Post by thread:[PIC] LED Matrix

Sorry everyone for starting a new article, but I'm new to this list and I'm just answering to some archived questions. Regarding to the matrixes, I've always noticed that most of those led signs are usually 5x7 characters. I have no idea why. even 44780-based LCDs When I was 13 or so, I made a program in QBasic which controls the parallel port and drives a 5x5 matrix. There are many approaches to this, but at the time i used this one: The columns are driven by the port bits, and the row is selected with a 4017 (this IC does a "rrf"-like operation on its outputs, that is, it selects 1 of 10 outputs, in sequence). with this approach you need 5 bits for the columns and 2 bits for the rows: one pulse strobes the IC and switches rows and the other pulse resets the output to 1 -- in case it losses sync, which it does quite often). But there is a more interesting approach, and I think this is the reason why characters are usually 5x7: If you use a 3-to-8 line decoder, such as the 74HC137, you use 3 bits (upper 3, for example) for selecting the row, and you have 5 bits left for your character!, so using a simple table lookup which looks like, say: 00000100 00101010 01010001 01110001 10011111 10110001 11010001 you can get a nice "A" character: notice the first 3 characters are just a binary count, and the other 5 are the data to be sent. first 3 bits are sent to the 74HC and the rest to the leds. if you sweep occidental-handwriting style (left-right, top-down) you can have more displays and room for whole words...). but i'm guessing, if the line is too long, you may need latches to avoid flickering. take a look at some of those ticker-style 1-line led signs, especially popular in the late 80s, and you will notice: characters are monospaced, 5x7 sweep is top to bottom (which may consist with a 3-to-8 decoder) when the text is sliding, it looks kind of slanted (i guess that happens because of the sweep) to reduce flickering, we can steal some ideas from the TV: many of you know about interlaced scanning. first you draw odd lines, and then even lines. this would help in reducing flicker and keep brightness. So what do you people think? is this a good idea? Hernán Freschi

> Regarding to the matrixes, I've always noticed that most of those > led > signs are usually 5x7 characters. I have no idea why. even > 44780-based > LCDs In the goode olde days when displays were very very very expensive to make and dot matrix printers were trying to take over from fixed character printers (such as daisy wheel, golf-ball, chain, band, etc) and also very very expensive, it was desirable to make characters using as few dots as possible. It was determined that 5 horizontal by 7 vertical was about as small a character matrix as you could get without doing excessive violence to the appearance. You can go smaller and be recognisable BUT it is immediately obvious that you are doing strange things. 5 x 7 usually looks really bad but the characters are instantly recognisable. You CAN produce largely recognisable characters using a 7 segment display! OK for the initiated to recognise but essentially unusable by Joe Average. RM

yeah but even today highway signs for example still have 5x7, and also many other kinds of displays, but we are way beyond that age. there must be some important advantage of 5x7 which makes it still the "standard". maybe because signs like that are designed to be readable (and those 5x7's look quite ok), and nobody cares about nicer characters? i remember the clock at some building in Buenos Aires (i think it was the Itaú Bank). It was huge, and had some beautiful numbers, all round and stuff. (obviously it was WAY beyond 5x7. ) www.itau.com.ar/institucional/img/empresa_edificioitau.GIF here's a drawing of the building (couldn't get a picture) but you can see that the clock is one story tall. it looks pretty neat, and i saw it from several blocks away. hjf On Tue, 18 Jan 2005 15:55:05 +1300, Russell McMahon <spam_OUTapptechTakeThisOuTparadise.net.nz> wrote: {Quote hidden}> In the goode olde days when displays were very very very expensive to > make and dot matrix printers were trying to take over from fixed > character printers (such as daisy wheel, golf-ball, chain, band, etc) > and also very very expensive, it was desirable to make characters > using as few dots as possible. It was determined that 5 horizontal by > 7 vertical was about as small a character matrix as you could get > without doing excessive violence to the appearance. You can go smaller > and be recognisable BUT it is immediately obvious that you are doing > strange things. 5 x 7 usually looks really bad but the characters are > instantly recognisable. > >

On Jan 17, 2005, at 9:36 PM, Peter Johansson wrote: >> i think 5x7 is used because I think 5x7 is used because long ago, when EPROMS weren't big enough to hold a bit-matrix character set, someone decided to burn a masked rom containing such a character set. 5x7 was what fit, and that ROM became the defacto industry standard for quite a long time (I think the first relatively public and available font bigger than 5x7 was the font used on the IBM Monochrome adaptor in the original IBM PC, although by that time terminals with larger "proprietary" fonts were pretty common.) There's probably also something in there about fitting a reasonable number of rows and columns on a near-standard video monitor. About 240 lines of video, huh? BillW

> >> i think 5x7 is used because > Back in 1973, at Georgia Tech (USA), a buddy and I wrote programs in Fortran and Algol that would print a banner on computer printout paper. We'd submit the programs on Hollerith card decks with the Data cards being the message we wanted printed. Each letter in the banner would print in a 5x7 matrix using the same letter for each dot. So, the letter A would print with letter "A"s, the B would print with the letter "B", etc. Like this: (view in fixed font) A BBBB A A B B A A B B AAAAA BBBB A A B B A A B B A A BBBB Anyway, we experimented with different matrix sizes and discovered that a 5x7 matrix seemed to fit all letters (upper case only) pretty well so that's what we used. Having an odd number of rows was good so you could have a line in the middle of the character for "A", "B", "E" etc. BTW, the computers we used were Univac 1108 and Burroughs B-5700. Carey Fisher

>Anyway, we experimented with different matrix sizes and discovered that a >5x7 matrix seemed >to fit all letters (upper case only) pretty well so that's what we used. It also allows you to do a reasonable representation of most lower case, and sort of allows descenders on g, j, p, q and y, enough to make them discernable for what they should be. To get properly decent looking characters you need a 9x7 matrix for proper descenders, and Motorola used to make a very nice ROM which had a 9x7 matrix mapped into a 9x16 space, giving excellent descenders, without requiring the full 9x16 space in the rom, it required a 9x7 plus 1 bit per char to signify it was shifted for the descender.

Alan B. Pearce wrote: >>Anyway, we experimented with different matrix sizes and discovered that a >>5x7 matrix seemed >>to fit all letters (upper case only) pretty well so that's what we used. >> > > It also allows you to do a reasonable representation of most lower case, and > sort of allows descenders on g, j, p, q and y, enough to make them > discernable for what they should be. > > To get properly decent looking characters you need a 9x7 matrix for proper > descenders, and Motorola used to make a very nice ROM which had a 9x7 matrix > mapped into a 9x16 space, giving excellent descenders, without requiring the > full 9x16 space in the rom, it required a 9x7 plus 1 bit per char to signify > it was shifted for the descender. > Came in on this one late. The following article http://www.swtpc.com/mholley/RadioElectronics/TV_Typewriter.htm was one of several Don Lancaster did in the early '70s when I was in High School. Got me interested (more interested) in electronics. Jim

Carey, On Tue, 18 Jan 2005 08:43:02 -0500, Carey Fisher - NCS wrote: > Back in 1973, at Georgia Tech (USA), a buddy and I wrote programs in Fortran > and Algol that > would print a banner on computer printout paper. We'd submit the programs > on Hollerith card decks > with the Data cards being the message we wanted printed. Each letter in the > banner would > print in a 5x7 matrix using the same letter for each dot. So, the letter A > would print with > letter "A"s, the B would print with the letter "B", etc. There's a coincidence - I wrote a program to do this in Fortran, in about 1975! I was an operator on an IBM 370/135 (at the lower end of medium sized mainframes at the time), and did this just for the fun of it during quiet night shifts. It was also useful for printing posters... I tried various formats and the 5x7 matrix "looked right" when printed on a line-printer, and gave 16 characters across, 6 lines per page. Cheers, Howard Winter St.Albans, England

what I want to know is how did they do those electronic signs in the 50's or 60's, with lightbulbs. was it some kind of punched card triggering a mechanical switch? On Wed, 19 Jan 2005 02:49:02 +0000 (GMT), Howard Winter <.....HDRWKILLspam@spam@h2org.demon.co.uk> wrote: {Quote hidden}> Carey, > > On Tue, 18 Jan 2005 08:43:02 -0500, Carey Fisher - NCS wrote: > > > Back in 1973, at Georgia Tech (USA), a buddy and I wrote programs in Fortran > > and Algol that > > would print a banner on computer printout paper. We'd submit the programs > > on Hollerith card decks > > with the Data cards being the message we wanted printed. Each letter in the > > banner would > > print in a 5x7 matrix using the same letter for each dot. So, the letter A > > would print with > > letter "A"s, the B would print with the letter "B", etc. > > There's a coincidence - I wrote a program to do this in Fortran, in about 1975! I was an operator on an IBM > 370/135 (at the lower end of medium sized mainframes at the time), and did this just for the fun of it during > quiet night shifts. It was also useful for printing posters... I tried various formats and the 5x7 matrix > "looked right" when printed on a line-printer, and gave 16 characters across, 6 lines per page. > > Cheers, > > > Howard Winter > St.Albans, England > > > -

> > with the Data cards being the message we wanted printed. Each letter in the > > banner would > > print in a 5x7 matrix using the same letter for each dot. So, the letter A > > would print with > > letter "A"s, the B would print with the letter "B", etc. > > There's a coincidence - I wrote a program to do this in Fortran, in about 1975! I was an operator on an IBM OK. I'll admit that I did this in BASIC sometime in that time-frame. However, we had a teletype (110 baud) rather than cards. -- D. Jay Newman ! Polititions and civilations come and jayKILLspamsprucegrove.com ! go but the engineers and machinists http://enerd.ws/robots/ ! make progress

On Jan 18, 2005, at 6:49 PM, Howard Winter wrote: >> Back in 1973, at Georgia Tech (USA), a buddy and I wrote >> programs in Fortran and Algol that > > There's a coincidence - I wrote a program to do this in Fortran, > in about 1975! Mine was in PL/C, on a univac 90/70, in 1977... We were using "virtual" card decks by then, but I laboriously entered the 5x7 character matrix on real punched cards, which I carefully saved in case I ever needed them again. I think I still have that deck. Somewhere. BillW

>There's a coincidence - I wrote a program to do this in Fortran, >in about 1975! I was an operator on an IBM 370/135 (at the lower >end of medium sized mainframes at the time), and did this just >for the fun of it during quiet night shifts. It was also useful >for printing posters... I tried various formats and the 5x7 matrix >"looked right" when printed on a line-printer, and gave 16 characters >across, 6 lines per page. I think pretty well everyone involved in that sort of field has done it at some stage. The one I remember doing was something similar, to print a legible name on the beginning of teletype punched tape, so you knew what program or data was on the tape.

Howard, You bring back memories. I operated the 370/135 (and before that various 360s going back to the humble Burroughs 263) and the last mainframe I operated was a 370/148 with 2 MB(!) of RAM, all of them the size of a refrigerator or two. I also printed those Happy Birthday and Merry Christmas signs, each letter taking up a page of the fan-fold (11x14?) paper and rotated 90°. Everything was Cobol and Fortran in those days. Cheers {Original Message removed}

damn! everyone in this list is that old?? ;) hjf On Wed, 19 Jan 2005 06:23:20 -0500, Tom Smith <.....peregrynKILLspam.....optonline.net> wrote: > Howard, > You bring back memories. I operated the 370/135 (and before that various > 360s going back to the humble Burroughs 263) and the last mainframe I > operated was a 370/148 with 2 MB(!) of RAM, all of them the size of a > refrigerator or two. I also printed those Happy Birthday and Merry Christmas > signs, each letter taking up a page of the fan-fold (11x14?) paper and > rotated 90°. Everything was Cobol and Fortran in those days. > > Cheers > > {Original Message removed}

> damn! everyone in this list is that old?? ;) No. But the first "real" computer I had anything to do with was an IBM BCD mini that took two passes to run the FORTRAN compiler because it didn't have enough memory to hold the whole compiler at one time. I was only a student - I only got to peer at it through the glass windows of the computer room. 1st microprocessor - SC/MP ! I still have a box of punch cards, just for memories sake. And an 8" floppy disk drive. And some core memory. And ... Time for uphill both ways in the snow stories again I see :-) RM

Essentially a ticker tape type mechanism with mercury. Pull the tape over a bed of mercury with the contacts on top, the holes punched in the tape would allow the mercury to touch the contact for as long as the hole was open. Mercury has a high surface tension, and lots of things float on it. -Adam Hernán Freschi wrote: {Quote hidden}>what I want to know is how did they do those electronic signs in the >50's or 60's, with lightbulbs. was it some kind of punched card >triggering a mechanical switch? > >On Wed, 19 Jan 2005 02:49:02 +0000 (GMT), Howard Winter ><HDRWspam_OUTh2org.demon.co.uk> wrote: > > >>Carey, >> >>On Tue, 18 Jan 2005 08:43:02 -0500, Carey Fisher - NCS wrote: >> >> >> >>>Back in 1973, at Georgia Tech (USA), a buddy and I wrote programs in Fortran >>>and Algol that >>>would print a banner on computer printout paper. We'd submit the programs >>>on Hollerith card decks >>>with the Data cards being the message we wanted printed. Each letter in the >>>banner would >>>print in a 5x7 matrix using the same letter for each dot. So, the letter A >>>would print with >>>letter "A"s, the B would print with the letter "B", etc. >>> >>> >>There's a coincidence - I wrote a program to do this in Fortran, in about 1975! I was an operator on an IBM >>370/135 (at the lower end of medium sized mainframes at the time), and did this just for the fun of it during >>quiet night shifts. It was also useful for printing posters... I tried various formats and the 5x7 matrix >>"looked right" when printed on a line-printer, and gave 16 characters across, 6 lines per page. >> >>Cheers, >> >> >>Howard Winter >>St.Albans, England >> >> >>

> > Russell McMahon wrote : > > > 1st microprocessor - SC/MP ! > I guess this is going to turn into a "My First Computer" thread! My first computer was an NCR Century 100 running Fortran and BASIC. You had to swap hard disk packs to change compilers. (They had handles on them.) You also had to boot the machine from the front panel using switches labelled in hex. My first micro was the RCA 1802. Nice little chip! Used it to decode RTTY and send ASCII to a dumb terminal program running on something I can't remember right now. Programmed it by entering hex code on keypad. Wrote ASM code on napkin and assembled it by hand. Sometimes just wrote it in hex. Carey

Carey Fisher <@spam@careyfisherKILLspamncsradio.com> wrote: > > Russell McMahon wrote : > > > 1st microprocessor - SC/MP ! > > I guess this is going to turn into a "My First Computer" thread! > > My first micro was the RCA 1802. Nice little chip! > Used it to decode RTTY and send ASCII to a dumb terminal > program running on something I can't remember > right now. Programmed it by entering hex code on > keypad. Wrote ASM code on napkin and assembled > it by hand. Sometimes just wrote it in hex. OK, hands up everyone who's done a real-time embedded app with an 8008! Back in 1976, I was in high school. The father of one of my friends had put together a Mark-8 (from the Radio Electronics article) and had an idea about using it to monitor delta-connected 3-phase power factor correction capacitors for blown fuses. The capacitors each had a series fuse, and they were packaged together in a unit with just the three corner leads coming out. Sometimes only one fuse would blow, and this was difficult to detect and isolate. The basic idea was to put current transformers on the three leads and use the micro to watch for the phase shifts that occured when one or more fuses blew. Anyway, he allowed me to do some of the programming. I was coding and assembling by hand in a notebook (using a fountain pen -- little personal quirk of mine at the time). Code was entered using the fine tradition of front-panel toggle switches. We basically got it working on the bench, but I don't know that it ever really became a product. -- Dave Tweed

On Wed, 2005-01-19 at 09:03 -0500, Carey Fisher - NCS wrote: > My first micro was the RCA 1802. Nice little chip! > Used it to decode RTTY and send ASCII to a dumb terminal > program running on something I can't remember > right now. Programmed it by entering hex code on > keypad. Wrote ASM code on napkin and assembled > it by hand. Sometimes just wrote it in hex. Hehe, the funny part is almost nobody would do it that way these days, yet almost every "hacker" in a movie does! Except Trinity, she used nmap and common stuff... :) TTYL ----------------------------- Herbert's PIC Stuff: http://repatch.dyndns.org:8383/pic_stuff/

> "OK, hands up everyone who's done a real-time embedded app with an > 8008!" No, I am too young for that. But when I was ~16y I discovered the Adam Osborne books on microprocessors and I found an ad in some magazine for the 8008. So I started designing my computer on paper, including level shifters and demultimplexers and whatnot. And of course I tried to write what you would now call a monitor program. But lacking any background in either programming or anything beyond simple electronics I (of course) got nowhere. But you've got to start somwehere :) Wouter van Ooijen -- ------------------------------------------- Van Ooijen Technische Informatica: http://www.voti.nl consultancy, development, PICmicro products docent Hogeschool van Utrecht: http://www.voti.nl/hvu

i feel so young. 1st computer, commodore 128 ( 5 years old or so). discovered PICs at age 12 from electronics magazines, but 1st programmed them at 17 when the local electronics store finally brought these to my town. well actually, built example circuits with precompiled software. I actually first wrote for PICs in college, for Computers Architecture class 4 years ago... hjf On Wed, 19 Jan 2005 17:49:02 +0100, Wouter van Ooijen <KILLspamwouterKILLspamvoti.nl> wrote: > > "OK, hands up everyone who's done a real-time embedded app with an > > 8008!" >

4004's!? I heard those were made from leather with iron pins... and the 1's and 0's were all hand carved from oak trees. That's why the old chips smelled different than todays chips when the magic smoke is released :) I'm not quite that old... I learned on the SC/MP chip. Ah the memories, I was 17, it was 1979 and I was a senior in high school (with hair!). I got to leave school at noon everyday to go to my electronics job at a small local company. Jack, the chief engineer had built a SC/MP based development system. I did programming on notebook paper. I had all the hex opcodes memorized. I especially liked counting backwards in hex to compute jumps. I found out years later about assemblers - Wow the jumps get computed automagically! SC/MPs are gone now. I moved on from that small company 20 years ago - sadly the company closed last year. Jack, the chief engineer I worked for is 78 now - we meet for lunch every few months. Quick somebody get me a kleenex. -Mark > > "OK, hands up everyone who's done a real-time embedded app with an > 8008!" > > Yep, and who's done PID on a 4004? > > (shit, I'm old....) > > John > > -

part 1 1165 bytes content-type:TEXT/PLAIN; charset=X-UNKNOWN; format=flowed (decoded quoted-printable) On Wed, 19 Jan 2005, [ISO-8859-1] Hern�n Freschi wrote: > what I want to know is how did they do those electronic signs in the > 50's or 60's, with lightbulbs. was it some kind of punched card > triggering a mechanical switch? It was a motor-driven wheel or drum with pins that made contacts on a row of switches for the more complex ones. But most simple signs used relays, and a few used mercury thyratrons. There are also schemes to build ring oscillators directly with high voltage gas discharge lights (the letters themselves). They only required capacitors and resistors besides the letters themselves but they were unstable (speed depends strongly on voltage and temperature). I wonder what used to drive the moving sign in Times Square, however. No motor & pins there, that was serious equipment. Most people do not realise how much power is involved in those lights. A 5x7 matrix done with 20W bulbs will draw 700 Watts per character displayed with all lights on and there are seldom any with less than 10 characters! And the inrush current of all those bulbs going on all together ? Peter part 2 35 bytes content-type:text/plain; charset="us-ascii" (decoded 7bit)

Ha!! You young whippersnappers! I started programming professionally in 1964 on an IBM 1401 - was THRILLED when I moved up to a 7090!!! Old Man Dennis Hernán Freschi wrote: {Quote hidden}>damn! everyone in this list is that old?? ;) > >hjf > > >On Wed, 19 Jan 2005 06:23:20 -0500, Tom Smith <RemoveMEperegrynTakeThisOuToptonline.net> wrote: > > >>Howard, >> You bring back memories. I operated the 370/135 (and before that various >>360s going back to the humble Burroughs 263) and the last mainframe I >>operated was a 370/148 with 2 MB(!) of RAM, all of them the size of a >>refrigerator or two. I also printed those Happy Birthday and Merry Christmas >>signs, each letter taking up a page of the fan-fold (11x14?) paper and >>rotated 90°. Everything was Cobol and Fortran in those days. >> >>Cheers >> >>{Original Message removed}

Dennis, You have me by a few years. I learned data processing in the USAF in 1966 picking up a smattering of Cobol, Fortran (before Fortran had numbers like IV, 77 or 90) and something called Report Program Generating (RPG). I began on IBM tab equipment in mid-'66 and continued through the end of '68. The precursor to the PC was the 407. I can still hear the relays clacking away :). In 1968 I took a machine language course on a 1401 then moved up to the B263 as big as a fridge with 4k of ferrite donut memory and "communicated" with the operator via an led matrix on the front panel. When a program went to normal completion it read hex 969. We did a monthly report, which had 150,000 cards input and they had to be sorted, run through the reader, resorted and run through again. It took four or five of us two or three days of 14-hour shifts to produce the reports. Some years later I picked up a credit card size pocket organizer with 4k of memory and carried it in my shirt pocket as a reminder of the progress in memory size. IBM produced the first PC in 1981 but they were outrageously expensive ($10k or so). I waited more than a decade before buying my first used PC and before long had a room full of them. But during that time I kept up with the field to some extent. I still have a McGraw-Hill book somewhere on Theory of Programming showing different schemes to implement assembly language. By the mid '90's the Internet had become affordable and Moore's law was helping bring down the cost of new PCs. So I picked up some Pascal and C and now I've come full circle playing with assembly language on PICs. Cheers {Original Message removed}

Russell, On Thu, 20 Jan 2005 01:41:07 +1300, Russell McMahon wrote: > I still have a box of punch cards, just for memories sake. I may have some somewhere - or they may have been thrown out in a purge at some time... > And an 8" floppy disk drive. Hey, I have a pair of these! Anyone want them? 100k single-sided as I remember. > And some core memory. Now I never did get my hands on any of that - still has some advantages over semiconductor memory, non-volatility being the main one. It always amused me that you could turn on a machine and carry on using it where it left off! > Time for uphill both ways in the snow stories again I see :-) Just getting warmed up... :-) Cheers, Howard Winter St.Albans, England

Howard, They must have been neons. Round with a flat surface and yellow or amber in color. You're right...diodes came along just a wee bit later. The card reader had white bulbs that shown through the holes in the cards to detectors. I don't know what kind of bulbs they were. Once when the engineer was away for a few days one of the bulbs blew so I grabbed his soldering iron and replaced the bulb and we were back in business. He gave me a shot when he checked out the repair...I was supposed to rotate the bulb to maximize its output by orientation. This is going waay back for me. :) Cheers {Original Message removed}

Most theater dimmers have the option to do preheating. When you have a $200 bulb thats only expected to last 40-50 hours, you preheat it. -Adam Jinx wrote: {Quote hidden}>>And the inrush current of all those bulbs going on all together ? >> >> > >Would they have had or been able to (economically) do zero- >crossing (assuming AC supply) and filament pre-heating ? Does >anybody bother with pre-heating or don't you need it if you could >do zero-crossing ? I'm presuming pre-heating would give better bulb >life-time with DC > > >

On Thu, 20 Jan 2005, Jinx wrote: >> And the inrush current of all those bulbs going on all together ? > > Would they have had or been able to (economically) do zero- > crossing (assuming AC supply) and filament pre-heating ? Does > anybody bother with pre-heating or don't you need it if you could > do zero-crossing ? I'm presuming pre-heating would give better bulb > life-time with DC No zero crossing there (we are talking early 1930's I think). The guys who had the contract to sell them the lightbulbs must have smiled so much they got creases on their faces from that alone. There is a site on the net documenting the Times Square sign and I read it a couple of years ago but I cannot find it now. I do not remember if it was very technical. Preheating always helps but it burns power. Not that they cared at the time. Maybe they used resistors for that (one per bulb). Peter

Hey, Tom! Hate to bust your bubble, but the old IBM computers used lamps that sounds remarkably similar to what you mention. They were incandescents!!!! You had a pushbutton (usually inside the front panel, which was hinged) you could push that would light all the bulbs so you could check for a burned out one. The front panel has what seemed like hundreds of these little bulbs!! Just an aside of days gone by!! Dennis Tom Smith wrote: {Quote hidden}>Howard, >They must have been neons. Round with a flat surface and yellow or amber in >color. You're right...diodes came along just a wee bit later. The card >reader had white bulbs that shown through the holes in the cards to >detectors. I don't know what kind of bulbs they were. Once when the engineer >was away for a few days one of the bulbs blew so I grabbed his soldering >iron and replaced the bulb and we were back in business. He gave me a shot >when he checked out the repair...I was supposed to rotate the bulb to >maximize its output by orientation. This is going waay back for me. :) >Cheers > >{Original Message removed}