Pages: [1]   Go Down
Author Topic: Acceleration Measurement  (Read 298 times)
0 Members and 1 Guest are viewing this topic.
Offline Offline
Newbie
*
Karma: 0
Posts: 38
View Profile
 Bigger Bigger  Smaller Smaller  Reset Reset

Hi,

I would like to try to measure horsepower from an accelerometer. If I collect acceleration data WRT time and perform some integrations, I will be able to calculate velocity and position. From there with some drive train calculations it should be pretty easy to calculate torque and horsepower. I need one axis and a 2 g accelerometer. I would like to collect as much and accurate data as possible to do my integrations. My plan is to use a memsic 2125 accelerometer. It is low cost, meets my g requirement, and axis requirement. From my reading there are 100 hz and 400 hz models? The accelerometer puts out a pulse signal, so I wouldn't even have to perform an ADC.

What is the best way to collect this data? Should I look at getting a SD card shield? How many samples per second can the arduino uno collect? 16 msps?
Logged

Offline Offline
Edison Member
*
Karma: 58
Posts: 2078
View Profile
 Bigger Bigger  Smaller Smaller  Reset Reset

Did you really say: 16 Mega samples per second ?
Or: 16 milli seconds per sample ?

By the way, that sensor is an expensive freaky weird old and outdated sensor. Why would you use a frequency output, when the Arduino has a I2C bus, a SPI bus and analog inputs.
There is a simple sketch for it: http://arduino.cc/en/Tutorial/Memsic2125
(But I still don't like that sensor)

Don't set high hopes on calculating the position. The MEMS sensors are noisy and drift.
http://en.wikipedia.org/wiki/Microelectromechanical_systems
I did see a few projects that calculate the position. Perhaps one was on www.hackaday.com
« Last Edit: September 16, 2013, 02:35:32 pm by Erdin » Logged

Dubuque, Iowa, USA
Offline Offline
Edison Member
*
Karma: 45
Posts: 2472
View Profile
 Bigger Bigger  Smaller Smaller  Reset Reset

There was a post on hackaday where an accelerometer was used to track the position of a person's shoe as they walked. The only reason this worked is because the user could count on their foot being motionless every step and at a known position (on the ground!).

But no, you can't hope to put an accelerometer in a car, for example, and have any sort of sane result beyond the first two seconds.

Practical rate of recording samples would be around 1000-10000 samples/second. Commodity MEMS accelerometers are only capable of outputting data at ~1000 samples/second.
Logged

Pages: [1]   Go Up
Jump to: