DIY focus collimator

G

Greg_E

Veteran
Just "finished" my DIY focus collimator device. Been wanting to build one for a while to check the back focus adjustment on our ENG cameras at work, and to use to set up lenses after I've fixed them and check mount swap and adapter length, etc. When I bought my GH2 a while ago, it came with a bunch of misc. lenses, so for this project I choose a decent lens in an FD 100f2.8 breach lock that I couldn't convert to EF. Everything I've read says that you should use a long lens for this type of creation, so I thought the 100mm would be good. Last week I decided enough was enough and I was going to build this thing. Found a suitable (non-functional) Canon T50 on ebay for $12 shipped and when it arrived I wasted no time tearing it apart.

This first two things that must be done are to get the mirror out of the way, and get the shutter out of the way:





After removing the prism, the mirror is held up by aluminum tape. The shutter was brute force removed and cut off where I could get on it.

Then I needed to create a pattern and print it, I used transparency film and printed with a laser printer, there are other more simple ways to get this done too.



I'll need to add a light behind this, but you could also use paper and put a light in the viewfinder space and front light it. I may move back to this and front light, depends on the kind of trouble I have fitting a light behind the transparency material


Looks like this:



Shooting down through the lens (at infinity):



On my rig:





And the as seen from the GH2 taking camera:



and focus mag:



What you can't see in these pictures is that if everything is at the right distance, you can clearly see the texture of the transparency material and the individual dots of toner melted together. That's how I know I'm at the correct distance with this lens and this adapter (LensTurbo). (edit: I guess you can see the texture a little bit)

That's about it, go forth and modify! More info can be found on the web with explanations of why this works. I can repeat this here if needed.
 
Cool, you basically built your own version of the NULL Lens & NULL Target that I helped bring to market. I'd suggest working on that target design some. A Siemens Star will only ever tell you so much, but you could do line pairs or dot patterning that will give you a direct read on acuity. That's especially useful when you are dealing with a camera & lens system that needs shimming instead of adjustable flange depth. Cool!
 
Yeah, there is a lot I'd like to do with the target, but the resolution of the printer really starts to become an issue. That was why I needed to put the 4 way division circle in the middle, the lines just became a jumbled mess in the print. Really need a photographic process to make a really high resolution target (take a picture on film). If you have suggestions on target patterns, I can try to make them and experiment a bit.
 
If you do line pairs then you need them in both vertical and horizontal to check astigmatism and spherical aberration. Corner targets are also good to look at falloff, parallelism and field flatness. Diagonal lines are also useful for acuity.
 
If you live in a fairly large city, there's likely still graphic service places that will print film at 1200 dpi.

Use a vector file (like illustrator) and you'll be resolution independent (no pixelation). I used to have 35mm slides printed from photoshop files, even up to 8x10 portfolio transparencies, and the quality was fantastic. Someone must still do that??
 
I'm using a CAD program and it is for all intents resolution-less. I may try running one through a different laser, but I'm pretty sure I'm already getting 1200x1200 on these prints. The film does resolve a little better than paper too.

Here's another try, trying to make it look "nice" while providing some features.



The diagonal lines need some work, they get a little thin, and I improved the star with better detail so a smaller dead spot. Hard to see but the upper right angle lines are twice as thick as the lower right angle lines.
 
Pretty good, but I think you should add something to the corners of the frame, way outside that circle. It's important to see what's going on outside the center. If there is an issue with parallelism, the center could be in perfect focus but one corner or a whole side could go soft.
 
I'll need to cut more of the shutter curtain out of the way. That said, most lenses may not fill the entire image area, the above picture before I used focus mag. shows very little of the image area. I'll have to try a wider lens on the target and see if it still performs well, have a few 50mm kicking around.
 
Think I have the final configuration... Added an Electro Luminescent panel for a back light with AA battery inverter, seems to be plenty bright on the receiving end at f2.8 and ISO 160.







The backlight allows me to put the film back/pressure plate back in place to close everything up.

Here is what I see with the 100mm on the target and 24mm + Lens Turbo on the camera:



Now with a 50f1.8 on the target camera:

 
That is a sweet rig. Very useful. I have an old Olympus OM-1 I might have to convert.
 
Decided I need a few screw on ND filters for this rig, our ENG cameras are too sensitive to let this work in the best way. Had to throw in 1/16 ND in the camera body to get the iris open and I'd really rather set back focus with the clear glass in the path because you are more likely to need the critical focus indoors in poor light. I was thinking of making a dimming EL inverter, but that is more work than it's worth for this tool when I can buy some cheap ND filters to screw on the few times I need them.

Wish our ENG cameras had a 15mm or 9mm rod mount on them, would be a lot easier to set this up. The two tripod method isn't much fun.
 
You should make a jig for this. A tripod quick release plate for the camera that's hard mounted to a board of wood. hard mount the collimator rig to the board on a worm gear to raise/lower it. You could also add a rail to move it near/far from the camera lens. A trip to a good hardware store should get you all you need for not much money. Very basic tooling.
 
Hmm, that is a good idea. I do have some old VCT plates that I could screw down to "something" and I have some extra cheap rails that I could attach at that correct height to the "something" too. Would save a ton of time each time I go through the cameras. We don't need to go through them too often, I've put tape on the adjustment ring so the students can't make the mistake of turning the FFD adjustment when they are trying to get the Macro setting. Yes that really was a huge problem. We will soon be requiring both a math and a science, hoping that elevates our average because things were a lot better back in 2003 when I started this job.

But every now and then it needs to be checked, already found one lens that has an off center or tipped element that I really should send back to Fuji for repair. Not uncommon for the lenses to hit things like door frames (who can be bothered with putting the camera back in the case just to move it?).
 
A trip to the basement offered up the plate to attach everything on, looks like all I need is something to bolt on that will clamp a set of rods and maybe allow them to have the height changed a little for fine tuning. This is the mount for our oldest teleprompters back from the 1980's and huge CRT's. One of those never throw away something that might be useful.

 
Last edited:
I think the final final version, so much easier than a chart on the wall. Used an ND4 filter and that seems to be just about right for our cameras in 1280x720/59.94p mode. Had an extra plate for an old tripod, decided I might as well mount it on one once in a while. Plastic feet for bench top operation. Now the target can stay at the (close to) 15mm LWS standard height to fit on my rig and the work rig.

 
@Greg_E The email I have for you isn't working. I have an update on the 35-105 conversion. I've made some good progress.

35-105.jpg

fdtoef.com
 
Nice. I'll shoot you a message. Glad the lens wasn't too damaged to make this work.

For the rest of you, the 35-105 f3.5 is a parfocal zoom with a constant aperture, you can understand why I was interested in getting this converted. Need to pick up a 20-35 f3.5 which is also parfocal. Both lenses have a backfocus adjustment under the zoom rubber ring, google should help show the process.
 
Kicking this back up because I just got done putting some repaired lenses back on cameras. This collimator has been one of the best "cheap" builds I've done in a while. Really makes life easy when you have a bunch of cameras to check. Also nice when setting up a speed booster. This tool let me confidently send one of the repairs back because it wasn't focusing at the right point, infinity aligned with the 10 meter mark on the barrel, and the flange focus was pretty far off to match. Fuji turned that one around in like 2 days since it was warranty and upon return it was perfect again.

The only thing I would do differently is to try and find a light source that was less powerful. I may drop an ND gel between target and backlight one of these days. Not a big issue on cameras that have electronic ISO adjustment, but on our ENG cameras even 0db gain is too high.
 
You must log in or register to reply here.
Back
Top