Hi Nic,

 Very good questions, most people ignore the calibration issue for sensors.

The internal temperature sensor of the microcontroller on both Tmote
Invent and Tmote Sky does require calibration.  TI recommends a
one-point calibration; however, we have found that a two point
calibration can yield less than 0.4C error.  How do we calibrate?
Very simply:

Place all your nodes in the same location and read the temperature,
presumably using Delta.  Store the raw ADC reading and the current
temperature from a reference device (we use a Tmote Sky with the
Sensirion SHT11 external sensor as the reference).

Next, put your motes into a fridge running Delta.  Allow the
temperature to stabilize (about 10 minutes). Record the temperature of
the reference mote and the raw ADC readings of each of the motes.  Now
your calibration coefficients are simply a matter of calculating
y=mx+b.

For light sensors, you have the same situation.  Put all of the
sensors in the same location and record the light levels.  You then
need to record the reference light levels.  Remember that light
sensors that are photodiodes have a logarithmic response.

Also, your initial email has a slight inaccuracy--the S9067 is used in
Tmote Invent, and the Tmote Sky uses the S1087 and S1087-01.  Each
sensor has a different response.  Also keep in mind that Tmote Invent
has a variable gain implemented as a variable resistor, whereas Tmote
Sky sensors have a fixed gain.

For the accelerometer, common calibration is performed by placing the
device on its edge with the X-axis facing towards the ground.  This
gives you the zero point for the Y-axis and the 1g acceleration for
the X-axis.  Repeat for the Y-axis to complete the calibration.

-Joe

On 6/14/06, nic <[EMAIL PROTECTED]> wrote:
Hi,

Part of what I'm currently doing will require the comparison of the
sensor readings taken from multiple sensors of the same repeating event.
The sensors I'm using are a mixture of tmote sky and tmote invent
devices. I was hoping that some people might share their experiences in
calibrating a number of such sensors so that the differences I observe
in the readings aren't due to calibration issues.

I'm just using the temperature and light sensor. So I'm using the msp430
internal temp sensor and the particular light sensor that is present on
both the invents and the sky devices, which I was told was the Hamatsu
S9067-01.

So if anyone has any experiences that maybe relevant I'd love to hear
them. For example - any simple methods found to be effective, perhaps
any intuition as to the extent of such issues I'll see - and so on.

Thanks for your help.

Nic

_______________________________________________
Tinyos-help mailing list
[email protected]
https://mail.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help

_______________________________________________
Tinyos-help mailing list
[email protected]
https://mail.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help

Reply via email to