Project E is a bit of a complicated project and so will require some significant R & D work. This is my electronics work area right now…. It’s absolutely in chaos, but I don’t have a lot of time so it will have to do and I clean it up as I move forward.

Here is my first little project that I put together to learn how to control a bank of LED’s with a microcontroller. I found a tutorial here on how to build this on Make’s weekend projects. It’s an animated cube of led lights. Pretty neat little project for getting started with controlling led lights with a microcontroller.

If your an At Home Imagineer and don’t know about Make magazine yet, you should seriously take a look and consider subscribing… or at least read their feed. There are some awesome project tutorials posted there.

So…. we have our first little clue to what Project E is… It is going to require controlling led lights with a microcontroller.

Here I am working on one of the prototype robot bases for Project E. I’m not finished yet… I still have to put the controller board on the robot base to control the servos below that drive the wheels. This robot is based on a tutorial here. This is just the R&D phase so I’m not sure if this is the robot platform I’m going to go with. I need 4 robot bases for Project E. I’m not sure if this protype is going to have a enough power.

Here is the underside of the bot. It has 2 drive wheels and one caster. I think the weight on the bot is going to require me to put two casters on each bot though. You can also see that I made my own servo mounts out of L-shaped aluminum stock. The body is made out of a sheet of HDPE plastic. I also had to take apart and modify the servos so that they would continually rotate and I don’t think I did a very good job at that. I think using better servos with more torque would be suitable for Project E, but we’ll just have to run some more test and see.

Ok… Project E has an amazing amount of part requirements! I was in serious need of some more room on my electronics workbench for working on project E. Boxes and boxes of parts have been coming in and my work area was getting extremely unmanageable. Since I did not have any walls behind my workbench to fasten any organizers to, I decided to do what any sane person would do and build a wall and bolt it to the workbench.

It’s not a real workbench unless you have carriage bolts in it.

So I found this inexpensive bin system at Home Depot that I decided to use. They have an assortment of different sized bins that hang on rails that you fasten to the wall. Here you can see the rails that I had to fasten to the wall. The plastic bins hang from these rails.

And finally you see here my wonderful new electronics workbench. For a quick weekend project done on a whim I think it turned out really well and did not cost me more then about $70 on plastic bins. Now maybe I can get some real work done!

I wrote a piece of software called Light Animator. I designed it to help me animate the hundreds of led lights that are going to be part of Project E. It should be capable of animating hundreds or even thousands of lights on a time line. I’m going to release it in the next couple weeks under the GPL so anyone can download it and extend it. It is extremely limited at what it can do currently, but over the course of the next few months as I complete Project-E, I will be enhancing it with all sorts of features. This Christmas I even plan to use the same software to animate my Christmas lights!

I just put together a small circuit that allows me to control an array of led lights with a shift register chip. I found out how to do that in this tutorial. It’s working out awesome. My next step is to expand it to 16 led lights, and then create a test animation with Light Animator and load it on the arduino.

More parts for Project E and more clues as to what I am building! I have a Daisy MP3 player kit from Make, 2 20 watt amplifier kits from Ramsey (although I think I might upgrade to the 40 watt kit), 2 different types of fm receiver chips (I have no idea which one I am going to use yet), and finally a gi-normous amount of copper wire (little clue, I’m probably going to need a lot more).

This is a prototype copper grid for me to test sculpting with copper and to act as a test bed for 30 leds. Also here is a prototype of a led assembly for the project. I’m going to have to make a couple thousand of these. Oh Boy!

I have decide to use the arduino board to power project E. In fact I think I may end up using 9 arduino boards in Project-E. I estimate that Project-E in total will have 13 microcontrollers. Here you can see I have a very simple circuit that I did based on this tutorial. It’s from the Make weekend podcast. These guys are awsome; however, Bre is no longer doing them so we’ll have to see what Make does in the future.

The arduinos come in many different shapes and sizes. You can even just grab the schematics online and build one yourself. Next to my main test board you can see a smaller version of the arduino which I am planning on using in project E.

Here is my start on the first bot! The start of my very first wire sculpture. Hope I don’t screw it up!

Wow… the frame of my first bot I think is coming together. I still have to finish up the roof, create the front and smoke stack. But so far, I’m pretty happy with it.

Here I have a sort of custom arduino that I built. At least I used the schematics for an arduino to get a nice simple avr platform up and running. I am just coding in straight C though using avr-libc though as opposed to using any of the arduino libraries or utilities. Here you can see the stk500 development board next to my board. I am using it along with avr-dude to program my board.

I added a a serial output to my development board to help with debugging the chip. The tutorials at sparkfun were really helpful!

Ok…. This was a big deal for me. For some reason it took me a while to get a TWI (I2c) interface up and running properly and it held me up from doing some blog posts. But it’s working now! Here I have two simple avr platforms up and running. The one on the left is the master and the one on the right is the slave. To control 300-500 LED lights I decided that it was a job for more then one microcontroller. So I will have a master LED controller and several slave LED controllers. The master will push the animation frames down the bus and each of the slave controllers will control a bank of shift registers connected to 64 LEDS.

Here is a close up of the Master Controller. Not really much different circuitry then before. I will be posting the code up here eventually. That is where most of the fun is! Don’t be fooled by the connector up to the top left. It’s not Ethernet. That’s actually a serial connection to my computer for debugging. I use Ethernet patch cables instead of serial cables often now because I can make my own cables then. Just slap a couple of custom RJ45-DSUB-9 adapters onto an ether cable and you have a serial cable. I picked that up from Make Magazine (Issue 01). Pretty Handy.

It’s great to be a geek!

Here is an update of my light control circuitry. I have an Arduino mini at the bottom right acting as my primary controller. I’m not using the arduino environment to program it though. I’m just using it because it’s a nice small atmega platform that has the crystal already on the board. I’m just coding in straight C and using avrdude to load the chip (which really isn’t all that different). I have it currently wired to three other amega48 microcontrollers which are programmed to act as slaves on an TWI serial line. I then have two of those wired up to MAX232 chips to stream the serial data that the slaves receive to my desktop computer so that I can verify that they are receiving the correct commands. I can’t claim that I entirely know what I’m doing… but I sure am having fun!

Here is another shot of my bench setup.

This is the gumstix that I bought to be the brains of the robotic swarm. In this picture I have a gumstix on top, a tweener board in the middle, and a roboaudiostix on the bottom. I’ve been looking for a reason to buy one of these for a while.

Thank goodness for companies like Adafruit Industries! One of the pain in the butt parts of project-E was going to be that I wanted to use Xbee connect all the robots together. Xbee is awesome, but it has two challenges to integrating it into common circuits. One is that the spacing of the pins is not .100 like bread boards and prototyping boards are so you are going to need some sort of pcb to make it fit and also that the Xbee chips are 3.3v and are not 5v tolerant. Well Adafruit makes this wonderful kit that does all the voltage level shifting for you so you can use a Xbee chip with a 5v circuit and with a nice breakout for breadboarding!?

Here is my completed kit. It was pretty easy to put together.?

Here is my kit with the Xbee chip in place.

To get started with Xbee I am flashing all the firmware on the chips and making sure my newly built kits from Adafruit Industries are working correctly. In the picture I am using an arduino board without the chip as a USB to TTL-232 converter to talk to the Xbee Chip. I later ordered a couple of cables from Adafruit Industries that do this for you. They come in real handy!

If you couldn’t tell already why I had the helping hands in the previous picture… here is a closer look. It’s holding one of the wires connected to the Xbee directly to a point on the Arduino board that wasn’t exposed on one of the connectors. Hence why I found it a little nicer to order the USB-TTL232 cables. It worked just fine… but those USB-TTL232 Cables come in handy for so many things. I often use to debug my microcontrollers by hooking a max232 chip up to the microcontroller. I would then use a USB to RS232 cable to hook my computer to the MAX232 chip. Now I just plug the USB-TTL232 into my computer and hook it up directly to my chip. Works great.

This is my new blackberry storm. For a while I was going to build a hardware based remote control for project-E… but as the design became a little clearer to me; the remote control became a little too complicated for a hardware based system. I thought the touch based interface on the storm would make a nice remote control. The entire remote can then be implemented in software which makes a lot of neat possibilities easier to do.

So far the storm has been pretty fun to develop for. It’s completely java based and very well documented. I know it’s not the geekiest phone out there right now but it was the best one for my purposes. My first choice would have been the G1 but unfortunately it’s only on a network that doesn’t have anywhere near the coverage that Verizon has. I don’t always like all of Verizon’s policies but it’s hard to beat their network coverage. I absolutely love the design of Apple’s phone, but they don’t seem to be very friendly towards developers and I find it difficult to understand how so many developers get excited over a phone that its company is so restrictive about.

Here is a look at my current work. Most of my work has been concentrated on gumstix development for the last couple weeks. I have the custom code written and communication up and running between the gumstix and the roboaudiostix via i2c. I’ll start posting my code soon… I just have to figure how I want to start posting the code to the blog.

I also have bluetooth communication working between my blackberry storm and my gumstix although the code has to be finished on the gumstix. The custom code is working on the blackberry but I simply used minicom running on the gumstix to verify my data coming across so far. My next step will to write the custom c code that runs on the gumstix to handle the incoming messages from the storm.

Here is a closeup of my new gumstix board. I now have a verdex pro XM4-bt to use as the main computer for the parade. I changed to a new board for two reasons. One is that I needed enough memory to store the wav file for the parade and the new verdex pro boards have an on-board SD card slot on them, and the other is the control system is now a blackberry storm. The best way way to control project-E with the blackberry is via bluetooth so it just made sense to have a board with built in bluetooth.