Compass belt status:

I figured out the problem I was having last week with the accelerometer: Turns out the I2C lines on the Nanocon board are brought out to two different headers. I hadn't noticed this, so I was trying to run I2C on two lines which I was already using for another purpose. I shuffled around the GPIO lines a bit to free those up, and I was able to talk to the chip. It's still behaving a little bit strangely in that it doesn't always initialize correctly. It's not likely to be I2C bus corruption because I don't see any flaky data coming back once the chip is up.

I tried hooking up an accelerometer to the microcontroller for the compass belt but no luck. It's supposed to communicate via I2C, but I'm having trouble getting I2C to work on the AVR board. I discovered from the schematic that SDA and SCL have on-board pull-ups activated by i/o lines B6 and B7, which I hadn't noticed, but even with them enabled something is pulling SDA down. There are some ESD protection diodes that I thought might have gone bad and been shorting to ground, but I pulled those off and still no luck.

See the recent notice about the site being back up. This is a total cop-out, but it's not looking like I'm going to get to an interesting stopping point on the "real" project, which was getting the microcontroller to talk to this accellerometer.

I was doing some cleaning and moved the box running this site. In the process, I managed to loosen the power cord, which very rapidly power-cycled the machine. I turned it off for a bit, started it back up, and things seemed okay. Some time during the morning, it crapped out. When I got home I did some diagnostics and determined the RAM had gone bad. My working theory is the rapid power cycling damaged the RAM somehow, and running with bad RAM caused a bunch of filesystem corruption.

"Of course, you have backups, right?"

That's, apparently, a project for another week.

I did get everything back up and running, so far as I can tell.

Working with MySQL was a bit of a pain. I found some quirks with mysqlcheck.

Crazy-late on this one, but here are some photos of the completed version 1.0 compass belt:

(Naj is modeling above.)

Compass belt update.

Ack. I'm late for the second week in a row.

Slow news week. Naj and I made a wire harness for the motor wires, but we're havint to redo because I didn't think far enough ahead. It brought all the wires to a single plug in front, which meant you had to shimmy into it. We started moving the circuit from a protoboard to soldered perfboard but only got about halfway.

[Above, Xander helps clear up a common misconception]

What's a compass belt?

Last week I had the circuit mocked up and the motor assemblies constructed. This week I got the motors sewed into some elastic, and give it a spin for the first time.

I added a brief explanation of what a compass belt actually is and does to the MOPTAW wiki: https://bandgap.rsnsoft.com/mediawiki/index.php/Compass_Belt

Here's the Wired article that inspired it: http://www.wired.com/wired/archive/15.04/esp.html

I've got some assembly photos of the compass belt up on my Flickr account:

As I mentioned in this update, I found the correct magnetic field model to be using to compute magnetic declination and magnetic inclination. Right now the program is busy crunching the values at quarter-degree latitude and longitude increments. It's been crunching away for about 30 hours, and is a bit over halfway done.