There’s been a couple of threads on the Dope which have referenced it here of late. I’m wondering if it would be possible to build one today and have it work with current broadcast signals. I realize that the quality of the picture would probably pale in comparison to that of a modern TV, but it’d be kind of nifty to have one.
It would be possible to build a mechanical television receiver today, using patent specifications. But it definitely could not pick up modern television signals.
How mechanical television worked:
http://www.mztv.com/how.html
In 1996 at Toronto’s MZTV Museum of Television, a working mechanical television system was demonstrated for the first time since the 1930s:
Here’s a GIF animation of a mechanical TV image:
http://users.pandora.be/ON1AIJ/nbtv_008.gif
Unless you have some high-powered signal processing CPUs in the front end, your mechanical TV can’t deal with NTSC television signals. The Baird standard had 30 vertically-scanned lines, rather than 625 horizontally scanned lines.
Apparently people still stumble across wax record albums with the recorded data for mechanical TV images. See:
http://www.dfm.dircon.co.uk/articles.htm
Here’s an 8.75 minute .RM file of a section of a recent remake of the first broadcast made-for-TV play, 1930 “The Man with the flower in his mouth”
http://www.dfm.dircon.co.uk/mwfihm.HTM
Grant Dixon’s page has good info and a link collection:
http://www.nbtv.wyenet.co.uk/
A steam powered TV! I like that! 
SILLYNESS ALERT
[sillyness]So that’s how those stone-cased mica-screened prehistoric t.v.'s worked in the Flintstones movie. Probably had a little trained monkey cranking the scanning disk inside. [/sillyness]
So don’t use the Baird standard. Is there any reason that a mechanical television could not be built according to the NTSC standard?
?CPUs??? bbeaty?? ?Can’t pick up a modern signal??? Walloon??
Quite off target.
Of course you can decode a TV signal to feed into a mechanical TV. 1930’s tech could do it, so no CPUs are needed at all.
The issue is the mechanics. NTSC is 30fps of 525 lines. (Pal is 25fps, 625 lines). Only half the lines are sent at a time, etc. but let’s ignore that part.
So you need a disk with ~500 holes (no need to have holes for the blanking interval) that can rotate 30 times per second. (Or twice as many holes and half as fast.) All doable. But …
You run into the issue of of “image retention”. Phosphors in CRTs retain their glow for about half of a frame. Not so with a mechanical system. In the classic systems you saw the dots for a line, then the next line, then the next. It relied on the eye’s image retention. But for 500 lines, that won’t work. You can draw each line simultaneously, slightly staggered, but that would require buffering of each line. Simple hardware but *500. Again, the eye would see an the image within a line “appearing” and shifting from left to right.
Now, you can add multiple holes per row and simultaneously show several segments of a scan line. More simple hardware. A lot more.
The bottom line: it’s “image retention” that is the headache and while in principle solvable, it would be a pain to fix.
BTW, That was for B&W, multiply hole count by 3 and put suitable color filters over each group.
I’m afraid it’s not the fist tim since the 1930’s. I saw a newly-built pair of mechanical TVs at an exhibit at the Whitney Museum in NYC back in the 1980s. I’m sure other folks must have tinkered with the same things in the intervening seventy years.
There’s a big reason. Think about it. Those old sets had only 30 holes, and the perimeter length of the scanning wheel is proportional to the number of lines and the size of the raster. Let’s say that the raster or “view screen” is two inches across. The scanning disk has one hole for each scanning line, with spaces between holes equal to the raster width. 625 holes, each spaced 2" apart, gives a disk with a perimeter of 104 feet, diameter of 33 feet, spinning fast enough so humans will see a raster. The hobbyist standard is 12Hz, pretty “flickery” compared to 60Hz modern television, but it lets the 33-foot disk turn fairly slow, only 720RPM. Better make it from carbon fiber, since the centrifugal action is going to be pretty fierce. With a perimeter speed of 820MPH it will also exceed the speed of sound and pump a huge amount of wind, so you might want to put it in a vacuum chamber. The sides of a flat vacuum chamber 33ft across will have a few hundred tons of air pressure, so you’d better use a spherical chamber for strength. It will occupy a four storey building.
All that, for only a two-inch “picture tube.” A mechanical NTSC TV camera would be of similar size.
The above is putting the signal onto a single wire (or a single radio channel.) If the disk scanned many holes simultaneously, and used lots of parallel channels for each image, it could be made proportionally smaller.
As I was originally thinking, if we keep the line count and screen size the same as those old mechanical sets, then we stay with 30 lines at 12Hz with a 20" disk, then we need a “slow-scan converter card” such as used by amateur radio hobbyists. It A/Ds and buffers the NTSC video in RAM and averages several lines together to create a much slower, few-lines raster output which is compatible with the old mechanical TV sets.
Another way to do it is half mechanical, half electronic, as with the “Private Eye” glasses-mounted computer display. That device has a vibrating mirror for the horizontal scan, and a couple of hundred tiny LEDs scanned fast electronically for the vertical.
Hey, I saw that same art exhibit in 1989 at the ICA in Boston. It was by that guy who was famous for nailing his hands to the hood of a VW Beetle back in the 1970s. Aha! “Chris Burden” (found by search terms +crucified +artist +“vw beetle”) Here’s a blurb from the show:
http://neil.franklin.ch/Usenet/alt.folklore.computers/20001226_Mechanical_Television
There is another way to generate TV signals via mechanical scanning-the German Telfunken system used a scanning wheel which projected a “flying spot” of light onto the scene to be transmitted.Banks of photelectric cells picked up the resulting reflected light. Of course, that meant that the scene had to be done in semi-darkness.
Mechanical ain’t the way to go!
From the article you cite:
Hey, when I saw the exhibit in NYC in the eighties (before 1989) it was already broken! I wonder if it ever worked? It wouldn’t surprise me if he went to all the trouble to put something like this together and never actually got the damned thing running – artists seem to do things like that. But it would be a much better piece of conceptual art if it did work.