PDA

View Full Version : Accelerometer help



Sunflower
03-05-2010, 03:11 AM
I bought a Memsic 2125 from Parallax and can not resolve accuracy better than +/- several degrees. Is this normal or am I doing something wrong?

kf4ixm
03-05-2010, 05:20 AM
How do you have it wired up? what does your code look like? i would really need to see a schematic of how you have it connected and attatch your code to your post to be of any help.

Sunflower
03-05-2010, 06:04 AM
The wires are simple MEMS Xout to Zin pin, 5V and GND. Battery power supply on MMC carrier board with capacitors. No vibration and stable room temperatures. Output is +/- 2 brads when level. Is that the expected performance of the MEMS?


Zin PIN 11 ' Tilt Z Accelerometer MEMSIC 2125

work4 = 0
FOR work.HIGHBYTE = 1 TO 64
PULSIN Zin, HiPulse, zRaw
zRaw = (zRaw */ Scale) '$0C0=Scale pulse to uSeconds 0.75 BS2p
zRaw = zRaw/10
work4 = work4 + zRaw
NEXT
zRaw = ((work4 / 64) - 500) 'Binary SIN -127 to 127
DEBUG SDEC zraw," "
NEXT

hover1
03-05-2010, 07:32 AM
An Accelerometer outputs in terms of g's, not degrees.

Jim


Sunflower said...
I bought a Memsic 2125 from Parallax and can not resolve accuracy better than +/- several degrees. Is this normal or am I doing something wrong?

kf4ixm
03-05-2010, 07:43 AM
i think you're using code for the compass module instead of the accelerometer code found here...

http://www.parallax.com/Portals/0/Downloads/src/prod/3rd/Memsic2125Demo.zip

Sunflower
03-05-2010, 08:17 AM
I left out the "8" factors for g force as I only need the SIN(angle) for relative tilt. From Memsic2125Demo.zip below. MEMS resolution is discussed in detail in small parts of a degree tilt, but are the actual measurements only good to a couple of degrees on each axis?


GOSUB Read_G_Force ' read X and Y

brads = (xmG / 8) ATN (ymG / 8) ' calculate angle
degrees = brads */ 360 ' convert to degrees

Read_G_Force:
PULSIN Xin, HiPulse, pulse ' read pulse output
pulse = pulse */ Scale ' convert to uSecs
xmG = ((pulse / 10) - 500) * 8 ' calc 1/1000 g
RETURN

hover1
03-05-2010, 08:48 AM
@Sunflower

I stand corrected. I have not followed the Basic Stamp software for this unit. I'm more familiar with the interface on the Propeller Chip.

The demo software does indeed report back degrees of tilt.

I will have to load up my BOE and try this out.

Jim

Sunflower
03-05-2010, 09:01 AM
Using ATN from two pins (X&Y) does smooth out the data to +/- 1 degree. I am running out of pins and would like the Z axis from one additional MEMS to one Stamp pin to measure dish tilt +/- 45 degrees, well within MEMS capability. But the Binary SIN rapidly fluctuates +/- 5 (a couple degrees) on just one axis. I guess I can accept that and still love the MEMS.