Gyro Question
ajward
Posts: 1,130
Hi Guys...
Haven't been around for a while because of... well, general laziness. :-(
Anyhow, I was wondering if the Gyro modules Parallax offers/did offer are sensitive enough to detect the rotation of the Earth. I can't go into the software I used because I don't remember where I put it on the hard drive. However I remember when using the LISY300 with the demo software, there was a constant drift I couldn't explain. When I used the currently offered module with some demo software, there was no drift... that I couldn't explain. :-)
Anyhow, I'd appreciate any wisdom on the subject!
Thanks,
Amanda
Haven't been around for a while because of... well, general laziness. :-(
Anyhow, I was wondering if the Gyro modules Parallax offers/did offer are sensitive enough to detect the rotation of the Earth. I can't go into the software I used because I don't remember where I put it on the hard drive. However I remember when using the LISY300 with the demo software, there was a constant drift I couldn't explain. When I used the currently offered module with some demo software, there was no drift... that I couldn't explain. :-)
Anyhow, I'd appreciate any wisdom on the subject!
Thanks,
Amanda
Comments
I don't think the L3G4200D is sensitive enough to detect the Earth rotation (at least not reliably).
The L3G4200D has 16-bits resolution. Since this resolution is both positive and negative this leaves 15-bits available to measure the magnitude of rotation.
15-bits gives a range of 0 to 32,767.
The most sensitive setting of the L3G4200D measure rotation up to 250 degrees per second. 250 degrees per second spread over 15-bits gives a sensitivity of 0.007629 degrees per second per bit.
The Earth rotates once every 23 hours, 56 minutes and 4.0916 seconds or 86164.0916 seconds. 360 degrees / 86164.0916 seconds gives us angular speed of 0.004178 degrees per second.
The rotation speed of the Earth is less than the single bit sensitivity of the gyro.
While the Earth's rotation isn't larger than a single bit, it is larger than half a bit so it's (theoretically) possible the L3G4200D can detect the Earth's rotation. I imagine there would likely be noise from various sources which would prevent one from detecting the Earth's rotation in a statistically significant way.
My guess is the software you're using is better at calibrating the gyro so you're not seeing the same sort of drift you previously noticed.
Orientation is everything ... basically to see the rotation of the earth, the axis needs to be oriented so that it is parallel to the Earths axis of rotation... as you go from parallel to perpendicular, you end up canceling out the rotational forces otherwise detected from Earths rotation.
Also, you need to structure your data capturing method in a way that you don't truncate anything or at least as little as possible. For example, if you assume random noise causes the LSB (Lest Significant BIT) of your data to be HIGH 50% of the time and LOW the remaining 50%, than anything outside of that 50/50 could be viewed as rotational bias that should be included as valid data. Many "code examples" I see truncate the LSB or try to average it which might explain why on some demo software you don't see the sensitive rotation of the Earth. The only time you should average the data is when retrieving the current value. The RAW data should never be averaged, only added continuously to an accumulator.
The long answer ... (see the link)
http://forums.parallax.com/discussion/127868/gyro-lisy300