10×10 LED Matrix

[KopfKopfKopfAffe] just finished a 3-year labor of love resulting in this 10 by 10 LED Matrix. This trumps the Shiftbrite table from earlier today by bringing an actual 100 LEDs to the display. These LEDs cost much less than the Shiftbrites, but since they don’t have their own on-board controller this project requires much more back end work. A total of 25 ATmega8 microcontrollers drive this display, in turn controlled via an RS-232 connection to a computer.

We love the Lemmings animations. It made us realize that this would be a great candidate to play Super Pixel Bros on. Check out the video after the break for snapshots of the hardware as well as videos of the matrix at play. You may also want to peruse the translated work log.


33 thoughts on “10×10 LED Matrix

  1. wow hackaday you have really outdone yourself here. A bunch of LEDs connected to a microcontroller, literally nobody has done that before and this is such a unique and highly skilled hack!

    man is really too bad it didn’t have an arduino connected to it instead

  2. Awesome project! Flashing lights impress everyone.

    I’m kind of new at this hardware thing.
    So there are 3 LEDs per cube. One Atmega8 for every 4 cubes. Using 12 dio lines on each MC.

    Is a separate MC dio line really necessary for every LED? With only 300 LEDs to address- couldn’t you use 1 MC and about 38 8 bit shift registers and save almost $100 on MCs?

    Perhaps they wanted every four cubes to manage its own logic? They wanted a refresh rate faster than the 50 Hz to fill the 300 bit shift registers on a 16 kHz clock?

  3. My other question for the community:

    Why the animosity against Arduinos? Why is it better to spend $15 etching and building your own board based on an Atmega MC vs buying an Arduino for $20 based on an Atmega MC?

    Is it just sour grapes that they lowered the barrier to entry or is there a rational reason not to use an Arduino?

  4. @Sobachatina : there’s nothing wrong with arduinos when they’re used cleverly. But most of the time, the project presented with an arduino are… quite stupid.
    Etching a PCB costs like $1 or $2, certainly not $15.

  5. Three years and 25 Atmegas…yes, that’s pretty much what I would expect doing it the hard way. It only took me one hour to build my matrix and another couple hours to measure the table and build the box to hold it.

    @Sobachatina: the simple shift register approach doesn’t work if you want brightness control of each pixel. Using microcontrollers as remote PWM is highly configurable and he got to choose his own update rate and communication protocols.

  6. @Sobachatina

    And to round it out properly.
    For anything posted here its comments will consist of;

    A bunch of people claiming you could buy something prebuilt for slightly less/more so why bother (they will later hate on anything using an *duino).

    A bunch of people claiming you could do it using 50x the components, a kite, a key and some bad weather (they will later hate on anything using technology designed after the vaccuum tube including *duinos).

    There’s a lot of people who don’t comment and just appreciate a project for its engineering (or sometimes its lack of engineering) whether or not it’s a “hack.”

  7. TJ: I think there are a lot of us like that. And I don’t think the comments are a good example of the majority of HaD readers(since many of us are quite), just the loud ones. Though I will admit I have seen a few projects that do seem overkill, but regardless people use what they have on hands. I have run into projects like that where for the most part it wouldn’t make sense to buy the parts to build it, but I happen to have them laying around.

    Besides if this doesn’t interest you (like some of the posts I saw recently) don’t read them.

  8. @j
    Yeah. I think it’s great when someone actually posts a constructive comment about how it could be done better, or how a different set of similar components could get the job done more easily.

    Unfortunately though, most of it is just noise from people who don’t actually have any projects.

    Although I might suggest that the copyright date in the footer be updated sometime this decade? d8)

  9. @Sobachatina

    There are definitely those insecure readers who feel it necessary to beat their chest to establish their moral and intellectual superiority by pointing out that any such project utilizing Arudinos could be better implemented on platform-x for reasons y and z.

    But honestly, it seems more likely that this “arduino animosity” is nothing more than Hackaday’s very own meme.

    (im in ur projekt, replaecan ur arudinoez)

  10. ah yes the TLC5947, that no one seems to have in stock

    Digikey will let you buy 1 for about 6 bucks, or 2000 @ 2.22$

    I guess if you need 2000 … but the mega8 is still more economical for this application

    and who cares, when you design your uber thing you can use whatever you feel like

  11. Thats a nice big matrix he’s got there :) I like shiny stuff, I’m impressed :)

    2 suggestions to make it better:
    1. ninjas (everything is better with ninjas)
    2. Lasers (everything is better with lasers, sharks optional)

  12. An Arduino? Really? How plug and play, I bet yoru a windows user, hell make that a Macbook user. I could have done this using anarray of RC Oscillators some salvaged NPN transistors, and LED’s taken from old model VCR’s found at the dump. What a waste!!!

  13. Apart from the arduino and 25 discussion: i like the result.
    Would be cool to have a wall filled with those blocks and then let slow random patterns wander over this wall – i call my invention “mood wall”(c)(tm)(r).
    This function could be implemented into the atmegas, without the need of a computer connected to the wall.
    But regardless of that: the lemmings rock!
    I will keep an eye on that project…

  14. One reason not to use Arduinos is because of the overhead in the code and less actual control over what is happening. Why not just learn C and the actual workings of an AVR uC. This also means you can more easily transfer to other platforms (such as ARM).
    Besides that, projects built with Arduinos always tend to end up being a big cable mess with breadboars (yuck, never ever use those besides for very early prototyping).

    On the totale number of ATMega8s used, I’m having my doubts. I once build a 16 RGB Led array with 8-bit color depth and 200Hz refresh rate. This all with one ATMega644 at 16MHz and some shift registers. I don’t know if you could make a 100 Led matrix this way, but you might get close. A nice trick I used for the PWM was Bitcode Manipulation (http://www.batsocks.co.uk/readme/art_bcm_1.htm).

    But in the end the easiest thing to do might be taking a LPC17xx ARM processor running at 100MHz. These only cost $5 and run like hell.

    All summed up, most important is that you actually finished your project (most people never get there), that the end result is nice and that you learned something along the way.

  15. IMHO If you are going to do this much effort you should just make a LED cube, those are much nicer than a boring (in my eyes) low-res monitor emulation.
    Still, it does bring some rightfully earned pride I guess to have created it when you make something like this.

  16. bro if there are 3 leds per cube and at total around 300 leds, 25 atmega8 controllers can be replaced by a single atmega32 and some logic multiplexing which could bring down the entire cost and time.it works as I had done it to my 8x8x8 led cube which worked alright.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.