This reminds me of something I did some years back. I am somewhat myopic (4.5 diopters or so). I wanted a clock I could read from across the room, without my glasses, for my bedroom (so I could read the time upon waking up without having to find my glasses). Initially I would just roll over and haul the bedside clock up next to my face - but then when I met by now-ex and found myself no longer sleeping alone, there was sometimes someone between me and the clock. The largest stock clock I could find had digits only 1.8" high, I think it was. I picked up one of those and found that, as I'd hoped, the display was on a separate board, with wires forming a kind of coarse ribbon cable between that and the main clock board. I cut the wires and probed the display part with power and a resistor suitable for LEDs, to determine the pinout. Then I cut a piece of 1/8" plywood to the size I wanted the clock face to have (something like 8" by 24" - I'd have to go measure to be sure). Then I laid out (digitally) and printed out (piecemeal, on multiple sheets, with registration marks) a clock display 88:88 made up of five LEDs per segment. Using those sheets as a guide, I drilled out the plywood to take round LEDs at the right places. Pushed the LEDs through, soldered them up appropriately, and all that was left was to build the rest of the case and wire up connectors (I used DB25s) to the display and the clock board. I now have a nice wooden clock on my wall, with the electronics board from the original tucked in one corner driving nice big 7-stroke digits. I tried to interpose additional electronics between the clock and the LEDs, because the LED brightness was so variable I could only assume I was near the limit of the clock's drive capability (driving five LEDs instead of the two or so it was designed for - 7.5V drop instead of about 3V). I never managed to make it work; it appears that power-line AC direct from the power transformer is used to drive the two common poles (I forget whether they were wired common cathode or common anode), and that made it difficult to get the switching to behave - I don't totally understand the problems; I punted to driving the LEDs directly and accepting the brightness variations rather than putting the time and effort necessary into figuring out what was causing it. This is different from what you're doing in some rather significant ways, such as being pure 7-stroke instead of the more rounded digits you're looking for, but perhaps the experience can help. I would actually prefer to sand it (or sandblast, or whatever) to a matte finish instead of a smooth finish, for better light diffusion - the same reason they frost light bulbs, basically. /~\ The ASCII der Mouse \ / Ribbon Campaign X Against HTML mouse at rodents.montreal.qc.ca / \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B

Jules Richardson wrote: You haven't said what your viewing conditions are -- darkened room, direct sunlight, etc. There are different display technologies that can be exploited in each. I built a "display" with 4 ft digits. Very hard to read "up close" :> Is there a reason for the aesthetics? Or, "just because"? You can buy 14 segment displays (look like British flag). Your "segments" may well end up being LED "bars" -- which usually have 3 or 4 discrete lamps in them for more uniform intensity. There are displays that use no "static" power (i.e. only to change state). You can buy 5x7 LED arrays that are end stackable - though this is overkill as it allows for the creation of full graphics. Some vacuum flourescents come with "curved" corners and larger sizes (look at some of the early pinball machines). You also need to consider (in your power/intensity budget) how you drive the displays. I.e. if you multiplex them, intensity goes down (as does power -- though not proportionately). I assume you want to do this on a shoestring budget?

On 10/4/2006 at 12:58 PM Don wrote: Or you could try multiple lengths of EL wire. Very easy to work with, but since it's driven by high-voltage AC, the driving part could get to be expensive. OTOH, curves are a snap and and there's a wide choice of colors. Cheers, Chuck

On Wed, 4 Oct 2006, Jules Richardson wrote: How about a mock-nixie made from electro-luminescent wire? -- David Griffith dgriffi at cs.csubak.edu A: Because it fouls the order in which people normally read text. Q: Why is top-posting such a bad thing? A: Top-posting. Q: What is the most annoying thing in e-mail?

On Wed, 4 Oct 2006, David Griffith wrote: howzbout: making 7 segment displays from flourescent tubes? anywhere from about 6 inches to 16 feet per digit or howzbout: rig up a multibit input (parallel port?) on a PC, wire that to the existing 7 segment displays, and write trivial code to display an image of 7 segment displays on the PC's screen (or projector!) In ~1979, my partner in a computer business (Elcompco) negotiated a deal with Lawrence Hall of Science (the museum, not the labs) to make a bunch of power supplies for Kim-1's. I still stumble over them. But, they also needed a classroom display. We took a picture of a Kim-1, enlarged and mounted it, and wired up rows of plain LEDs replacing the digits in the enlarged picture, and put a real Kim-1 behind the poster. I wanted to make LARGE keypads on the poster, but my partner veto'd the idea. -- Grumpy Ol' Fred cisin at xenosoft.com

Fred Cisin wrote: I made a display out of ~2 ft tubes -- 4 ft digits. The problem is they are slow to start -- so not useful for a clock display counting seconds, etc. But, for a more "static" display, this wouldn't be an issue. However, they are high voltage devices which complicates their control (presumably, nowadays, you could use small inverters to control individual tubes instead of ballasts) Or, a mechanical actuator on an X-Y positioner to physically place placards in the desired locations on the "display grid". Personally, I would find the action of mechanically updating the display to be far more engaging than the "wonder" of a machine that plays soduku. Just like watching a chess *robot* is far more interesting than watching a chess *program*! An X-Y positioner approach also has the appeal of being reusable for other sorts of displays -- not just "decimal digits". I.e. you could display placards of chess pieces. or "checkers" (noughts and crosses?). Or any similar "board game".

On 10/5/06, Jules Richardson <julesrichardsonuk at yahoo.co.uk> wrote: I like the idea of the 9-layer plexiglass sandwich displays (you don't need '0' for your application), but I _have_ seen 3"-4"-tall 7-segment displays. We have a "mission control" type clock here (not GPS controlled, but does have an NTP client onboard). I haven't measured the LEDs personally, but the clock is over 3' long. A quick google search on '4" 7-segment display' gives this as the second hit... http://www.elexp.com/opt_0165.htm $8 each (q. 10) doesn't sound bad, but 81 of them is a fair quantity. Perhaps you could negotiate a slighty lower q. 100 price? This sounds like a fun project. Good luck with it. -ethan

On Wed, 4 Oct 2006, Jules Richardson wrote: I didn't see anyone else offer this suggestion, so maybe this idea would work... Use some plastic optical effects fiber with a single LED per numerical shape. Paint one end of the fiber, polish and heat shrink the LED to the other and bend the fiber into whatever shape you want. You could easily layer these. The only downside is that you'll probably have to buy a large spool of the fiber, but the plastic stuff is relatively cheap anyway. You won't want bare glass fiber that's normally used for cable splicing. It won't transmit much, if any, light though the sides. The plastic optical effects fiber will, as its made to do just that. If one strand wouldn't be bright or large enough, you could parallel and light more than one strand with the same LED too, particularly if you use a high intensity LED. (Yes, I'm still alive, just not really reading the list much these days. Still have my "collection", more of it is in storage right now do to space and time constraints, but I hope to change that this next year.) -Toth

Brent Hilpert wrote: You know, I'd forgotten about that. Our TR-48 uses such displays IIRC. It's certainly a possibility. As you say, there's some light bleed-through between digits, but that actually adds to the effect anyway! cheers Jules -- A. Because it destroys the natural flow of conversation. Q. What's wrong with top posting ?

der Mouse wrote: I don't think it's that big a deal either, but in this situation I'm trying to keep things as simple as possible. My gut feeling for Sudoku was that: - each display cell needs 5 bits of data; 4 to encode the digit (or blank) and one to dictate whether the user could change the contents of the cell during gameplay. - 'address' is basically formed from the XY position of the cursor (possibly via some form of lookup table so that 'holes' aren't needed in memory). - n starting grids exist in some form of ROM; one of these should be selected at random and copied to 'game memory' at game start. In other words, there's not really any kind of CPU or usual buses in any kind of classic sense. However, using a SBC of some form would certainly speed up development, so I've not ruled it out. I've not sat down and seriously thought about a CPU-less design yet, though - I'm more concerned at present with whether the display side of it can be done (cheaply). -- A. Because it destroys the natural flow of conversation. Q. What's wrong with top posting ?