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

Figuring a 10.5 inch mirror

  • Please log in to reply
65 replies to this topic

#1 dodo

dodo

    Explorer 1

  • -----
  • topic starter
  • Posts: 85
  • Joined: 12 Dec 2017

Posted 25 March 2019 - 01:26 AM

Hello everyone!

After finishing my 8.5" f/9 mirror with the assistance from fellow CNers waytogo.gif, I have been working with this mirror trying to use all whatever I have learnt from my first venture. This 10.5 " was actually fine ground two years ago when unfortunately during pouring the pitch lap I broke the glass tool and stopped the project. Now, I have cast a concrete tool and polished out the mirror and have done some parabolization using the standard Texereau strokes.I give below the details of the mirror as under-

D=265mm; ROC=3121.025 mm; Thickness of mirror= 19mm

I am using a 1 inch thick full sized tool and MOT. I have some ronchi/foucault pictures

inside.jpg outside.jpg focault centre.jpg focault 70%.jpg

I have also uploaded a ronchi video at  https://www.youtube....eature=youtu.be

I am however not sure how far/near I am from completion so I am unable to proceed any further.

So any assistance to how I will proceed will be of great help.

I will try to post some numbers from my Foucault test later. Thank You.

Forgot to add that this is plate glass.

 

 

Attached Thumbnails

  • focault edge.jpg


#2 dodo

dodo

    Explorer 1

  • -----
  • topic starter
  • Posts: 85
  • Joined: 12 Dec 2017

Posted 25 March 2019 - 02:02 AM

Here are the readings and results from the foucault test

10.5 inch.jpg

Are the ronchi and Foucault results pointing at the same issue? 

I am not sure but it appears to me that the ronchi is showing a better mirror and Foucault is showing it quite worse.confused1.gif 


Edited by dodo, 25 March 2019 - 02:07 AM.


#3 dogbiscuit

dogbiscuit

    Messenger

  • -----
  • Posts: 407
  • Joined: 18 Oct 2018

Posted 25 March 2019 - 03:53 AM

First thing to check if you think Foucault analysis doesn't look right...

Is your tester light source fixed or moving?

FTA setup shows fixed.



#4 dodo

dodo

    Explorer 1

  • -----
  • topic starter
  • Posts: 85
  • Joined: 12 Dec 2017

Posted 25 March 2019 - 04:38 AM

Yes. The setup is fixed source.

I am confused that while ronchi looks decent - at least to my untrained eyes, the foucault analysis shows such a poor result !



#5 dogbiscuit

dogbiscuit

    Messenger

  • -----
  • Posts: 407
  • Joined: 18 Oct 2018

Posted 25 March 2019 - 04:41 AM

How many lines per inch or mm is your grating?

 

I put your data into FTA.  What version of FTA are you using?

My version graphs the same as yours but best fit conic and strehl don't match.

And your analysis shows a negative strehl.  Something wrong there.



#6 dogbiscuit

dogbiscuit

    Messenger

  • -----
  • Posts: 407
  • Joined: 18 Oct 2018

Posted 25 March 2019 - 04:59 AM

Are you sure your light source doesn't move with the knife?

Your knife readings are more plausible for a moving source tester, although there must be an error in the readings.

Maybe the tester or mirror is on an unstable stand, or some other instability while you take readings. 

How many knife readings did you do to get the average entry shown in the screenshot?

 

Looking back at your 8.5" thread I see you used a 100 lpi grating.

I put your 10.5" data into Ronchi for windows.  Your Ronchi images show a pretty good match to a parabola.

 



#7 dave brock

dave brock

    Surveyor 1

  • -----
  • Posts: 1671
  • Joined: 06 Jun 2008
  • Loc: Hamilton, New Zealand

Posted 25 March 2019 - 05:04 AM

I'd check your readings. Zones 2, 3 and 4 are basically the same. Doesn't fit with your images.
  • dodo likes this

#8 Pinbout

Pinbout

    ISS

  • *****
  • Posts: 22409
  • Joined: 22 Feb 2010
  • Loc: Montclair, NJ

Posted 25 March 2019 - 06:17 AM

I’d fix the edge before worrying Bout correction


  • dodo likes this

#9 dogbiscuit

dogbiscuit

    Messenger

  • -----
  • Posts: 407
  • Joined: 18 Oct 2018

Posted 25 March 2019 - 07:22 AM

I played with the knife readings based on the possibility you failed to count a turn of the 1/4-20 screw micrometer that equates to 1.27 mm per turn.

Applied the difference to your entered readings for zones 3 and 4. 

 

here is what I got.

 

Fixed source tester.dodoKeModified.jpg

 

It looks about right, but of course it could be (probably is) wrong.

But you probably did miss 1 or more full turns of the micrometer.

I notice that your readings for zone 2 and zone 4 are identical.  So maybe zone 4 is 2 full turns from zone 2.

I'll keep the previously  modified zone 3 and change zone 4 to be 2 full turns from zone 2.

 

dodoKeModified2.jpg

 

Not really much different.

Of course this is all just playing to see how your reading might have gone wrong.  Don't consider either of these to be accurate.

I looked at your Ronchi movie and it looked to be fairly stable so I doubt that tester stand or mirror stand stability are the reason for your bad readings.

Maybe not exactly as either of my examples but I'm pretty sure you missed counting a full turn or two of your micrometer.

 

Take some more readings.


  • dodo likes this

#10 dodo

dodo

    Explorer 1

  • -----
  • topic starter
  • Posts: 85
  • Joined: 12 Dec 2017

Posted 25 March 2019 - 07:56 AM

I’d fix the edge before worrying Bout correction

True, but I was thinking of grinding/bevelling off any residual TDE if at all that could be done.

 

How many knife readings did you do to get the average entry shown in the screenshot?

I took five readings in total. At first I took two readings then took the rest three in a different mirror orientation.

So, I'll now take readings with my moving source tester and see how the readings go.

Btw, is there any software which automatically takes measurements while I capture photos/videos of my ke images with my camera attached to a PC?



#11 Pinbout

Pinbout

    ISS

  • *****
  • Posts: 22409
  • Joined: 22 Feb 2010
  • Loc: Montclair, NJ

Posted 25 March 2019 - 08:13 AM

post-286701-0-67776300-1553494631_thumb.

 

dodos 10in.JPG

 

at a glance it looks pretty good

 

but look at the top left of your ronchi... its flat - undercorrected...look at diffract how the white line continues to curve out and yours goes straight.

 

dodos real ronchi.JPG

 

didn't look at the ke reading to see if it correlates but the ronchi in this case doesn't lie.

 

fix the edge, add correction to the outer zone and forget about the pimple in the middle.


  • dodo likes this

#12 dave brock

dave brock

    Surveyor 1

  • -----
  • Posts: 1671
  • Joined: 06 Jun 2008
  • Loc: Hamilton, New Zealand

Posted 25 March 2019 - 03:28 PM

When you took the foucault pics hopefully the knife edge wasn't lined up properly.
  • Pinbout likes this

#13 Dale Eason

Dale Eason

    Apollo

  • *****
  • Posts: 1167
  • Joined: 24 Nov 2009
  • Loc: Roseville,Mn.

Posted 26 March 2019 - 12:26 PM

Something is really wrong with the Foucault analysis.  Strehl can never be negative or greater than 1.  So I don't trust those results.  I wonder how the software computes Strehl.  The usual method for this kind of data is to estimate the Strehl form the RMS.  But that could have never created a negative strehl.

 

The Foucult images are prettry good and show a turned perhaps even a rolled edge that can and should be fixed.  It shows the outer regions of the mirror to have little correction (mostly flat).  It shows a slightly rough surface.  

 

Dale


  • dodo likes this

#14 dodo

dodo

    Explorer 1

  • -----
  • topic starter
  • Posts: 85
  • Joined: 12 Dec 2017

Posted 27 March 2019 - 12:40 AM

I have ordered a dial indicator to resolve this Foucault reading mess.

Got it here  https://www.amazon.i...1_TE_item_image.

I would first like to reconcile the ronchi images with the Foucault readings before attempting any retouching.



#15 dodo

dodo

    Explorer 1

  • -----
  • topic starter
  • Posts: 85
  • Joined: 12 Dec 2017

Posted 29 March 2019 - 09:13 AM

Untill my dial indicator arrives, I have taken some readings with my moving source tester (Stellafane type). All these readings are the zonal differences. 10 readings in total.Attached File  READINGS.xlsx   9.24KB   6 downloads

I have two software- the FTA version 2A and the Foucault 1.51. I input the averages. Now please look at the results that I get from each of them-

FTA V2.0A.jpg

Attached File  FOUCAULT1.51.docx   96.59KB   8 downloads

and I am no software guy shocked.gif

Now I am curious about one thing. See in the FTA screenshot the program places a value in the average reading column against Z1. But won't this be zero? Well I have entered zero in Z1 in both the software.

Also none of the average readings shown in the analysis are my actual average reading and so the difference column may not be true!

Or, am I seriously missing something?


Edited by dodo, 29 March 2019 - 09:17 AM.


#16 Pinbout

Pinbout

    ISS

  • *****
  • Posts: 22409
  • Joined: 22 Feb 2010
  • Loc: Montclair, NJ

Posted 29 March 2019 - 10:14 AM

how did you measure your RoC btw?



#17 dodo

dodo

    Explorer 1

  • -----
  • topic starter
  • Posts: 85
  • Joined: 12 Dec 2017

Posted 29 March 2019 - 10:55 AM

how did you measure your RoC btw?

I nulled the centre by placing the mask on the mirror using the moving source tester. The ROC was 122 7/8" from the mirror centre to the ke. 



#18 dogbiscuit

dogbiscuit

    Messenger

  • -----
  • Posts: 407
  • Joined: 18 Oct 2018

Posted 29 March 2019 - 12:07 PM

There is something wrong with FTA (Foucault Test Analysis) version 2.0A showing a negative Strehl.

Replace that with version 2.0B.

 

For all practical purposes I get the same anaysis with version B that you got with version A, except for Strehl.

 

There is some confusion in mixing/converting mm and inches.

 

Best to pick the one that matches your micrometer calibration and set the program for that measurement unit.  That way you don't need to convert.

 

Whether the Z1 is set to zero or not is not important except that you need to under stand why.  :-)

The wave front is calculated from measured difference from Z1 to Z2, Z2 to Zone 3, Zone 2 to Zone 4.  The zones have finite width large enough to be seen from the tester.  The knife edge is used to measure average ROC of the zone, and it is assumed that the measured average ROC for the zone is at some nearly equal distance from the inner radius and outer radius of the zone, and that is referred to in the programs as the zone's effective radius.

The mask has no Zone zero where the the exact center is both the inner and outer radius of the zone and also the effective radius, so we never really measure the ROC of the exact center when taking the test readings.  We start with Z1 measuring the ROC for it's effective radius that is some distance from the mirror's center.

When you calculate the desired knife movements for the effective radius of each zone, the results for each zone will be a distance as measured fro the mirror's center. But you never measure the very center of the mirror, and Foucault analysis doesn't calculate from center, it calculates from movements between adjacent zones.

So all this means that you may if you wish just enter readings directly from your micrometer and let the program do what it does.

Or you can adjust your readings to have Z1 knife reading be zero.  To do this you subtract the mircrometer reading of Z1 from the micrometer reading of each zone.  This leaves intact the differences between zones.

 

You may ask "Why zero the first zone?"

When you compare raw readings from the micrometer, in various test sessions the readings from for Z1 might be very different and of course each of the other zones.  It is difficult to look at the numbers and immediately see how far they are from the desired knife movements.

Zeroing the desired movements, and then each test session knife movements, you can look at the numbers and know immediately how well your numbers match the desired movements.

 

Some micrometers could be zeroed on the Z1 while testing, but the tester would likely be disturbed while zeroing.  It is best not to zero the micrometer, but just note the raw reading from each zone.  

 

Some consider it good practice to take readings in both directions.  Start at Z1 and working outward,  and then another set starting at the outermost zone and working inward.

 

In some cases the software will display the desired readings in the Z1 zeroed format, in some the desired readings are not zeroed but shown from mirror center.

 

All that is what has you confused. 

 

I see why there is a FTA 2.00 version B.   :-)

Delete FTA version 2.00A and get FTA version 2.00B.

http://freewareapp.c...lysis_download/

 

 

 

I like FTA, but there are a couple of things about the wavefront graph that I don't like.  One is that the zone boundaries are not shown.  Another is that the lateral scale from mirror center to edge is is divided into 5 equal distances from center and the distance shown for each is rounded too coarsely.  Otherwise It is a very good program.

 

You can also try FigureXP and see how you like that.  I like the graph better.  :-)


Edited by dogbiscuit, 29 March 2019 - 12:14 PM.


#19 dogbiscuit

dogbiscuit

    Messenger

  • -----
  • Posts: 407
  • Joined: 18 Oct 2018

Posted 29 March 2019 - 12:25 PM

Sometimes I have mentioned that the wavefront is calculated from the differences between zone measurements.

I should clarify.  The measured differences are not entered as zone readings (except I guess a program could be written to accept that entry format).  

If  the measured difference from each zone to the next was .025" and you zeroed the reading of zone 1 you would have these modified readings to enter inot the program.

Z1  0

Z2  0.025

Z3  0.05

Z4  0.075



#20 Pinbout

Pinbout

    ISS

  • *****
  • Posts: 22409
  • Joined: 22 Feb 2010
  • Loc: Montclair, NJ

Posted 29 March 2019 - 12:38 PM

 

There is something wrong with FTA (Foucault Test Analysis)

it doesn't match your ronchigram either


  • dave brock likes this

#21 dogbiscuit

dogbiscuit

    Messenger

  • -----
  • Posts: 407
  • Joined: 18 Oct 2018

Posted 29 March 2019 - 01:17 PM

it doesn't match your ronchigram either

and there is something wrong with his readings.


  • dave brock likes this

#22 dogbiscuit

dogbiscuit

    Messenger

  • -----
  • Posts: 407
  • Joined: 18 Oct 2018

Posted 29 March 2019 - 01:55 PM

Concerning Foucault1.5.1

The program's units of measure is mm.

You have entered mirror data in inches.



#23 dodo

dodo

    Explorer 1

  • -----
  • topic starter
  • Posts: 85
  • Joined: 12 Dec 2017

Posted 29 March 2019 - 02:25 PM

Thanks dogbiscuit for the detailed explanation. waytogo.gif

I may detail my procedure of taking readings from the screw dial. My dial consists of 50 divisions with each division measuring 0.001". For example while measuring the central zone Z1, if the dial reads 33 and Z2 reads 21, I take the measurement as 0.038". Next if I read Z3 as 46 and Z4 as 28, I count the divisions from Z2(21) to Z3 (46) as 0.025" and from Z3(46) to Z4 (28) as 0.032". Now considering Z1 as zero I have the following readings 

Z1 as 0

Z2 as 0.0338"

Z3 as 0.025"

Z4 as 0.032"

But from what you have explained in post#18 and 19, my correct readings should have been 

Z1 as 0

Z2 as 0.038"

Z3 as 0.038"+ 0.025"=0.063"

Z4 as 0.063" + 0.032"=0.095"


  • dave brock likes this

#24 dodo

dodo

    Explorer 1

  • -----
  • topic starter
  • Posts: 85
  • Joined: 12 Dec 2017

Posted 29 March 2019 - 02:31 PM

So I was wrongly under the impression that the readings are the differentials between each zone. Now I realise that the difference is with respect to the central zone. 

Thanks again for making that clear. 


  • dave brock likes this

#25 dodo

dodo

    Explorer 1

  • -----
  • topic starter
  • Posts: 85
  • Joined: 12 Dec 2017

Posted 29 March 2019 - 02:35 PM

I will revise my readings accordingly and see what comes up hmm.gif

And yes I will get the newer version of FTA 


Edited by dodo, 29 March 2019 - 02:40 PM.



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