Accelerometer and gyro for datalogger, what to buy?

A buddy of mine introduced me to the arduino concept the other day, and I think this sounded like a fun base for a project.

I'm looking to make a datalogger for my motorbike, sort of formula 1 / moto gp style logger that logs the force applied to the dampers(fork and shock) during for example a track run.
So in that case I need a sort of gyro, to get the angle of the bike, and a accelerometer to get the force applied.
I have ordered a ATMEGA2560 from ebay, and need some help in regards to what other components to buy, what will fit/ be compatible etc..

I need a gps to log the speed and distance traveled, accelerometer and gyro to log the suspension data, and a sd card module to store the data.

Could anyone give me any advice on what components to buy, and not?

I like this one: Pololu - MinIMU-9 Gyro, Accelerometer, and Compass (L3G4200D and LSM303DLM Carrier)

I am still learning how to use it.

If you're quantifying the effectiveness of the dampeners then you want to measure the jerk on the motorcycle. By "jerk" I'm not referring to the rider, but rather changes in acceleration ;). Google will provide lots of leads in how to measure jerk.

If you were going to attempt to run this test by simply looking at acceleration you would not get useful data. If you're lazily sitting in a chair feeling the consistent 9.8m/s acceleration due to gravity you don't normally consider this a bumpy ride. If you're flying in an airplane in turbulent weather you're still being accelerated by the pull of gravity, but you describe it as a bumpy ride because you're feeling change in that acceleration.

Food for thought, after riding what Chagrin said, you might need accelerometers for the axles and one for the bike. In this way you can see what forces the axles are seeing verses the frame of the bike therefore, the effect of the dampers.

I'm a little new on here, but I like the above. With the 2560 you have 6 interrupts, which means you can log both axis of each of the three accelerometers pretty close to simultaneous. The only downside is accelerometers are kind of expensive. I bought a Memsic 2125 from Radioshack which cost me $38. Ebay doesn't seem to be much cheaper when you consider having it now vs. waiting a week or two. But, if you did get them all, you could log the data on your SD card and compare the values next to each other rather easily on the computer. My guess is you would want to see how much the body of the bike's pitch is affected by the force that changes the readings on the axles of the bike. Good idea though, I bet it could even work on a car, with a device on each wheel and one on the body.

If you're going to use a gyroscope in this project to measure the back and forth tilt of the bike with respect to the earth then you'll have to deal with the natural drift of the gyro (all gyros are imperfect in their output and alone, after any period of rotational changes, cannot tell you their orientation with respect to the earth). An accelerometer would be used to provide a point of reference with respect to the earth because it detects the consistent normal force of gravity. That then takes you into dealing with Kalman filtering, but on a motorcycle with its associated bumpy ride screwing up the accelerometer that will inevitably complicate the issue.

If you want to use the gyroscope to record the rate of change in pitch that would be much simpler, but again I'm not sure how useful that information would be. If you were talking about a dirt bike and how it handles over "whoops" it might be useful -- but not for a street bike.

Chagrin: Thanks for the advice so far.. The bike is a supermoto/dual sport, but the logger is going to be used only on tracks.

I was thinking of using the gyro for measuring the angle the bike is tilted during cornering, and accelerometers to get the forces applied, and then using that angle and forces to calculate the forces through the suspension. Putting accelerometers at the axles would probably work best, but it will be a bit tricky and expensive to do. I was thinking that as long as i could get the forces applied to the suspension in corners, then the graph I make afterwards will tell me how the spring rate etc. is working based on the "smoothness" of the graph through the corners and braking. The accelerometer/gyro would be placed approx at the center of mass of the bike.
e.g. a pointy curve would indicate that the suspension is a bit stiff/hard
a smooth curve would indicate that the suspension is good.
a flat/pointy curve would indicate that the suspension is too soft.

Am I making sense here? :cold_sweat:

Again, this is just a for-fun project, I dont really know if this will help me set up any suspension better, and I don't care, just want to give it a try. :slight_smile:

The memsic 2125 that neato3000 suggested seemed to be something that could do all that I want?