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

number of lines to use in DP test

  • Please log in to reply
19 replies to this topic

#1 RichA

RichA

    Surveyor 1

  • *****
  • topic starter
  • Posts: 1,940
  • Joined: 03 Jun 2010
  • Loc: Toronto, Canada

Posted 10 July 2020 - 11:53 PM

I remember someone once told me maximum 3 lines to really get an idea of figure, roughness.  Even using 1 thick band and letting it pan across the image works well.  But I've seen people using 6 and 7 bands.  Is that too many?

 



#2 Pinbout

Pinbout

    ISS

  • *****
  • Posts: 24,500
  • Joined: 22 Feb 2010

Posted 11 July 2020 - 06:09 AM

5 at the most, 2 at the least. 1 if you want to use it like ke to see any zones pop.


  • DAVIDG and Jeff B like this

#3 MKV

MKV

    Cosmos

  • *****
  • Posts: 8,884
  • Joined: 20 Jan 2011
  • Loc: Florida, USA

Posted 11 July 2020 - 08:22 AM

But I've seen people using 6 and 7 bands.  Is that too many?

Yes it is. As Pinbout said up to five is considered okay in ATM community. I think five is pushing it.  I don't know where this "rule" comes from, and what "science" it's based on, but there's not doubt that fewer is better. I prefer 2-3 bands max. 

 

Fewer bands show more curvature and zonal errors are also easier to detect.

 

This can be shown using a 6-inch f/8 parabolic mirror, whose OPD error will be 1/4 wave pvw at ROC. It's obvious with 2 bands and not at all with 8 bands using a 100 lpi screen. 

 

ronchi No of bands.png  

 

My suggestion is: when you get two or three straight Ronchi bands using the Ronchi test (regardless if it's at ROC or in autocollimation), switch to the knife-edge test for a final assessment.


  • PrestonE and Jeff B like this

#4 Pinbout

Pinbout

    ISS

  • *****
  • Posts: 24,500
  • Joined: 22 Feb 2010

Posted 11 July 2020 - 08:39 AM

Yes it is. As Pinbout said up to five is considered okay in ATM community. I think five is pushing it.  I don't know where this "rule" comes from, and what "science" it's based on, but there's not doubt that fewer is better. I prefer 2-3 bands max. 

 

Fewer bands show more curvature and zonal errors are also easier to detect.

 

This can be shown using a 6-inch f/8 parabolic mirror, whose OPD error will be 1/4 wave pvw at ROC. It's obvious with 2 bands and not at all with 8 bands using a 100 lpi screen. 

 

attachicon.gifronchi No of bands.png

 

My suggestion is: when you get two or three straight Ronchi bands using the Ronchi test (regardless if it's at ROC or in autocollimation), switch to the knife-edge test for a final assessment.

the title said in DP which is 2x's as sensitive. 

 

and as your working the mirror you want to scale the sensitivity as you get closer to what is possible.

 

that's why Ceravola uses ronchi in his ross, then ends up using a ke.

 

its called scoping. 

 

and of course he follows up his ke/ross test with IF which proves with a ke and ross he gets to 1/60~. 


  • LarsMalmgren likes this

#5 Jeff B

Jeff B

    Fly Me to the Moon

  • *****
  • Posts: 6,905
  • Joined: 30 Dec 2006

Posted 11 July 2020 - 08:42 AM

In DPAC I like to use 3 bands, one in the center...because it looks nice...and it gives good resolution, especially towards the edge with a 133 LPI screen.  I also use the screen as a knife edge of sorts to look at the surface.  

 

My example is an 8" F7.5, both in single pass Ronchi and DPAC.  I love DPAC as it tells you almost instantly if you have a good mirror system.  The problem, though, is finding a decent, large aperture flat.  Mine is 12".

 

Jeff

Attached Thumbnails

  • Fagin OTA, Inside, Null, Outside, DPAC.jpg
  • Fagin, F7.5 Inside ROC.jpg
  • Fagin, F7.5 outside ROC.jpg
  • Fagin, F7.5, At ROC B.jpg

  • PrestonE, LarsMalmgren, Pinbout and 1 other like this

#6 Pinbout

Pinbout

    ISS

  • *****
  • Posts: 24,500
  • Joined: 22 Feb 2010

Posted 11 July 2020 - 08:47 AM

here's a 6in f8 qrt under in dpac, 133lpi. 4 lines and 2 lines. outside focus

 

6f8 dpac.JPG

 

6f8 dpac qrt 2lines.JPG

 

 

and the same 6in f8 qrt under outside RoC, 133 lpi (not dpac...duh)

 

6in f8 qtr under roc.JPG


Edited by Pinbout, 11 July 2020 - 02:18 PM.

  • Jeff B likes this

#7 Jeff B

Jeff B

    Fly Me to the Moon

  • *****
  • Posts: 6,905
  • Joined: 30 Dec 2006

Posted 11 July 2020 - 09:16 AM

Danny, a dumb question.  How did you do a mirror DPAC without a central obstruction?

 

Jeff



#8 Pinbout

Pinbout

    ISS

  • *****
  • Posts: 24,500
  • Joined: 22 Feb 2010

Posted 11 July 2020 - 09:20 AM

Danny, a dumb question.  How did you do a mirror DPAC without a central obstruction?

 

Jeff

computer generated... i could put a hole there is you would like. ha ha ha

 

but there was a guy in AU said he did... I forget his setup... it was posted here... but that was like 6 -8 years ago when he was working on his 18-20 mirror.


Edited by Pinbout, 11 July 2020 - 09:21 AM.

  • Jeff B likes this

#9 MKV

MKV

    Cosmos

  • *****
  • Posts: 8,884
  • Joined: 20 Jan 2011
  • Loc: Florida, USA

Posted 11 July 2020 - 09:25 AM

and the same 6in f8 qrt under outside RoC, 133 lpi 

How can it look corrected with two bands if it shows undercorrection with four bands? 



#10 Jeff B

Jeff B

    Fly Me to the Moon

  • *****
  • Posts: 6,905
  • Joined: 30 Dec 2006

Posted 11 July 2020 - 09:53 AM

Magic.


  • MKV likes this

#11 MKV

MKV

    Cosmos

  • *****
  • Posts: 8,884
  • Joined: 20 Jan 2011
  • Loc: Florida, USA

Posted 11 July 2020 - 11:48 AM

Magic.

LOL, I trust rayracing. Two-band Ronchi looks curved in autocllimation.

 

6in f8 sphere dpac_a.png



#12 DAVIDG

DAVIDG

    Cosmos

  • *****
  • Posts: 9,640
  • Joined: 02 Dec 2004
  • Loc: Hockessin, De

Posted 11 July 2020 - 12:25 PM

  There is no reason not to test at 3 lines or less,  all  your doing is changing the position of  the grating which takes 3 seconds !  You also want to watch has you move the grating from one side of focus to other and back again. The lines can look pretty straight when stationary but as you watch  them move you can pick up bowing  that you miss when they are stationary. A  well corrected mirror,  lens or optics designed for an object at infinity will show dead straight lines no matter how many and on both side of focus. So as you move the grating from ones side of  the focal plane to other the lines should stay straight just the number will decrease as you approach focus. 

    As for testing  a concave mirror without  a hole in the flat,  I do it all the time. You set up the test in a  Newtonian configuration with a diagonal. 

 

               - Dave 


  • Pinbout likes this

#13 Pinbout

Pinbout

    ISS

  • *****
  • Posts: 24,500
  • Joined: 22 Feb 2010

Posted 11 July 2020 - 02:11 PM

How can it look corrected with two bands if it shows undercorrection with four bands? 

a 6in f8 -1 conic undercorrected by 1/4~ is almost a sphere so straight lines at RoC



#14 Pinbout

Pinbout

    ISS

  • *****
  • Posts: 24,500
  • Joined: 22 Feb 2010

Posted 11 July 2020 - 02:16 PM

How can it look corrected with two bands if it shows undercorrection with four bands? 

Outside RoC is Obviously not in dpac

 

and your showing inside of focus. why? outside always shows what you really have. Inside is always backwards.


Edited by Pinbout, 11 July 2020 - 02:19 PM.


#15 Steve Dodds

Steve Dodds

    Owner - Nova Optical

  • *****
  • Vendors
  • Posts: 502
  • Joined: 02 Apr 2007
  • Loc: Utah

Posted 11 July 2020 - 02:21 PM

What I do is get it to 3 lines outside then scoot it inside 3 lines and compare the two.  Back and forth back and forth, significantly more sensitive than just looking at one side.


  • Pinbout and sopticals like this

#16 Pinbout

Pinbout

    ISS

  • *****
  • Posts: 24,500
  • Joined: 22 Feb 2010

Posted 11 July 2020 - 03:07 PM

6in f8 1/16th under dpac 133lpi  flipping in and out... inside is about -6.5mm outside is about 5mm

 

click on the pic to see it animate.

 

6in-f8-16th-under-dpac.gif


Edited by Pinbout, 11 July 2020 - 03:08 PM.


#17 MKV

MKV

    Cosmos

  • *****
  • Posts: 8,884
  • Joined: 20 Jan 2011
  • Loc: Florida, USA

Posted 11 July 2020 - 04:12 PM

a 6in f8 -1 conic undercorrected by 1/4~ is almost a sphere so straight lines at RoC

Still the bands are curved. 6-inch f/8 conic -1 at ROC, 133 lpi.

 

6in f8 at ROC 133 lpi.png



#18 Pinbout

Pinbout

    ISS

  • *****
  • Posts: 24,500
  • Joined: 22 Feb 2010

Posted 11 July 2020 - 04:15 PM

Still the bands are curved. 6-inch f/8 conic -1 at ROC, 133 lpi.

 

attachicon.gif6in f8 at ROC 133 lpi.png

A qtr wave under ... if you please 



#19 MKV

MKV

    Cosmos

  • *****
  • Posts: 8,884
  • Joined: 20 Jan 2011
  • Loc: Florida, USA

Posted 11 July 2020 - 04:23 PM

A qtr wave under ... if you please 

6-inch f/8 parabola is quarter wave pvw off at roc as sphere is 1/4 wave pvw off at prime focus.

 

6in f8 cc_1 roc_a.png



#20 MKV

MKV

    Cosmos

  • *****
  • Posts: 8,884
  • Joined: 20 Jan 2011
  • Loc: Florida, USA

Posted 11 July 2020 - 04:37 PM

Here is the IF  wavefront analysis for a parabola at roc and sphere at f for 6 in f/8

 

6nf8 cc -1 roc pvw.png

6in f8 sphere at f pvw.png

 

 




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