accelerometer / gyroscope

TheGideon

hey people

I'm working on a project where I need to check if a beacon is correctly in place between two railway tracks.
thats why we need to monitor three axes and check their angles.

We bought with the Basic Stamp a Memsic 2125 Dual-axis Accelerometer.
(item code: 28017)
http://www.parallax.com/Store/Sensors/AccelerationTilt/tabid/172/CategoryID/47/List/0/SortField/0/Level/a/ProductID/93/Default.aspx

We knew it only would monitor two axes but we took it to gain some experience with the Basic Stamp module.

So a month later we decided to buy the MMA7455 3-Axis Accelerometer Module.
(item code: 28526)
http://www.parallax.com/Store/Sensors/AccelerationTilt/tabid/172/CategoryID/47/List/0/SortField/0/Level/a/ProductID/585/Default.aspx

but it is not precise enough for our project. The max angle on the x-axis is about 2°. Our test results from the sensor where varying way to much.

I would like to know what sensor we can use best.
The results from the Dual-axis accelerometer where good, so we maybe can take that one and place a compass next to it.
Or is it better to try an other more precise accelerometer or maybe a gyroscope.

is there someone with an advice about these sensors?
I would like to hear them.

Greets
Gideon

Franklin

I'm working on a project where I need to check if a beacon is correctly in place between two railway tracks.
thats why we need to monitor three axes and check their angles.

What kind of enclosure are you using and what degree of precision do you need? How do you attach the device to your installation?

Duane Degn

Theoretically the MMA7455 should give you about 0.9 degrees of resolution (I now realize I did this math incorrectly the other day when I stated it would give 0.45 degree resolution). I think for your purposes, you'd want an accelerometer that could give better resolution (as you just said).

This accelerometer can be set to use all 14-bits of resolution at +/- 1G (14-bits over 2G) so if I did my math right this time (I didn't last time I tried this) the accelerometer should provide a resolution of 0.007 degrees. Assuming you could keep your instrument still enough. The math (I think) is arcsine(1 / 8191). 8191 is the max value of 13 bits (you loss one bit because of the 2G range).

I hope someone correct me if I'm wrong.

I didn't look through all of SparkFun accelerometers, I just did the math on the first one that looked like it would work.

TheGideon

Thanks for your reply!

we're probably going to try this one.
I sure can use this information.

we will keep thinking forward!