Originally published at: A pixelated, animated Times Square billboard, circa 1940 - Boing Boing
…
3 Likes
Although the animation might have been originally created in the earlier decades, that video with a Bulova Accutron ad on display could not have been any time before 1960 when Accutron was introduced.
6 Likes
Ah. I was expecting someone to explain the technical feat because I don’t know what kind of analogue supercomputer would have been around to create this in 1940ish.
Still would have looked supercool in 1960 because it does now too!
2 Likes
Apparently Douglas Leigh came up with billboards like this in the 1930ies:
And you wouldn’t have to use a supercomputer; something very much like a Jacquard loom or a player piano would have sufficed.
5 Likes
My guess would be a line of about 6,000 switches acted on by a large, rotating drum with pegs in it to operate them.
2 Likes
I can take a guess at how this worked. The data was no doubt stored on punched paper like a player piano roll.
The simplest way to read it would be spring loaded contacts that complete a circuit if there is a hole in the paper. But how do you flip from one frame to the next without smearing/blurring as the strip moves? You need something that can store a frame and load a new one all at once.
In the 1930s, the answer was relays. One per bulb in the sign. In its simplest form, a relay is an electromagnet that pulls a set of contacts closed when energized. You can have multiple pairs of contacts, and you can make a relay latch by powering a second coil from one pair of contacts, keeping the rest closed as well. To release the relay, you have to cut the power to the latching coil. That makes it a digital memory cell: it retains whatever the input was, even after the input is gone, until power is cut.
Assume you have a big loop of paper with all the frames punched into it. When a frame came into position in the reader, a single “load” contact somewhere outside of the frame hits a hole and closes a relay that applies power to all the other contacts in the frame. Wherever there is a hole, power is applied to a latching relay, turning that bulb on.
As the frame starts to move out of position the “load” relay opens and the other relays no longer get power from the reader. No more relays will be triggered by holes in the paper as they move, but the ones that were previously triggered stay latched and the bulbs stay lit.
Right before the next frame move into position, a “clear” contact hits a hole and causes a different kind of relay to pull a pair of normally closed contacts open. This cuts all power, releasing the latched relays. A fraction of a second later, the “load” relay is triggered again and the next frame is latched. Because bulbs keep glowing for a while, there’s no flicker.
5 Likes
The animations are stored on a film loop (the video switches from sign to film at 2:36), then they scan it somehow, probably related to early television technology, into a 1-bit stream.
How do they play that stream onto the sign matrix? 
A line buffer with at least one vacuum tube per line pixel is a whole lotta tubes! A similar number for the line selector.
Each bulb would only be lit for an instant, but persistence of vision would be their friend.
By the 1960’s it is very possible they used film and a flying-spot scanner (or just a whole lot of photocells) along with flip-flops made with tubes. Computers had been using them for a while.
Back in the 1930s it would have likely been electromechanical, I described how that might work above.
3 Likes
A latching relay per pixel would solve a lot of problems, but at ~50 lines and a similar number of pixels per line, that’s around 2,500 relays, ouch.
I wonder what the frame rate is? I don’t see any artifacts as an update sweeps across the frame. (Maybe they have a frame buffer by doubling the relays? Double-ouch!)
1 Like
I think the footage here is from the 1960s so it might use a system like you described. It may even be transistorized.
Someone else mentioned similar dot-matrix signs existed in the 1930s, and they likely had a much lower frame rate.
If you look there are some pixels that seem to flicker between on and off, suggesting there is some kind of optoelectronics that occasionally get a value right on the threshold.
This topic was automatically closed after 5 days. New replies are no longer allowed.
