Collimators, Autocollimators and Infinity Focus

[ic-racer]

Collimator: There are numerous devices used in many fields called "Collimator." With respect to film photography and achieving proper infininity focus on the film plane, a collimator is a device which projects non-diverging light rays to simulate objects at infinity ( a distance of 1000 times the focal length).

This is very handy when one does not have a convenient distand object within view of one's workshop. With a collimator, there would be an image at infinity, on which one would focus the test camera. One would need to observe the film plane and the infinity stop of the lens (and rangefinder, if present) to make sure all are in agreement.

There are a number of resources on the internet describing using a camera as a collimator. Mike Elk has a great write-up of this technique: https://elekm.net/zeiss-ikon/repair/collimate/

 

[vandergus]

I use the same technique on my workbench. One detail that always seems to get glossed over in the tutorials is to make sure you place your target, the ground glass behind the lens you are adjusting, on the inner rails. That's the film plane. The outer rails are for the pressure plate and typically sit about 0.15 mm behind the film plane. There's also some debate about whether to factor in some amount of back focus. Because the film isn't pressed flat against the film rails it often sits and few hundredths of a millimeter behind the film plane. I don't really have an opinion on that, though.

 

[ic-racer]

Autocollimators: The autocollimator goes one step farther. Not only does it project an object at infinity, there is a coincident viewing system (frequently using a beam splitter) that allows one to view the projected image.

For example, an autocollimator used for camera testing and repair, projects a star pattern on the film plane through the camera's lens. Via the insturment's eyepiece, the image can be observed to see if it is in focus.

 

[ic-racer]

I use the same technique on my workbench. One detail that always seems to get glossed over in the tutorials is to make sure you place your target, the ground glass behind the lens you are adjusting, on the inner rails. That's the film plane. The outer rails are for the pressure plate and typically sit about 0.15 mm behind the film plane. There's also some debate about whether to factor in some amount of back focus. Because the film isn't pressed flat against the film rails it often sits and few hundredths of a millimeter behind the film plane. I don't really have an opinion on that, though.

Excellent point! And, yes, getting infinity set 'correctly' is a real art, requiring compromises.

 

[ic-racer]

This is a schematic of the National Camera C-6005 made in Japan by Pearl:

 

[ic-racer]

A couple of Gokosha collimators: (https://magicflexcamera.com/)

 

[ic-racer]

There are some downloadable rescorces on the internet that describe the operation of these autocollimators.

 

[Andreas Thaler]

Excellent topic!

I've always wanted to know something about this, but couldn't find a general introduction.

 

[ic-racer]

This is a good introductory description of how it works:

 

[ic-racer]

In addition to infinity and focus scale setting, the autocollimator can check for parallelism. For example, one would want to ensure the front stardard or lens mount is parallel to the film plane. In a 35mm camera it would be checked like this image below. A similar method can be used to set the front detends on a 4x5 or other view camera.

 

[ic-racer]

Autocollimators can also be used to determine the focal length of lenses. For example, there was a question about measuring camera eyepiece diopters.

 

[Mr Bill]

Excellent topic!

I've always wanted to know something about this, but couldn't find a general introduction.

I agree; thanks ic... ; fwiw we (where i worked) used to have a couple of those older Gokosha autocollimators; the one that looks like a 1950s sci-fi ray gun) in the in-house camera repair shop. Lots of things you can do with em. Expensive, though; as I recall they sold for upwards of $5,000 US from the North American distributor circa 1990s.

Andreas, the first excerpt looks like it's from an old Edmund Scientific booklet series from about 1960s. They used to have them online somewhere on their website. Today... I think there was a spinoff company that still has the pdf files online. I consider the main author of that series a master of explanation, making things understandable to experimenters who are untrained in such things.

Update... I found this in a quick search: https://www.cloudynights.com/topic/...scopes-edmund-book-now-available-free-access/

 

[ic-racer]

The focus scale of the c6400 collimator looks like this. In this case the displacement from infinity is inbetween 2.3 and 2.4 mm.

 

[ic-racer]

Expensive, though; as I recall they sold for upwards of $5,000 US from the North American distributor circa 1990s.

I found this For Sale listing in the 1978 Camera Craftsman magazine.

Cost for this used c6400 collimator in 2024 US Dollars would be: $4,540

 

[ic-racer]

Just to review, there are two places to make readings with the device.

1) The vernier focus scale (focal length, curvature, focus settings and infinity)
2) Vertical or Horizontal displacement of the reflected image to the reticle (parallelism)

 

[Andreas Thaler]

@ic-racer @Mr Bill Thanks!

 

[ic-racer]

In terms of design, most of these devices for cameras are pretty similar in their specification.

A longer focal length of the collimator lens allows for measurement of smaller angular displacements.
A larger objective lens on the collimator allows for a greater range of angular displacement measurements.

 

[ic-racer]

If one wants to measure flange-focal distances (without a camera body), one needs a calibrated lens base.

I think these are very hard to come by. I'm still searching through "Camera Craftsman" issues to find more information.

 

[ic-racer]

This is one of the few pictures I could find. This one is for a horizontal cine camera/lens collimator.

One does not need this if one has a camera body that is of a known flange-focal distance. In that case, the collimator, after focusing it, will read the offset and then the flange-focal distance can be calculated from the known distance of the camera body.

 

[OAPOli]

I'm currently building a DIY autocollimator. It's a WIP and I'll make a thread when it's done.

What's really neat about this device is that you can place a film in the camera and use it as the reflector. This way any bulge or concavity can be inspected and compensated. 120 film in particular can bulge outwards. And in principle you could test for different films and whether the bulge changes over time.

Ground glass or plain acrylic screens also yield a reflection, so you can finely match the position of the focus screen to that of the film for reflex cameras. Unfortunately the fancy machined screen (e.g acute-matte) don't. I'm not sure how one would calibrate those. You can always check the projected target directly on the screen with a magnifier, but it's not nearly as accurate as catching its aerial reflection on the autocollimator.

 

[reddesert]

From the Edmund Optics technical documents archive linked above:

https://www.edmundoptics.com/ViewDocument/CollimatorsCollimation.pdf

https://www.edmundoptics.com/ViewDocument/Collimating Systems.pdf

Note that photographers and camera technicians tend to use "collimating" to mean setting the focus distance accurately (so that infinity matches a parallel bundle of rays). While telescope users often use "collimate" to mean aligning an optical system so that the optical axes match, for example getting the optical axes of the primary mirror, secondary mirror, and eyepiece of a telescope lined up in position and angle. Both of these uses are correct, but there is sometimes confusion. They reflect (haha) the different alignment needs of typical photographic vs telescopic systems.

 

[Alex Varas]

In addition to infinity and focus scale setting, the autocollimator can check for parallelism. For example, one would want to ensure the front stardard or lens mount is parallel to the film plane. In a 35mm camera it would be checked like this image below. A similar method can be used to set the front detends on a 4x5 or other view camera.

View attachment 382258

I wonder what kind of glass plate is it used, it would be very handy to know the parallelism is well set with the autocolimator.

 

[OAPOli]

I wonder what kind of glass plate is it used, it would be very handy to know the parallelism is well set with the autocolimator.

I think plain glass would work (although maybe with some flatness certification). In the setup illustrated above you would see the main reflection from the mirror at the film rails, as well as fainter reflection(s) from the glass surface(s) lying on top of the flange mount. If the planes are parallel, the reflections will be co-axial. If not, they will be shifted.

 

[Mr Bill]

What's really neat about this device is that you can place a film in the camera and use it as the reflector. This way any bulge or concavity can be inspected and compensated. 120 film in particular can bulge outwards. And in principle you could test for different films and whether the bulge changes over time.

Yes, I had played with this sort of thing with one of our Gokosha autocollimators (I forget the fl). These had a threaded collar that allowed fine (and precise) adjustment of the objective lens so that different distances could be mimicked, looking into the front of the collector. With the collar set at zero it was an infinity target, and could be verified via a flat mirror in front of the autocollimator. If the film position changes one simply rotates the autocollimator barrel until focus is again achieved (via the autocollimator eyepiece). Then read the distance off the barrel. Using the classic formula 1/f = 1/u + 1/v one can figure how far the film surface moved. Note: f = autocollimator fl, u and f are the conjugate distances of the lens.

As a note the autocollimator is mainly looking only at a small area near the center of the film.

 

[ic-racer]

I wonder what kind of glass plate is it used, it would be very handy to know the parallelism is well set with the autocolimator.

The parallelism is comparative. The reflecting mirror at the image plane does not have to be exactly perpendicular. One needs to ensure that the fainter image from the glass on the lens mount overlaps the brighter image from the film plane.