Afocal setup with small USB sensor?

[patta]

There are many threads about afocal setups with DSLR; and about using the phone camera.
I'm putting together something in the middle, an USB camera with lens (like a webcam), looking into the eyepiece.

I couldn't find this setup described before, sound strange. If somebody know about that, or have suggestions, please tell me!
Before embarking in the tedious manufacturing of a proper mechanical adapter, and maybe expensive/useless purchases..

The goals are to:
1) use inexpensive, fast USB camera with small sensor
2) use compensating eyepieces
3) avoid hard-to-find low magnification photo eyepieces
4) get the full field of view

A CCTV lens is mounted on the USB camera, focused at infinity, and they are placed above the eyepiece, just like the phone camera. The CCTV lens can be chosen so the angular field of view matches the eyepiece. Then Eyepiece + CCTV lens work as an adapter/reducer with chosen magnification.
The image quality of the unstable prototype seems acceptable for my modest standards; more or less what I get visually from the eyepiece. Next it just need some stable mechanical mount. See illustration.
I think this setup could be very suitable for playing around with objectives and compensating eyepieces from different brands and ages, that is my main hobby lately.

A valid alternative would be to get a second-hand phone and mount it as permanent afocal camera
that may be good; but the phone camera lens has usually much wider FOV than the eyepieces, so need to crop.

.

Afocal_USB_camera.jpg (154.36 KiB) Viewed 9574 times

.

Afocal camera pairings.jpg (123.33 KiB) Viewed 9574 times

[MichaelG.]

The smaller Moticam cameras use afocal ... and their C-mount lenses seem to work quite nicely.

MichaelG.

.
P.S. __ I’ve never used him, but : Ted Pella lists the Motic parts
https://www.tedpella.com/cameras_html/camera1.htm.aspx

[patta]

Oh yes, looked now, the Moticam eyepiece mount that's exactly the same setup
Great, it is official!
They also sell separately the eyepiece mount adapter and a dedicated 12mm C-mount objective for reasonable $$. Hm, maybe I'll get it instead of 3D printing plastic.
But I'd like to keep the capability of swapping the M12 lenses so to fit different eyepieces

[MichaelG.]

Grab some of these adapters : https://www.ebay.co.uk/itm/2-x-S-Mount- ... 4858946999

That’s the listing that I ‘won’ but they appear to have more

Silly-cheap, but nicely made.

MichaelG.

[Hobbyst46]

The afocal USB camera idea has been mentioned I am sure, but was never tested in depth. Thanks, patta, for working on this project.

[MichaelG.]

The afocal USB camera idea has been mentioned I am sure, but was never tested in depth.

My Moticams work very well; but they are old and slow, and only 800x600 resolution … so that doesn’t really qualify as ‘tested in depth’.

MichaelG.

.
Note: As patta has already deduced : it’s helpful to use a reasonably long focal length lens on the camera.
… Too wide-angle will see the insides of the eyepiece !
Motic’s choices are very sensible: 8mm, 12mm, 16mm, according to the sensor format.

[Hobbyst46]

My Moticams work very well; but they are old and slow, and only 800x600 resolution … so that doesn’t really qualify as ‘tested in depth’.

Well, I only meant that I did not see a comprehensive test, images taken afocally with various compensating eyepieces and objectives, images of stage micrometers that demonstrate the presence (or better, the absence) of aberrations, etc. At any resolution.
So I look forward with anticipation.

My own inexpensive USB camera arrived with a 0.5X reducing lens, no 1X lens.

[MichaelG.]

My Moticams work very well; but they are old and slow, and only 800x600 resolution … so that doesn’t really qualify as ‘tested in depth’.

Well, I only meant that I did not see a comprehensive test, images taken afocally …

That’s fine … I was just clarifying that I don’t consider my Cameras good enough to do an exhaustive test.

MichaelG.

[patta]

Exhaustive testing, it will take some time, "too many projects" here too...
first i have to fix properly the USB camera that is messing up with the drivers.
If the drivers get fixed, I'll do some shots to see how this afocal setup stands against eyepiece projection and other methods

[MichaelG.]

For what little it is worth … this is a [slightly cropped and re-sized for the forum] Stereo pair, taken using the two Moticams afocally on my Nachet NS50
.

CBEDF4AF-4F09-4232-8C36-B5718FF05B31.jpeg (101.46 KiB) Viewed 9533 times

.

MichaelG.

[Scarodactyl]

Looks like it works pretty well.
I have a Leitz '0.1x' c mount adapter that must go over an eyepiece for afocal. I haven't tried it but it makes sense that some of the big dogs would have tried this approach

[patta]

About that last Leitz 0.1x, I have a confusion;
there are several types of "camera adapter":

one type is bundled with current cameras, from big dogs & chineses; typically 0.5x or 0.3x;
my understanding is that it is a complete relay lens, meaning, it is the whole assembly [eyepiece + M12 cctv objective] as above, takes the image from the intermediate plane and project it to the sensor.

the other type, like this Leitz 0.1x, is a different thing: it needs to be mounted after the photo eyepiece
the photo eyepiece was designed for, like, 35mm film format, and this 0.1x extra adapter focuses the image nearer and smaller, to fit into a small CCD chip
it works as "close up lens" for the photo eyepiece.
or maybe it is a second relay (the photo eyepiece being the first) with 0.1x magnification?

while the Moticam adapter is just the M12 CCTV objective, that takes the image from the visual eyepiece (projected toward infinity) and focuses it on the sensor.

...

For what little it is worth.... using the two Moticams afocally

well at least shows that the Moticam system works, do not vignette, no horrid haze or reflections, they're both in same focus... how is the field of view compared to visual? Almost the same or considerable crop?
I've looked at MoticEurope and they sell only the 12mm lens, at a reasonable 50 Eur. No other focal lengths. Motic US, different catalog! The afocal seems to be discontinued there..

[Scarodactyl]

the other type, like this Leitz 0.1x, is a different thing: it needs to be mounted after the photo eyepiece
the photo eyepiece was designed for, like, 35mm film format, and this 0.1x extra adapter focuses the image nearer and smaller, to fit into a small CCD chip
it works as "close up lens" for the photo eyepiece.
or maybe it is a second relay (the photo eyepiece being the first) with 0.1x magnification?

while the Moticam adapter is just the M12 CCTV objective, that takes the image from the visual eyepiece (projected toward infinity) and focuses it on the sensor.

These two options are the same in this case, because Leitz photo eyepieces also project the image to infinity--their official systems were mostly or all afocal in that era.
Olympus did use the other option, resizing the image from a projective eyepiece, but I don't know how that works (which is unfortunate because it would helo a lot with other formats poorly covered by olympus bh2 era stuff by default)

[MichaelG.]

[…] how is the field of view compared to visual? Almost the same or considerable crop?

.
Sorry, I can’t recall at the moment [must have made notes ‘somewhere’ ]
I will check later.

MichaelG.

[patta]

, because Leitz photo eyepieces also project the image to infinity--their official systems were mostly or all afocal in that era.
Olympus did use the other option, resizing the image from a projective eyepiece

Aha, big ignorance from my side, I thought that all photo eyepieces were made to be projective, focusing directly on the photographic plate. I always wondered at exactly what distance this plate should be!
So instead there were/still are, both projective (focusing at finite distance) and afocal (focused at infinity) photo eyepieces, that instead need a camera lens.

[MichaelG.]

UPDATE:

Quick & Dirty check on the NS50, with Zoom set at about 1.25 on a scale of 1 to 4
Note: Camera sensor is 1/4” and the Motic Lens 8mm

Visual field width 19mm gives Photographic field width 10mm
.. no significant barrel or pincushion distortion, and no significant Chromatic Aberration
There is a little softening noticeable at the edges, so presumably a slight curvature of field

Purely my subjective assessment, because behaviour may differ with other eyepieces.

MichaelG.

[PeteM]

A classic approach to afocal was to use a Nikon Coolpix with a screw-on 10x objective. Several articles out there on this approach. Objectives with a 28mm thread worked perfectly as relay lenses. These included a plan Leitz with Leitz corrections, a Nikon model that seemed fairly neutral, and third party ones with neutral corrections.

The last Coolpix that worked this way was the "5000" with 5mp resolution -- better than most USB cameras and a decent sensor for the time. The Coolpix models would allow zoom of the image, remote shutter release, and direct transfer to a PC.

It's still a viable approach IMO - the image quality beats most dedicated USB camera eyepieces up to about the $300-400 range and offers better image size matching.

One can often find the used bits, including the camera, 10x photo relay lens, etc. for under $200.

[patta]

Ok, so the theme has been explored more than I knew...
Good to know, the Moticam is keeping only the central part of the view; likely to avoid the periphery of visual eyepieces, that is usually not so good.

I'll go further on my road, USB camera + cctv lens in afocal over eyepiece.
Reason is, I have already all components: an USB camera, an unreasonable hoard of compensating eyepieces and CCTV lenses.

Price and practicality-wise, not sure it is optimal. Buying the components for this setup (cctv lens, eyepieces, adapters, camera excluded) will be going to cost ~100US$ or more; and old visual eyepieces, won't be necessarily great for photography.
-The Nikon Coolpix way looks proven and convenient
- Decent eyepiece adapters, like 0.5x C mount, cost 50 US$, and that's it
- The Moticam adapter (Lens, C mount, eyepiece mount) is 60 + 12 Euro...
- At 100 US$ one can start buying true used camera adapters of the specific brand, like the Leitz 0.1x, or new knock-offs
- Then true proper microscope cameras start at 400$, good ones 1000$ and +, so if one gets a good camera, would be worth to invest some money also in the dedicated adapter, instead of fiddling with glue and vintage eyepieces.

This setup may have some niche application. Let's see how it works.

[patta]

Here, a photoshoot test. The subject is well boring, but at least is flat and could keep track of magnification.
I've tried different sensor/lenses, all in afocal; with the same objective + eyepiece, Leitz NPL 16x 0.40 and Leitz Periplan 10x (visual eyepiece).

The actual field is pretty small, could be covered by just 2 Megapixel. Other things like color, contrast, noise, sensitivity, live view capability, raw image, that's another story, I don't know yet. It's basically the first time I'm trying to take decent large photos....

First illustration is a collection of the whole images, so to see the how the field of the eyepiece (circle) is projected on the sensor, with vignetting or cropping.
Second illustration, from the same images as before, center and field crops of the 10 micron spaced bars of the calibration slide. Resolution of the objective should be 0,7 micron, but from the images isn't clear to me whether it is achieved or not.

Images 1-2-3, the setup as post #1, an USB camera, 5 megaPixels, with M12 CCTV lenses of different focal length.
Image 4, from the phone camera
Image 5, a modest small webcam "HD", 0.9 megapixel, built-in lens.
Images 6 and 7, an APS-C DSLR with prime lens, 35 and 50mm respectively.

The USB+ M12 system works well, easy to mount and center. The M12 lenses are compact and with wide aperture, so it is easy to capture the eyepiece exit pupil, which is often an issue. The camera instead is likely a bit crappy, or I don't know how to use it.
The phone, works easily also, but the image is a bit small and needs heavy cropping, so we lose a little bit of resolution. Also the phone always adds its own image processing, like sharpening, smoothing etc, that I don't like for this use.
The small webcam, it is hard to mount because the lens is so small, it needs to be collimated exactly with the eyepiece. Then, it works, not so bad for its 8 $. Apart for the date timestamp!
The APS-C camera, of course give better images; but it is hard to focus, because needs to push high ISO and the live view is dark and noisy. Also, the 50mm lens is Autofocus, the front of the lens is delicate and moves, mounting the camera there to the microscope, will likely result in damages. I liked more the 35mm, that is an old, sturdy manual focus.

Now I'm thinking that a standard surveillance camera may be just enough for live-view and filming: FHD - 2 Megapixel, decent speed and light sensitivity.

I tried to sort out the parameters of the system, but stumbled in some basics; like, what is the focal length of the eyepiece?? or of the objective? I would like to get the diameter of the exit pupil, that is where afocal gets coupled.
Here link with details of the test and calculations attempt
https://docs.google.com/spreadsheets/d/ ... sp=sharing
I'll play in the future with some different objectives / eyepieces, try to get larger field, and maybe publish larger images in some Dropbox or blog
.

Afocal_photoshoot_1_1000px.jpg (121.74 KiB) Viewed 9345 times

.

Afocal_photoshoot_2_crops.jpg (130.41 KiB) Viewed 9291 times

Edit updated photoshoot 2 with more green squares
.

[Hobbyst46]

Thanks for pursuing and posting this project!!

There have been such projects with focally mounted USB cameras; yet, for example for the old Zeiss microscopes, afocal would be interesting. The stage micrometer photos are very appropriate, since they can provide visual impression of the various aberrations. Although white or grey background is needed. I tried to identify CA in the non-cropped 5MP+25mm photo, by changing the background to white (MS Office picture manager), but did not see any.

[Dubious]

Glad you are pursuing this, as I am also looking for solutions. Those shots from the 5MP USB with 12mm look pretty good.

[patta]

About the color, in principle was brightfield with white light... but each camera has its own idea of "white". I posted the images as shot, no adjustments.
The 5Mp USB, in particular, at one point I've smashed its glass protector / infrared filter, so is working without it, and everything is a bit reddish.

As far as sharpness/resolution, yes, seems all right, and the 12mm M12 lens does really a good job (it's the most expensive of the three, at 20$. The 8mm was 2$!). Absence of aberration, we have to thank mostly Leitz, their NPL objective + Periplan eyepiece combo works very well. (an the field is pretty narrow).

But, big issue for this sensor, I'm not able to control ISO and exposure, goes only automatic, and does some built-in image processing; probably I'm not taking true photos, but just grabbing a frame from the video stream. Imaging software like Toupview or APT, don't even see the camera.
I'll try a bit again, but fear that it's really not fit for photography.

About M12 lens afocal by itself, that I'm pretty happy.

[Hobbyst46]

My inexpensive (~$70) no-brand USB camera is limited but exposure time can be controlled, with either Toupview or MICAM 2/3. White balance too, though not convenient - it only sets the bkg color temperature.
It is a C-mount camera. Its original 0.5X reducing lens is fitted inside a cylinder, ~28mm OD and 48mm long, that screws on to the camera. Into the other end, a 23.2mm or 30mm OD tube is added, and is inserted into the photo tube of the head.
For afocal on the Zeiss, the M12 lens should be neither correcting nor compensating.
Unfortunately, ordinary Zeiss eyepieces do not have a top thread. I have a 12x16 KPL, low eyepoint. And a delaminated 10X18 WF, high eyepoint.
Which M12 lens (of the various models that are listed on eBay) do you favor ?

[patta]

I've got my M12 lens on Aliexpress, mostly from a shop called Fulekan; just because they claim to be manufacturers, not sure it's true, but at least they have a clear catalog with specs, and answer messages.
Edmund Optics, or other professional suppliers, they have 1000's of models too, but I haven't tried.

You can either get a bunch of inexpensive M12 lenses (the ~3$, get 8, 12 and 16mm) or one good (~25$, 12mm) The cheap ones, you can buy from any seller, ebay or so.
Specs of those M12 are focal length, rated resolution in megapixel, and supported sensor size, F number. (make sure the vendor is listing them..) Example: 12mm, 3MP, 1/2", F/2.4.
Also some have IR filters other not, I don't think you need it since is already on the sensor. Cost the same.
The good lenses are rated like or 8-12 MPx or 4K, the cheap ones 1-3 MP or HD. I tested several and there is some difference between good and cheap, but not so much, the cheap ones 3MP seem to already split ok the pixels of my 5MP sensor. The expensive ones, are cleaner.
Lenses from Edmund optics, are always rated only 1 MP but are also expensive. I don't know.
Do NOT get the "starlight" lenses (large aperture f/1 or f/1.6) as they are very long and clash with the eyepiece exit pupil.
May get instead of the M12, true C-mount lenses, but I haven't seen some suitable (either large or very expensive) The M12 are usually very compact instead.

Which focal length. As from previous shootout, you may choose, see the images in the post. Either get the full FOV of the eyepiece (image 2, 12mm), or just the central crop (image 3, 25mm), so you discard the eyepiece edge that's always a bit nasty.
Calculating which lens you need, 1) what is the angular FOV of the eyepiece? Measure it by taking a photo with the phone; so you can also see how good/bad are the borders. Otherwise, you may try to calculate it, FOV = 2* ArcTangent( ( Field mm/2 )/ Eyepiece focal length); but I'm not sure it actually works (what is the focal length of the eyepiece?).
This angle should be compared with the Diagonal Field of view of the M12+sensor. Which size is your sensor?
Mine is 1/2.5", meaning 4.3 x 5.7 mm, diagonal 7.1mm (I have no idea what 1/2.5" relates to..) so the diagonal of the angular FOV should be again FOV= ArcTangent( (Diagonal/2) / Focal length) ). edit: 2*
Simplified formula for the two FOV to match:
Field diameter / eyepiece FL = Sensor diagonal / objective FL

More on mounting
You will need a CS- M12 adapter, as post#4. Should cost ~1$ unless is super precise brass, rest is shipping & cuts.
I have mounted the camera over microscope photo port, with adapter as post #1, the eyepiece sits in the middle, while the camera is fastened (glued) to a tube around; so I didn't need to mount it to the eyepiece. May need an eyepiece clamp if you want to fasten the camera directly to the eyepiece, or those conveniently threaded eyepieces. A cardboard roll with some tape may substitute all of the above...

Have a look at the Moticam adapter, together with its eyepiece clamp, it is a ready-made afocal, more expensive but sure it's easier if you don't want to fiddle with homemade adapters; see link at post (edit) #2, near the bottom "Microscope Camera Adapters / Accessories"

[Hobbyst46]

Thanks, patta, for the detailed informative answer.

[patta]

You're welcome!
will be good to sort out a design for 3D printed mechanical adapter

[patta]

I've just realized, this M12 lens setup can be used also over projection eyepieces, as from post
viewtopic.php?f=9&t=7642
The projection eyepiece project to the distance/size it was designed for (like, at 120mm toward a 6x7 plate) and the M12 lens is screwed in so focuses past infinity, and can then reduce the large-format image to a small sensor. I'll try after the weekend.

[BramHuntingNematodes]

I use a 16mm lens with a 5x compens eyepiece and get only a small amount of vignette. I might look into getting a 7.5x eyepiece. It's worth playing around with the eyepiece if possible.

[patta]

I didn't work much on this setup; instead, went back to the fundamentals:
What is the eyepiece doing?

I've measured Focal Length, eye relief, field diameter etc of the random selection of eyepieces I have, since those basic things aren't so easy to find in catalogs and literature.
Not about aberrations or image quality, just the basic measurements with "tape and cardboard" methods.
This half-day work has been incredibly useful, both understanding the principles and now I'll know what distances and angles are needed to set up afocal or projection imaging, a significant improvement from pure trial & error.

Key discoveries:

• The focal length of eyepieces is normally defined against the 250mm reference length. A 10x has focal length 250/10 = 25mm. For all eyepieces I've tested except one.

• The intermediate image position, that influences conjugate distance, is often different from the "standard DIN" 10mm, depending on eyepiece age and brand.

• For eyepiece projection, it is possible to calculate (estimate) the distance at which a sensor should be placed. All eyepieces I've tested can be used as projective over APS-C dslr (my Canon 600D) without vignetting (sensor diagonal=max field). Only extreme eyepieces, super wide and with short eye relief, won't work well, but I don't have them.

• Using a visual eyepiece as projective, pushes down the intermediate image, so the field stop is out of focus and the objective's conjugate distance is changed. For a 10x/20 projecting over APS-C sensor, this displacement amounts to hefty 20mm. Thus, the eyepiece should be mounted with 20mm extra height to restore objectives conjugate distance & parfocality. Still, the field stop is out of focus and vignettes a bit.

• For the projection eyepiece "Nikon PLI 4", the 4 does not mean 4x but reproduction ratio 4:1; Aha! It is a "7x" according to the 250mm definition, and was designed to project a field of 11mm to full format film, 44mm diameter, hence the 4:1. The only other projective I have, "Wild 15x Phot" instead is really 15x, and was designed to project with about 7:1 reproduction ratio. If those reproduction ratios are not respected, the field stop isn't in focus, and the intermediate image is misplaced.

• Eyepieces with narrow field have less stray light. The field stop acts as efficient baffle. Eyepieces with large field, like 10x/20, show a lot of reflections from the walls, hence lower contrast. Also, expensive eyepieces appear to have less stray light than cheap ones, thanks to the money baffling.

• Eyepieces for amateur astronomy have much wider field than those for microscopy. They are also available in good quality and reasonable price. But don't fit in the 23.2mm microscope tube, nor in the 30mm. That may be a future project.

All in all my impression is that it is not good to use normal eyepieces for projection over small sensors (APS-C included); also old projection eyepieces are not working as intended, unless cropped significantly. Or I'll just make a baffle to work as field stop in the right place.

Focusing the eyepiece nearer than intended also displaces its aperture (the exit pupil) and may cause extra aberrations; but, maybe another time.
Also knowing the "cardinal points" of the eyepiece would come usfefl for calculating where to put the sensor etc.

Below the drawing with the parameters and the "intermediate image push down" effect when using the eyepiece as projective. And how to measure the eye relief / exit pupil (image of condenser iris).

Related thread abut projective eyepieces, where I've scribbled a lot of confused ramblings:
The [b]focal length[/b] is normally def ... 10 = 25mm.

The table with measurements (2nd sheet):
https://docs.google.com/spreadsheets/d/ ... sp=sharing

Eyepiece measurements.jpg (136.75 KiB) Viewed 8362 times

Eye relief.jpg (39.14 KiB) Viewed 8362 times