PDA

View Full Version : logging sensors that output millivolts



Brian Smith
06-08-2005, 01:33 AM
I· have a BS2 and I am currently using it and my laptop, with stamp plot lite,·to log x and y acceleration in a car. I am wanting to log tire temperatures with·3 infared temperature sensors. My problem is that I need to read milivolts from the sensor and I dont know how to do this with the BS2. Here is an example of the type of sensor I want to use
http://www.exergen.com/industrl/microirtc/Linear%20Reportr1.pdf
also this is a project I am working on at my university. Does anyone have a suggestion where I might be able to·buy these sensors at a low price or even donated?·I need something relatively small, under 5" long or so, and i need to read from about 70 - 250 deg. F.· If anyone has a suggestion on a different type (non voltage output) sensor, that might work too.

Chris Savage
06-08-2005, 01:44 AM
Hello,

·· Millivolts (Or any volts for that matter) would be read by connecting an ADC to the Stamp Module.· We carry 3 different ADC Chips, and there is plenty of example code out there for using them.· The number of bits of the ADC will affect the resolution, as will the reference voltage.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Chris Savage
Parallax Tech Support
csavage@parallax.com (mailto:csavage@parallax.com)

steve_b
06-08-2005, 01:47 AM
You might look in to Analog to Digital Converters. They read in an analog voltage and output a digital word that the stamp would ingest.
Lean towards the serial ones rather than parallel ones (less pins to use up on your stamp).

Depending how you do for 'noise', you might look in to a small buffer/amplifier to boost the voltage up. I looked at some of the graphs and you'd be reading less than a volt when standing still and I can see you'd want some good resolution as your "interested temp range" doesn't differ by much more than a volt.

** Someone beat me to the post....ah well, this is just another vote then!

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
·

Steve
http://ca.geocities.com/steve.brady@rogers.com/index.html
"Inside each and every one of us is our one, true authentic swing. Something we was born with. Something that's ours and ours alone. Something that can't be learned... something that's got to be remembered."

Paul Baker
06-08-2005, 02:03 AM
Many ADCs have a Vref pin where you can set the full scale voltage (maximum voltage of signal you're measuring), you can use this to increase the resolution of your measurement over the reduced voltage scale.

steve_b
06-08-2005, 03:09 AM
If he's using this in an automotive application, wouldn't there be more of a noise issue? And would connecting up Vref not induce the noise in to the chip....

This has come up a few times as far as power quality in a car....lots of filtering! lol

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
·

Steve
http://ca.geocities.com/steve.brady@rogers.com/index.html
"Inside each and every one of us is our one, true authentic swing. Something we was born with. Something that's ours and ours alone. Something that can't be learned... something that's got to be remembered."

Chris Savage
06-08-2005, 03:27 AM
Steve,

·· There's no doubt that there's going to be some considerations to be made, but that gets him started.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Chris Savage
Parallax Tech Support
csavage@parallax.com (mailto:csavage@parallax.com)

Paul Baker
06-08-2005, 03:31 AM
Yes noise would be a real problem, after looking at the link provided, these are just temperature sensors, much like thermocouples. The output needs to be compared with a cold junction then use the graphs provided to calculate the temperature. This is typically done using a chip such as this one (http://www.analog.com/en/prod/0,,766_811_AD595,00.html). The AD595 is designed for K type thermocouples so that particular chip may not work with the temperature sensor you are interested in, you have to compare the termerature curves to know for certain.

Ken Gracey
06-08-2005, 11:22 AM
Dear Brian,

If you continue to use the device you are using, take a look at how the bias and span can be set with an 8-bit A/D converter, as in our·Process Control (http://www.parallax.com/dl/docs/books/edu/ic.pdf) text. Using a precision potentiometer it is easy to get 8 bits of resolution·over a narrow·voltage span.

Also, see the section on op-amps in our Basic Analog and Digital (http://www.parallax.com/dl/docs/books/edu/baad.pdf) text.

But it seems like you want a non-contact method of measuring temperature, perhaps even pre-calibrated. My favorite approach would be to use the Melexis Infrared Module (http://www.melexis.com/mlxstore.asp?search=MLX90601&Submit=Site+Search&family=MLX90601). Linear output, no calibration, easy to use. It could be that the issues of·measuring a low-voltage output would be offset through the selection of a more appropriate sensor which provides an easy-to-use output.

Good luck on the project and be careful.

Sincerely,

Ken Gracey
Parallax, Inc.

Brian Smith
06-08-2005, 10:36 PM
Thanks everyone. You all really replied quickly!

The Melexis IR sensor looks like a nice way to go. Its cheap too, but I am not exactly sure how to read the signal. I figured the PWM output automotive model would be the best for me. Page 15 of this http://www.melexis.com/prodfiles/MLX90601_1.PDF explains the output signal.

T = 102.4 us (92.16 to 112.64)
T1 = 1/8 * 102.4 us = 12.8 us (no color)
and then I measure the high signal for the next 101.2 us and divide by 101.2 us to get duty cycle. (green)
any high signals after this are error signals (yellow)
right?

I think my biggest problem would be figuring out when the period started. could anyone post some code that would be used to read a sensor like this? If you cant tell, Im really new to this.