8×8

hello and welcome to the 8×8 development blog.

i’ve realized, for some time, that the blog has been quite difficult to understand. obvious questions like “what is 8×8?” has not been answered, and as a first time visitor you had to dig deep to find some sense. that’s the nature of blogs i guess; you write them backwards. so the most recent, and quite possibly the most geeky stuff, becomes the news on top.

well, this entry will try and help you out. i can’t offer any quick answers to the question above, but if you’re lost, look at the following pages/posts:

about 8×8
8×8: a short description
3d rendering of 8×8
why the huge touch screen interface?

i have been asked whether or not i support the use of my ideas by other people; if it was OK to copy this project, in part or in whole, or to take parts and ideas and modify them. yes, it is. in fact, i would be very happy if you did.

8×8 and the development blog is released under an attribution-non commercial-share a like creative commons license.

making 8×8 and making this blog has been my first personal experience in open development, and i have been thrilled by the positive response and creative feedback this development form has given. the internet never ceases to amaze me!

so keep postin’ them posts! i really appreciate the input/support, and i love the fact that my work with 8×8 can help, inspire, and eventually grow, through the works of others.

check out this link.

it’s really neat - seems like a real time translator for video -> 8×8 LED matrix. thanks to mik mo for the link.

the second last day of dedicated school work with 8×8 has come, and i must admit that i am nowhere near the 8×8 design, i originally planned. again theory and real world implementation proves to be two very different concepts.

there has been a lot of tech problems with implementing the max7219 to 12V incadescent lamp setup, which has taken most of my time during the three week construction period.

my main priority is now to complete the physical frame of the display and the mosfet/incadescent lamp setup. the whole sensory/input part of the project is left to further development. as well as a more powerful powersupply.

i will continue to develop 8×8, but the pace will be much slower. after tomorrow i will no longer have the privilege of working on 8×8 as a school project, and during next semester i will be writing my bachelor, so dedicated time for 8×8 will be limited.

i will, of course, still be posting on this blog with development news, when they arrive along the way.

its time to bend some metal! i’ve worked out a neat little lamp holder made from 3 electric couplers and simple steel wire.

now i just need to bend 128 pieces of wire, connect them with 64 diodes, strip 128 wires, fit it all in a nice 8×8 grid, solder the electronics and build the metal frame works. in 2 days, if possible. aiaiai.

to test whether the heat convection will be a problem or not, i have made a test setup of a vertical row of lamps. by only allowing a bit of air to come in at the bottom and front, limiting the outflow at the top as well, i can simulate the “chimney effect” that my construction will have and measure if the temperature at the top becomes critical.

since i am unsure of exactly how much energy each light bulb will end up receiving/consuming (see several posts on power supply issues) i have settled for an estimate of 4W each. this gives an acceptable brightness level, and it is more realistic than applying a full 10W to each of the bulbs.

measuring the temperature development just above the top light bulb renders the heat development graph below

at first i measured the heat development with a 5 cm gap in the front (simulating a perforated back plate construction). once the temperature had stabilized, i reduced the front gap to simulate minimal inflow from the back (solid back plate construction). as we can see, there’s not a big difference between a sideways ventilated system and a chimney. both systems stay well below the critical 60 degrees (thats when you start to burn yourself on your equipment!)

this is extremely nice. it means that an acceptably lit 8×8 construction can hang on a wall without causing a fire hazard!

things have been quiet on the blog for almost a week, but i have been busy sorting out my own mess:

relatively early in the design process i decided on using the maxim7219 LED controller. it seemed like the right/easy thing to do, since complete libraries for controlling it had already been written for arduino. but something peculiar happens when you buy a chip and download some code to take care of your dirtywork: you stop thinking about what you are actually doing! or at least i did.

i have done my light, heat and power calculations from the assumption that ALL 64 lamps will be on constantly (worst case scenario) - but this is impossible.

i am making a grid of lamps. i only have 16 wires. 8 posivite, 8 negative. now how can this connect to 64 lamps each with their own ON/OFF setting? only by scanning the grid. the maxim7219 goes through each row one at the time supplying positive voltage which flow is determined by the state of the 8 column pins at that exact moment (if the pin is low, power flows through the LED, if it is high, no current flows). this happens at 800Hz making it impossible for the human eye to register any kind of flicker.

but it happens.

which simply means: one light bulb can only be one 1/8 of the time, giving me a total light effect of only 80W in stead of glorious 640W. shish.

good thing is: i don’t need to worry about overheating the lamp anymore!

second bonus: i only need 1 PC power supply for the whole setup!

80W in stead of 640W was greatly disappointing for me. i have really been pulling my hair on this problem. but then i got a cool tip: nature of the incadescent lamp might save me. i have designed a MOSFET setup to boost my LED driver voltage to the required 12V/10W, but in theory it can go much higher than that. the 12V lamp power supply and the 5V microcontroller power supply are only sharing ground, so i can boost the power on the incadescent lamps as high as 24V (or perhaps even more). the incadescent lamp can take this load in peaks, and evens out the load to an average current and light/heat flow.

this kind of power is not going to come from a PC power supply, but it means that i can build 8×8 in a working, although not very bright version now, and optimize the power supply later on.

click here for a simplified schematic of my max7219 -> 12V (or MORE) incadescent lamp setup.

i have decided to scale down 8×8 a bit. main dimensions will now be 80×80cm. i am worried that the theremin pitch sensor will have too limited sensitivity range to cover to area from the two upper corners. so i’ve decided to use 4 pitch sensors on a smaller area in stead: one in each side bar, in stead of two in the top corners. this will allow my construction to be much more stable, it will be simpler and more beautiful and also it will add the functionality of two extra theremin pitch sensors.

and the numbers are great: 8×8: 80×80, 1 pixel: 10×10.

also, i am now programming 8×8 using processing. squeak is great, but i am more at home with processing, and time is very limited.

this is so much fun!!

processing translates what ever is in the processing window to an HEX encoded string sent over the serial port. arduino then translates the string from processing to the maxim7219 LED controller.

i’ve set up my processing to respond to keyboard input (and to create the visuals) and also to read a potentiometer value sent back over the serial port from arduino.

click here for the processing code and here for the arduino code
sound visualization in processing is built on the ess library by krister olsson.

like lo-res animations? like them big? well, here is your screen!

do you have a neat little pattern or a nice .GIF animation that just needs more attention?? do you feel like making a feature for 8×8 lo-res square-o-vision? send it to me: uglenbatman [at] gmail.com and i will bring it to the big screen!

the video above is, ofcourse, showing the LED 8×8 prototype. i still need to build the big one. but you can get the idea.

animation tech specs: 8×8 pixels. b/w (this might change - i’m working on 16 grey scales) preferred format: .GIF/QT

my previous heat calculations might be wrong… i have based my estimates on an isolated pixel. but 8×8 used pixels stacked on top of each other. the warm air is transported up and on to the light bulb above, which then heats the warm air further, and all of this is repeated 8 times.

this might cause the top lamp to overheat!

yay! the PC power supply works!

please note: the supply doesn’t go on, until the green wire is connected to ground. i didn’t know, so this had me fooled for some time.

the front end material should be transparent and light. it is the material which the user will interact with, so it it is essentiel, that it has a have nice feel to it.

initially i’ve thought of white vinyl, which offers a very cool and smooth surface. but sniffing around the workshop in the department of structural engineering, DTU, i found a type of plastic which in danish is called “kanalplast” (directly translated: “canal plastic”) as the name suggest, kanalplast offers a the double layered construction with lots of air canals inside it. the layered and linear build, refracts the light in a nice “pixelated” which i feel, suits the low-res aesthetics very well.

kanalplast is commonly used in covered patios, greenhouse and the likes, so you can buy it in your local building market.

the initial output i got from the theremin pitch sensor was quite noisy and caused many reading errors on the microcontroller. using a schmitt trigger i can make the signal quite square and “digital-read-friendly”.

ah, PC power supplies: cheap or free - castaway PCs are everywhere. and quite powerful. since my lamps are grouped in rows of 8, i might be able to used 1 PC power supply per row. perhaps even 2 rows per power supply. the result will still be a very large power supply, but at least it is CHEAP.

also i found this interesting web page om modding PC power supplies. and this one on making a more general desktop power supply, cheaply.

an alternative cold also be to use a car battery. they typically give you 12V 60A which is perfect, but they would only last for a limited amount of time. perhaps an hour or so. good lenght for shows in strange, non-powered places, but not acceptable for permanent installation/use.

lighting up 64 10W light bulbs initially seemed like a unproblematic thing to me when designing this 8×8. i mean, stage and film lamps easily go to a 1000W.

but they are 220V. since my lamps are 12V the current is 10W/12V=0,8333A per lamp. 0,8333Ax64 is 53,33 ampere! thats a lot. (you can compare it to a stage lamp: 1000W/220V=4,56A) buying a power supply which can supply this amount of current is expensive!! and they’re HUGE! pfff… this really sucks.

the final period of this project has come. the next 3 weeks, up until january 20^th, are dedicated to constructing and documenting 8×8. so i am preparing myself to spend a lot of time in here!

to feel more at home, i’ve brought my AC/DC coffee mug. i think it looks really cool, standing next to a DC power supply! shurely it will bring a good rock vibe to the project.

i’ve constructed a simple theremin pitch sensor from schematics found on theremin world. you can see how the frequency changes on the oscilloscope when my hand approaches the antenna/wire.

as a small prototype of the 8×8 display i’ve constructed a 8×8 LED matrix. the arduino board has limited outputs, so it is controlling the LEDs using a maxim7219 LED controller (which only needs 3 outputs: clock, data and load). the maxim7219 then decodes the signals from arduino and takes care of the multiplexing/demultiplexing of the 64 LEDs.

the pattern displayed is controlled by my computer, sending the data to the arduino microcontroller via USB/serial communication.

to control the light intensity of the separate lamps, i am using PWM. when using PWM we are actually not fading the lamps but turning them on and off so quickly so we perceive the average value as the light intensity.

below is video of a simple PWM setup, “fading” two LEDs