PSN-L Email List Message

Subject: Re: ADXL202 Sensor
From: "David Saum" DSaum@............
Date: Sat, 26 May 2001 18:55:48 -0400

> From:    Casey Crane 
> Date:    Thu, 24 May 2001 17:35:29 -0700
>How do you define "Counts of
> noise" ? I have an O'scope But how do I measure this random or
> psuedo-random stuff without a spectrum anylizer ? 

You don't need a scope, just the old eyeballs.  And you need to use the
gravitational field of the earth to calibrate your ADXL circuit board.

What is your resolution?  You can check that rotating your
sensor from + to - 1g, where "g" is the gravitational field of the
earth.  For example, Larry Cochrane set his ADXL05 up so that
his 16 bit a/c gave about 16000 counts/g.  Then each count 
represented about 0.06mg (thousands of a g).  See
What change in counts do you get when you rotate your board?
What does each count represent in your system in mg?  

Now how many counts +/- do you get when the unit is sitting
still?  Larry got about +/-30 counts, and since his resolution was
0.06mg, the noise level was about about 2 mg.    Just eyeball the
output for a minute or two to get a rough reading of the range
of the noise readings.  The ADXL202 spec sheet has a formula
that relates bandwidth (b) to noise level (n).  It is approximately
n=0.5*sqr(b) in mg.  So if your bandwidth is 15 hz, your noise
level is n=0.5*sqr(15)=~2mg.  If you see a lot more noise than that
you should recheck your circuit.

>         My AtoD is quiet as a mouse. I'm using an ADC0809 eight channel
> AtoD and am using but four of these channels. I'm using the printer LPT1
> interface.

Perhaps your resolution is much larger than 2 mg and that is why you
are not seeing any noise?  If you rotate your board from  +/- 1g 
and the a/d does not change by at least 1000, then you need more gain or
more a/d bits to make use of all the performance the ADXL202 is
capable of.

The ADXL202 has the same noise specs as Larry's ADXL05.  The 
newer single channel  ADLX105 has half the noise level of the
ADXL05 and ADXL202.  I recently built some boards with
the 202 and 105 chips but my results for both were similar to Larrys.
I did not see much improvement in the 105.  If the ADXL noise level
is about 2mg then the "shake maps" on the web suggest you
have to be very close to a 4m quake to see
anything with an ADXL based sensor.  


The Inexpensive Seismometer Project


Public Seismic Network Mailing List (PSN-L)

[ Top ] [ Back ] [ Home Page ]

Larry Cochrane <cochrane@..............>