Given the number of operations (> 24 floating point multiplies and adds) needed to get a useful reading out of the LSM303DLH compass module, it would make sense to put at least an atmega168 on the board, with a simple AVRISP-type programming header. That way a variety of I/O options could also be available.
Cheers, Jim Remington
We are thinking of doing something with a 32-bit processor for an AHRS-type product with a gyro also included. Do you specifically want just a compass, and is the main motivation that the price would be lower?
Yes, I like the accelerometer/magnetometer combination and have no need for the gyro and higher processing capabilities. I have an idea for a simple, low power hand-held device that determines only altitude and azimuth, and it bothers me that the main CPU has to perform so many operations just to get magnetic north. On the other hand, it bothers me that ST Microelectronics sells a “factory calibrated” chip that needs substantial individual calibration effort to actually be useful! Maybe another manufacturer will do the right thing.
There are already a number of AHRS boards available (in fact, I have the CHR-6dm board), so what might be the advantages of a new one from Pololu?
We do not have anything specific in mind yet, but if we did one, it would almost certainly be with a USB connection, allowing for easier configuration and firmware updates. I think we would push for a low-end option to get it as cheap as we can so that more people could play with them. Developing our own version is also a step toward integrating the capabilities into an affordable robot kit or at least making it a low-cost option.