On a Whim
by Mat in Vadnais Heights

It’s time for a little bit of nerd history. Most days I am all about high technology, but I am also a fan of clever low-tech solutions to classically knotty problems. Well, relatively low-tech in this case. Get ready for the bizarre yet weirdly comprehensible bit of analog tech called mechanical television.

These days, of course, your typical flatscreen TV has lots and lots of teeny little LEDs in each of the primary colors, which are toggled on and off 60 or 120 times per second. Do this in the correct sequence, and you can replicate any full-color moving image you like in real time. The first TVs, though, did not do anything like this. They were more of a weird collision between a zoetrope and a telephone, with many hot, bright lights.

Let’s see if I can describe this thing. Take an old LP record. We’ll call it R1. Grab a marker, ruler, and protractor and draw lines to divide it into 30 pie-slices.

Pick any line at random. Measure 4.75″ from the center. Drill a tiny hole, about 1/16″ across, on the line at that point. Proceed to the next line. Drill the same kind of hole, except place it about 1/32″ closer to the disk’s center. Repeat until every line has one hole.

You now have a spiral of 30 holes that span one full revolution of the disk. The innermost hole is about one inch closer to the center of the disk than the outermost. This is called a Nipkow disk, after Paul Nipkow, who patented it for image-transmitting purposes in 1884.

Now, on a second record, call it R2, cut a 1″ square window that begins 3.75″ from the center and ending, of course, 4.75″ from the center.

Place R2 on top of R1. If you put a light source behind this setup and spin R1 clockwise, keeping R2 stationary, you will see a tiny bright dot sweeping across your 1″×1″ screen. If the dot moves fast enough it blurs into a raster line (rastering is how CRTs and tube TVs worked, though the lines were generated differently).

If you spin R1 at 60 rpm you will see raster lines sweep across the whole little screen once per second. At 600 rpm, it’s 10 times per second. At 1800 rpm, it’s 30 times per second. Somewhere in there, your persistence of vision will blend all these raster lines into a single square image.

If you slow this back down to, say, 5 rpm, you can see clearly that at any given time there is just a single dot of light moving across the screen, top to bottom. If there is an illuminated object behind R1, then you will see the brightness of that dot vary as lighter and darker areas of the object are revealed. Now let’s speed it back up to persistence-of-vision territory and go on to the next bit.

It’s all (relatively) straightforward from here. What we need to do now is read the intensity of the light coming off the screen with extremely high fidelity. Fortunately, you can do this with the element selenium, which changes its electrical conductivity depending on how much light is falling on it. So … pass a current through a bit of selenium that’s positioned in the path of the light coming off the screen. What you get out of that is a modulated electrical signal encoding the brightness of the light at every moment. (This is what the original Bell telephone did to sounds via a crude microphone. We’re doing it to light via a hunk of selenium.)

Now we boost that signal using transistors or old vacuum tubes and transmit it (that’s the “tele” part of television) down a wire and out an antenna. We have an appropriate receiver at some remote location that picks up the modulated signal.

OK, but now that we have it at our remote location, what do we do with it? We pass it through a lamp that can vary its brightness extremely quickly. This recreates the bright/dark pattern recorded by the selenium. Then we pass that light through a second, identical Nipkow-disk-and-window setup which is spinning at the same speed as the original. The result: If you look at the teeny screen on the receiving end you will see that the moving-dot raster-lines perfectly recreate the moving pattern of lights and darks you’d see by looking at your original Nipkow disk’s screen. And persistence of vision again translates that into an image. A live, moving image that has been transmitted via electricity!

Now for the downsides. While this works as proof-of-concept, the problems are legion. First, the screen is tiny. You need a magnifying glass to view it properly. And the only way to make it bigger is to enlarge the disk. A mechanical TV with a viewable area of modern flat-screen dimensions would not fit inside your house.

Second, the image is dim. Only a tiny fraction of the available light makes it through the pinholes in the disk. So all your lighting on both ends needs to be painfully bright.

Third, the thing is insanely finicky about calibration. The screen must be positioned precisely atop the spinning disk, or the image won’t line up. If the rotation rate of the receiving disk is off even slightly, the image will “roll” or just disintegrate altogether. Any vibration in the spinning disk will blur the image. And a disk spinning at over 1000 rpm is gonna want to vibrate.

And finally, the overall image quality is just plain bad, and there’s no easy way to make it significantly better. For reference, U.S. tube TVs had 480 raster lines, and they are considered poor quality today. Our gizmo has only 30.

This doesn’t stop me from loving the thing. First, because it is an elegant (if impractical) solution to the problem. And second, because it launched a major industry predicated on a defect of persistence in the human eye. Which is hilarious.