

10" f:6 on 1" plate I don't think would have unusually large flexure, but  it's worth thinking about the test stand.

 

Some variability of your readings are to be expected.  Air currents affect the light path, tester stand, mirror stand and even floor flexures cause distance changes between mirror and knife edge that sometimes  far surpass the small distances you are trying to measure.  I've even had wind conditions strong enough to see shaking during testing even though I was testing on a concrete slab floor.  Outside walls are supported by the slab so no real surprise.

 

Testing in a stable environment with the mirror and tester acclimated to room temperature,  Taking care not to move around on the floor or in your chair thus flexing the tester support while taking a set of readings,  averaging multiple reading can help overcome some of these things and get accurate test results.

 

Weather is another significant factor for testing in a space that is not climate controlled.  Overcast, fog, steady rain,  rizzle rain are some things that frequently occur with steady temperature and fairly stable air.  If you can't control the conditions, wait until the conditions are good, or at least don't miss a chance to test when conditions are good.

 

It might be you are letting some bias affect your readings.  find the null by eyeball, then note the reading.  Don't think about if it is more or less than what you want.  Just write it down and continue with the test until finished.  Then analyze.

 

I think more likely trouble is testing during different parts of the day with different rising or falling temperature and some of these other things I mentioned giving you different results.

 

You are doing good.  Remember that all of these test results have been quite close to a good parabola.  You are trying to measure a few millionths of an inch.  It doesn't take much bending of light or bending of test stands or floor to throw you off a little.

 

Think about these things I've mentioned and maybe you can think of ways to make your testing more reliable.



#142 Pinbout

Pinbout

    ISS

  • *****
  • Posts: 24,742
  • Joined: 22 Feb 2010

Posted 26 January 2020 - 07:24 PM

lets see a 4lines ronchi...



#143 dogbiscuit

dogbiscuit

    Apollo

  • *****
  • Posts: 1,212
  • Joined: 18 Oct 2018
  • Loc: Barksdale

Posted 26 January 2020 - 09:45 PM

Don't do any of the previous suggestions for figuring until you are satisfied you have reliable test results.

You've had results showing under and over correction and very near 100% correction.

You might have a very good figure now.

 

Try to get some good Ronchi pictures.

Some good Foucault without the mask would help too.


  • Pinbout likes this

#144 Chris Adamczyk

Chris Adamczyk

    Sputnik

  • -----
  • topic starter
  • Posts: 43
  • Joined: 18 Nov 2019
  • Loc: Southern California

Posted 27 January 2020 - 09:24 AM

OK, I'll try and get a good Ronchi and a Foucault pictures. I was trying to test again last night, early evening as the temperature was dropping and spent a good hour plus collecting total junk, it wasn't until I was walking away in total frustration that I realized that the dropping temperature could be the issue. 

 

I took a 3 day weekend and was  testing at 5am

 

Chris



#145 dogbiscuit

dogbiscuit

    Apollo

  • *****
  • Posts: 1,212
  • Joined: 18 Oct 2018
  • Loc: Barksdale

Posted 27 January 2020 - 10:29 AM

Another consideration is smoothness and low friction of the lead screws.

If while taking a set of readings you bump the tester it only takes moving it 1 or 2 thousandths of an inch to cause variations of results of about the scale you are seeing.

 

Rough threads or too much friction require a heavier touch on the tester controls. 

Ideally any touching of your fingers to the tester would only be very light touch on the knobs to rotate the knob without exerting any pressures forward/back left/right that would alter the position of the tester.  Very very light touches to turn the knobs and impart only rotation to the knob.

 

A drop or two of oil on the threads might be enough to improve things.  Lapping the threads to a fine finish might help too.

 

I have seen some incredibly low priced micrometers.  Maybe you even have one in a junk box gathering dust.  It might be reasonable to buy one of those and saw the frame off and just use  thimble and the hub with compression nut for their smooth 40 turn per inch threads.  You could still use the dial indicator for readings.



#146 Jeff B

Jeff B

    Fly Me to the Moon

  • *****
  • Posts: 7,074
  • Joined: 30 Dec 2006

Posted 27 January 2020 - 10:48 AM

Looking good there Chris!

 

An inspirational thread for me.

 

Jeff



#147 Don H

Don H

    Apollo

  • -----
  • Posts: 1,227
  • Joined: 28 Dec 2006
  • Loc: Near Tucson, AZ

Posted 27 January 2020 - 05:23 PM

If you have a friend who knows how, have that person take a few sets of readings. The more you test and compare, the more certain you will be of your measurements. If both of your readings are close, you can average the results. If the are widely disparate, look for possible reasons for that.

 

Congratulations on being so close to a fine mirror. I always felt this was the point where it was most fun. And there were many times when I had a good enough figure, than wrecked it with a mistake, only to take hours to get it back where it was. But then when I got a little improvement, I am glad I kept at it every time I look in the eyepiece...

 

Another thing you could do, is make a 5 zone mask and take readings with it to compare to the 4 zone results.


Edited by Don H, 27 January 2020 - 05:26 PM.


#148 Pinbout

Pinbout

    ISS

  • *****
  • Posts: 24,742
  • Joined: 22 Feb 2010

Posted 27 January 2020 - 10:36 PM

Put the mirror in a scope

and star test it

 

https://youtu.be/QxUQJjjsdW4



#149 Chris Adamczyk

Chris Adamczyk

    Sputnik

  • -----
  • topic starter
  • Posts: 43
  • Joined: 18 Nov 2019
  • Loc: Southern California

Posted 27 January 2020 - 10:52 PM

Foucault 1.jpg Foucault 1.jpg Thank you all for so much encouragement, I may need to built the scope, sona tube or truss? Tonight I took some pictures Foucault and Ronchi, the really strange and worrying thing is to get a Ronchi outside of focus I have to move the testing stage back, why does it not stay in the same position as for the Foucault test is something drastically wrong??  

 

Attached Thumbnails

  • Foucault 2.jpg
  • Ronchi inside.jpg
  • Ronchi outside.jpg


#150 Pinbout

Pinbout

    ISS

  • *****
  • Posts: 24,742
  • Joined: 22 Feb 2010

Posted 27 January 2020 - 11:55 PM

 

? Tonight I took some pictures Foucault and Ronchi, the really strange and worrying thing is to get a Ronchi outside of focus I have to move the testing stage back, why does it not stay in the same position as for the Foucault test is something drastically wrong??

Your ke images look like your not performing the test correctly




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