"Which, for a clock showing HH:MM:SS, means 72 light bulbs to replace."
??? There are only ten digits, and you don't even need all of them everywhere. By my reckoning, you would need 3+10+6+10+6+10=45.
If you add the ability to display dd-mm-yy, you till would only need 3+10+6+10+10+10=49. I wonder what they did with the remainder. Spelling out the word "ALARM!"?
Or did they simply ship spare bulbs inside the product?
I would imagine they wanted 6 identical modules just for the manufacturing efficiency.
That still leaves the question of why 12 lights instead of 10. I would bet that the separator colons are actually produced by a separate mask+lamp, rather than having different masks for each digit on those two modules. I.e. to show "8:" you light up both the "8" lamp and the ":" lamp at the same time. That gets us up to 11. No theory on the 12th one though.
Almost certainly this. In one of the links by another commenter here, you can see a voltmeter where the guy changes the decimal. The dots are bizarrely high to be a decimal point, but right where you'd expect it if it's the bottom dot on the colon.
Suspect it's just that the devices are a generic component, the application for projecting digits is just one possible application of a projection indicator. The physical constraints of the design require a rectangular projection array. 12 projectors, a 3x4 array, fits in a reasonable aspect ratio area. A 3x3 array would not suffice for this application, a 4x4 array would be overkill, and a 5x2 array, while optimal, would not fit well in the space available.
These displays were made with many different mask options, including symbols and letters. I have a set where one mask is red (to give an optional red background to the displayed digit). For metering applications there were + and - masks. Etc.
Oh, those things. I used to use UNIVAC 1108 computers, which had the system clock on the console with those displays.[1] It took about a hundred small printed circuit cards to drive the thing. It actually was the system's time of day clock, read by the computer, not a remote display of a counter elsewhere. If the muffin fans under the console were not regularly cleaned, the clock would make intermittent errors, which caused the operating system to have problems. I once had to write software to detect this.
When I was a kid, my dad would sometimes bring home scrap electronic equipment from his lab and I'd have fun taking it apart. I remember one of those displays, with the arrays of tiny bulbs.
14 comments
[ 5.2 ms ] story [ 28.7 ms ] thread??? There are only ten digits, and you don't even need all of them everywhere. By my reckoning, you would need 3+10+6+10+6+10=45.
If you add the ability to display dd-mm-yy, you till would only need 3+10+6+10+10+10=49. I wonder what they did with the remainder. Spelling out the word "ALARM!"?
Or did they simply ship spare bulbs inside the product?
The detail for what the extra masks were for is unclear, but it does look like there were 12 per digit.
That still leaves the question of why 12 lights instead of 10. I would bet that the separator colons are actually produced by a separate mask+lamp, rather than having different masks for each digit on those two modules. I.e. to show "8:" you light up both the "8" lamp and the ":" lamp at the same time. That gets us up to 11. No theory on the 12th one though.
(I still write it like that, in handwriting.)
http://www.decadecounter.com/vta/tubepage.php?item=10
http://www.decadecounter.com/vta/articleview.php?item=989
One company that makes them is still around:
http://www.ieeinc.com/about-us
A lot of other interesting old display technologies here:
http://www.electricstuff.co.uk/count.html
[1] http://www.silogic.com/Athena/Univac%201108.html
Here's an interesting channel on youtube I came across and the newest video happens to be about the NIXIE tube. https://www.youtube.com/watch?v=jytF5bvPGcU
http://thegarage.jalopnik.com/project-make-the-weirdest-peri...