How to photograph a profile?

[patta]

Hello, asking for help:
I'm struggling to find a way to take decent images of object profiles against bright background.
The subject are currently screw threads.

I've tried some setups, either with diffuse backlight, or using a telephoto lens as "condenser" to produce collimated light.
As objective, macro lenses.

The diffuse light, worked only with stopped-down aperture of the objective (f/16, f/32...) and keeping the lightning "grazing" near the subject contour.
Otherwise get bad artifacts, and the profile gets illuminated also in front, not neat, glaring.
(first photo, bad; second photo, f/16 & grazing backlight)

The "condenser" illumination worked badly with the 90mm macro, a bit better with 150mm. Here is the opposite, I need to keep the aperture open otherwise the objective won't catch the whole size of the subject, unless I "defocus" the illumination.
(third photo: objective 90mm f/5; fourth, 150mm f/9)

Now the last image starts getting decent, but still missing some understanding of what's going on and if there are simpler ways instead of build unstable assemblies of objectives.
And I'm scared that the actual working aperture is from the illumination, that is very small (to get"parallel lightning"), so very poor resolution.

There are widespread "Profile Projectors" instruments for industrial inspection of parts (made also by Mitutoyo...); apparently they work neatly for this job, but I can't understand how. Either they have some huge lenses and telecentric tricks, or they're just using very small aperture, hence modest resolution.

Illustrations:
1 - a commercial profile projector
2 - best scheme of profile projector optics I've found
3 - the schemes I've tried so far
.

.

.

.

Large diffuse backlight, objective 90mm f/5.6

.

Grazing diffuse backlight, objective 90mm f/16

.

Collimated illumination pinhole + 400mm telephoto, objective 90mm f/5

.

Collimated illumination pinhole + 400mm telephoto, objective 150mm f/9

[rjlittlefield]

The standard solution is to use telecentric illumination plus telecentric imaging optics. For discussion see https://www.edmundoptics.com/knowledge- ... lications/ .

For sufficiently small subjects, you might construct DIY imaging optics using some technique like described at viewtopic.php?t=18323 . (The corresponding theory is at http://www.photomacrography.net/forum/v ... php?t=1472 .)

A similar technique can be used to produce telecentric illumination from a broad source. But your technique with a pinhole (or slightly larger aperture, see below), placed exactly at focus, with the illumination lens wide open, will accomplish a similar result with less trouble.

And I'm scared that the actual working aperture is from the illumination, that is very small (to get"parallel lightning"), so very poor resolution.

I would worry about that too. I think you will have to play with the apertures to get a good tradeoff for your application.

--Rik

[patta]

The standard solution is to use telecentric illumination

For sufficiently small subjects, you might construct DIY imaging optics using some technique like described at viewtopic.php?t=18323 . (The corresponding theory is at http://www.photomacrography.net/forum/v ... php?t=1472 .)

--Rik

Ho great thanks for the links, this old DIY telecentric of yours looks the way to go! I didn't read before...
Hand-holding a lens seems to work with almost any objective... need to find a way to install it at the right distance.

[soldevilla]

A reflector telescope is a perfect parallel ray illumination system, placing the focus in the place where the eyepiece is normally placed

[patta]

Took a bit of time, I've embarked full-steam in this tentacular Telecentric endeavour.
A lot of pages to read, many ring adapters to make...

Up to now, I've measured the position of the entrance pupil for most objectives in the house; scrubbed my secret stash of vintage glass to find some suitable single lenses to use as add-on (included telescope objective!); found or made (almost) all necessary ring adapters.

I'm heading hopefully soon to a wrap and will test the various telecentric configurations to see what works best. If you have comments/suggestions, tell me!
I try to publish as much details as possible here: https://wordpress.com/page/patta1072853 ... ss.com/873

Funny find as for now:
Telephoto zoom (70-300) and long macro with internal focus (Sigma 150) both become almost-telecentric at closest focus, with the entrance pupil pushed away at long distance, one meter.

Below images:
-Macro duel, measuring the position of the entrance pupil - the perks of "obsolete" DSLR.
- some of the adapter rings...
.

.

[patta]

Here, the (resized to 1000px) infographic I'm drawing for the telecentric add-on lens:
Hope are a useful integration to the host of information already written in the previous threads.
.

.

.

.

[patta]

Keep you updated (spammed) on the progress about this amazing topic:

I've studied a bit more how works the optics of true machine shop "Profile Projectors" a.k.a "Optical Comparators"
It seems indeed that are made like the illustration "Contour illumination" of post#1 (turned out to come from a Mitutoyo brochure);
at least the old, standard models. Here the summary:

The objective has no aperture by its own; it is not telecentric by itself.
The illumination, instead, it is collimated (telecentric) and with a very small f# like f/20.
The objective becomes then telecentric since it takes all this collimated light, and its "aperture" is where the illumination (telecentric) decides.
This setup is essentially identical to a brightfield microscope with the condenser iris aperture heavily stopped down.

Therefore a "true old style" profile projector can be made easily with the configuration 3 from post#1: collimated backlight from a pinhole source; long macro at 1:1 with aperture full open.

More recent Profile Projector instead seem to have a true telecentric objective with its own aperture, the illumination with NA large and controlled; that is where I'm heading.

I've put the notes on this Profile Projector here:
https://patta107285337.wordpress.com/20 ... -are-made/

[patta]

it keeps rolling

I've published the first tests batch for the "close up lens" telecentric as suggested in post #2
As objective, a 90mm macro, Tamron old model without IF nor floating elements.
As add-on lenses, some achromatic doublets of various power I had around.
Total magnifications from 0.65x to 1.9x.

I'll try more combinations next.

Getting the assembly telecentric is pretty simple (just measure the entrance pupil position and the extra lens FL)
While getting it sharp, that is though, specially at the corners. Best results with a weak extra lens placed close to the subject, so not much added magnification.
The nice Raynox lenses may be the best way to go (but I don't have them)

Here the link; not polished nor collimated as O'Toole's tests, hope is useful.
https://patta107285337.wordpress.com/20 ... mron-90mm/

edit: done the same for a 55mm macro lens, with low magnification.

[patta]

Some more infographic, for the other method to get a telecentric lens from a standard one: put a stop at its focal point. I haven't properly tested that yet; can just say that it works too.
.

.

[patta]

Nearing to a finish (ending time and enthusiasm)
I've tested the last method of "displaced aperture", and it works quite well for high magnification and NA (3x, NA 0.1)

[rjlittlefield]

I've tested the last method of "displaced aperture", and it works quite well for high magnification and NA (3x, NA 0.1)

Exactly which method is this? What objective are you using?

--Rik

[patta]

Thanks again for the enthusiasm-boosting ping! It is just the normal prime lens used inverted as macro, over bellows or over a tube lens, like the illustrations of previous post.
To make it telecentric, you put an aperture stop in the middle, some 30mm away from the prime's last lens, and that's it. Works with almost any prime objective, just the stop must be placed at the right position.
Image quality is like, slightly worse than, normal inverted prime without the extra stop. NA 0.1, was bit of exaggeration, corners not truly sharp, but works.
It may get better using some big enlargement lens (like, 100 or 150mm).

I'm more into low magnification now (1x, 0.5x) so those tests were quick and few.
Dump links of posts:
Illustrations and theory (same images as previous post)
https://patta107285337.wordpress.com/20 ... ft-method/
Test 2016 by Keith Cooper
https://www.northlight-images.co.uk/diy ... ro-lenses/
Test inverted 58mm prime and 55mm macro
https://patta107285337.wordpress.com/20 ... ure-shift/
Test two identical primes 1:1, bi-telecentric, symmetric, zero distortion, zero lateral CA, blah blah, (with tradeoff)
https://patta107285337.wordpress.com/20 ... -standard/
DIY telecentric homepage, slowly taking final shape
https://patta107285337.wordpress.com/te ... aging-diy/

[patta]

Enough for this year, time for holidays!
I've more or less wrapped up the results. The original issue, projecting profiles, maybe next year!

Attached here below the summary plot of all Telecentric setups I've tested. As you may guess, we want large field and high resolution.
Most are not so good, a few are ok for me. Details in the links above.
.

[patta]

Almost forgot, the "profile projector" is up and running since a while... I tend to abandon projects once they're done!

Here below the scheme: On the right, the camera with macro lens (Tamron 90mm here) and an add-on lens in front to make it telecentric (see previous posts)

The illumination is backlight coming from the left, another objective with add-on lens so it is telecentric too. It is kept there by an old film camera, with the back open; instead of the film there is a diffuser, illuminated by the lamp.

Illumination and imaging are collimated, so the camera registers the shadow of the subject. Very similar to a "brightfield" microscope, with the illumination optics on the left doing the job of the condenser. Both work at the same aperture.

It works ok for my purpose; magnifications I've managed, between 0.5x and 4x; the cylinders (screw) subjects however limit the usable f#, I'm going with f/8.
.

.

[patta]

Here below three sample photos of an inspiring microscope screw thread (that was the main initial purpose!)
All at nominal f/8
1 - The "collimated backlight illumination" as previous post
2 - same objective, but standard diffuse backlight
3 - some sort of poorly arranged side diffuse illumination

See also the 10 yrs old post about m26 threads, with well made diffuse illumination and higher magnification and NA
viewtopic.php?f=25&t=12404&p=78357&hili ... cam#p78357

1)
with some effort the background of the collimated illumination can be made more uniform

2)

3)