Hello rec.photographers, Yes, I really have made my own 900 mm lens. I thought the readers of rec.photo.advanced might be interested in a little writeup. First, the good news: The lens is very sharp, virtually distortion- free, and has zero chromatic aberration. It will focus anywhere from 12 feet to infinity, and I built the whole thing for under $300.00. And now the bad news: It weighs 23 pounds, is 7 inches in diameter and over 5 feet long, and has a fixed aperture of about f/11. Also, being a mirror lens, it has the usual mirror lens funnies in out-of- focus areas of the picture. But the other good news is that the pictures taken with it look pretty darned good. This all started one weekend last May. I was hiking at Pawnee Buttes in northeast Colorado, and tried to get some shots of a ground squirrel. But my Canon EOS 75-300/4-5.6 just wasn't long enough and I wasn't able to get close enough. On the way home, I spotted an antelope in the grassland near the road. But again, he was too far away for much of a shot, other than a snapshot to prove I saw him. The next day, at a bird santuary, I wasn't quite able to get close enough to a yellow-headed blackbird perched on a cattail. So I got on the net and looked up the Canon EOS faq, and went over my options for more focal length. A teleconverter would be the cheapest, but since I don't have the greatest or fastest lenses to start with, I think I would be disappointed. I could go for the 300/4 with and extender, but then I'd just wish I had the 300/2.8 with and extender, and that would cost more than all the rest of my camera equipment combined. Of course a 600/4 would be great, now were talking real money. So I got to thinking whether I could make something. I'm something of a tinkerer, and saw this as a chance to learn a little more about optics. In my job I work with people who are optical experts, but I'm not one myself. I thought it couldn't hurt to pick up a little more optical knowledge. I have no way to grind glass, so I got out the Edmund Scientific catalog to look at what stock components might be available. They have a variety of large mirrors, and that's what I started with. I'd been working on a little ray-tracing program anyway, so I set up a model using a 36-inch focal length f/6 spherical reflector. Thirty-six inches is about 914 mm, which seemed like enough length, and because a camera requires a field of view that includes off-axis areas, I needed a large enough mirror that after it was stopped down some, it could still have a reasonable aperture. By the way, the mirror part number in the Edmund Scientific catalog is J32,842, and it lists for $181.20. You can also get an enhanced aluminum version (J43,566) for $240.20. They'll take a credit card number and ship the mirror right to your door. It's almost like ordering a pizza. The ray tracing looked pretty good. At f/11 and focused far away, I estimate a maximum spot size on the film of about 16 microns. This compares very favorably with Ansel Adams' statement that spot size should be less than 1/1000th of an inch, which is 25.4 microns. I'm also told that most camera lenses have a sharpness of about "30/30", or 30 percent modulation at 30 line pairs per millimeter on the film. (See some of the recent discussions of Modulation Transfer Function.) My analysis showed that (depending on the quality of the mirror) the lens should be at least that good. One of our optical engineers here at work also looked at it with a commercial ray tracing program and agreed. Of course, this all depends on the quality of the mirror. Why a spherical mirror? Why not parabolic? Well, the honest answer is that I tried a spherical model first because I already had the equations worked out for reflecting rays off of spherical surfaces. When that looked good enough, I just stopped there. But since then, I have worked out the formulas for tracing rays off a paraboloid. As expected, the on-axis performance is quite a bit better. But the off-axis performance is quite a bit worse. So all in all, I think I'm better off with the spherical mirror after all. The layout of the lens is something like a reflecting telescope. The main mirror is in the back, and then a flat mirror is mounted in the middle of the tube at a 45 degree angle, and the camera looks in from the side. I scrounged a good quality first-surface flat mirror from some parts left over from a former work project. The mirror sits in the optical path and occludes part of it, which is what causes the out- of-focus effects. It also reduces the light transmission a little. One significant way my lens differs from a telescope is that the main mirror is movable for focusing over a wide range. In order to get a good close-focus distance, I had to provide for about 8 inches of focus travel. I also learned that the focus distance versus focus travel relationship is very nonlinear. As the closest-focus distance gets very short, the focus travel grows very rapidly. So I can focus from infinity down to 12 feet with about 8 inches of travel. If I wanted to focus down to 6 feet, I would need _much_more than 8 inches of travel. Another tradeoff to be made was the location of the aperture. Placing the aperture at the center of the mirror sphere appears to make for the best sharpness. But a 36-inch focal length mirror has a radius of 72 inches, and that would make the lens _really_ long. Also, the farther the aperture is from the mirror, the greater the area of the mirror that is used, and therefore more vignetting will happen or the aperture size will have to be reduced. Placing the aperture at the center of the sphere also would be a problem as focus was adjusted. To keep the aperture and mirror a fixed distance apart, the camera would have to move for focusing, and that seemed awkward mechanically. I decided that vignetting was bad, and lens length was bad, and that the sharpness tradeoff wasn't too bad, so I placed the aperture about 6 inches in front of the camera. Its distance from the mirror depends on where the main mirror is in its focus travel. Now that I had an optical design, I had to figure out how to make it mechanically. In real life, I'm a mechanical designer for Hewlett-Packard, so I just sat down at the 3D CAD system and went at it. The lens barrel is made of 6-inch black ABS plastic pipe. I made a focusing mechanism out of a leadscrew and some bearings from a Hewlett- Packard optical disk autochanger. One interesting problem in the design was figuring out how to place the focus knob close enough to the camera so that you could even reach it while looking through the viewfinder. I found a way, but the leadscrew is outside the diameter of the pipe in a sort of rectangular protuberance. Another problem was how to mount the 45-degree mirror and the camera so they were in good alignment with each other. There again, I found a way, but it required borrowing some special machining tools from our model shop machinists. After all the parts were designed, I spent several evenings in the machine shop making all the special parts. The requirement that I be able to make the parts myself (with my limited machining skills) put some added constraints on the design of the parts. To attach the camera, I just bought a T-mount adapter for the EOS lens mount. I had to modify it a little to open up the hole in the center so that it wouldn't vignette. Incidently, the 45-degree mirror is rectangular, rather than elliptical as in a telescope. The reason is that the field of view needs to be rectangular to cover the whole frame. A telescope needs only a circular field of view. It would be possible to cut my mirror down a little and make it trapezoidal shaped, as some of the rectangle isn't really used now. That would make it obstruct slightly less of the aperture, but it hasn't seemed necessary. The barrel is made in three sections. The main mirror and focusing mechanism are in a section that's a about 20 inches long. The camera mount and 45-degree mirror are in another section about 32 inches long. The aperture mounts on the end of that, and then another section attaches with a coupling. The last section is about 15 inches long and is simply a lens hood. Would you believe that a coupling for 6-inch ABS plastic pipe costs $16.00? It was quite exciting to put it all together for the first time and look through the viewfinder. The lens gives a correct image - it's not upside down or anything. But the first few pictures were pretty bad. The contrast was terrible. I determined that the problem was stray reflections off of the inside of the pipe. Even though the pipe is very black, the inside surface is very shiny - almost mirror-like. The solution turned out to be very simple. At first I thought I'd have to design some elaborate baffles to kill the stray reflections. But it turns out that simply painting the inside of the pipe a very flat black was good enough. I used Krylon Ultra-Flat Black spray paint. Getting the correct exposure is also interesting. In the EOS system, apparently the lens tells the camera what its maximum aperture is. The camera uses that in its metering calculations, and then calculates what shutter speed will be necessary at the aperture that will be used to make the exposure. But my lens doesn't tell the camera anything. What I've discovered is that in the absence of data, the camera assumes an f/1.0 lens. So to get the correct exposure, I put the camera in aperture-priority mode and set it for f/1.0. Then the automatic metering works and the camera picks the correct shutter speed. In manual mode, I would also have to leave the aperture set at 1.0. And now the pictures look pretty good. I went out a couple of weekends ago to try it out. It's a little awkward to use, but I've got some techniques that help. I don't think my Bogen 3001 is up to holding this lens steady, so I bought a bean-bag chair to use as a lens rest. It actually works pretty well propped on the tailgate of my pickup truck, or even plopped down on the grass. I'm learing to look through the viewfinder, hold a cable release in one hand, keep the other hand on the focus knob, and maneuver the whole thing with both hands. And if I take the lens hood section off, I can carry everything in the cab of the truck rather than in the back. So I'm calling the whole project a success. I've got a workable (if awkward) 900 mm lens, I only spent $300 on it, and I learned a _whole_ lot in the process. And what's next? I suppose I could put a motor on it to turn the focus knob. It's pretty slow right now to move from close-focus to infinity or back. And then once the motor is there, maybe I could decode the command signals coming from the camera and make it auto-focus! I suspect that at f/11, this would only have a prayer of working in very bright sunlight. Or I suppose I could figure out how to have the lens tell the camera it's a fixed-aperture f/11 lens, so I wouldn't have to remember to lie to the metering system. But as things are now, I _can_ interchange apertures to get a little adjustment, so I think I'll leave the metering situation as it is. Or I could build a mongo tripod out of 2 by 4's. But for awhile, I think I'll just play with it and see what seems important. And now, as a reward for those who've read all the way through this long posting, here's an offer. If you'll email me your regular mailing address, I'll send you some photos of the lens itself, as well as a couple of pictures that were taken with it. I'll do this at my expense for U.S. residents if there aren't too many requests. If you're outside the U.S., then email me and we can make some kind of arrangements. It will take at least a couple of weeks to send out anything because I want to catch the next full moon to be one of the pictures. [ Note added 11/4/94: Please note that this offer has expired. I guess you'll have to make do with online images. Thanks.] Later, Dave Boyd Hewlett-Packard, Greeley, Colorado P.S. Come to think of it, maybe I'll get that teleconverter anyway. Surely I can think of some use for an 1800mm f/22 lens!