Complementary Filter Success!

Back to tackling this old problem of how to get clean sensor data. Previously, my tests were using simple low pass filters, which, while simple, produced major lag. Recently got hit by the "OMG I MUST CODE A KALMAN FILTER FOR IMU" bug, and started doing classes with Khan Academy, Youtube etc. in physics and statistics. I finally get what a co-variant matrix is... sort of. But a Kalman filter is really beyond my ken at the moment, and decided, hey, that can take longer to study. It's not like world peace will magically appear if I learn how to code one. The suggested alternative, is the complementary filter.

While I do have my electronics kit and kaboodle with me in Vancouver, one BIG component is not working - the tip of my soldering iron appears to have died >.> And the replacement is apparently in the mail, for over a week now. So I can't really do much more than drool appreciatively at the new 9DOF sensor board on my table. In the mean time, apart from doing my studies above, I found some sample data (google for "sample data freescale") online and started messing around with them in Processing.

After going through many complementary filter links, I found one that finally "clicked" for me. While I mostly understand how to implement it (especially after the physics classes), after reading so many forums and sites, I'm pretty sure I'm missing some of the finer key points.

In other news, I've purchased a Teensy 3.1, which I hope will arrive sometime soon. This is my 2nd time playing with arm chips, can't remember why I sold off my Arduino Due.