Category: Sensor Workshop

My sensor report, “Soft Sensing: Using Conductive Fabric and Thread as Input“, can be found on the wiki for Tom Igoe’s Sensor Workshop class.

In this assignment, data travels from the sensor to the microchip to the computer to the server to a text file. Very exciting! Setting up the PHP and text files was pretty straightforward. Doing the whole “chmod” thing took a little while because I’m not quite familiar with terminal commands. For the pic/Processing setup, I started out with a trying a simple call & response just to make sure that everything was working. I then moved on to sending from Processing to the network. All code came from Tom’s Network Datalogging Suite.

The intial success that I had in the lab with a potentiometer was very excited. I then took the setup home and used a photocell measure the sunrise Sunday morning. I set it up to take readings every 30 seconds and left it running for a couple hours. I was pleased with the results.

The only real problem I ran into was a discrepancy in time:

This turned out to be an issue with the PHP code (“M” vs. “m”). I plan to redo the photocell thing so I can have readings with proper time stamps.

PRODUCT PAGE DATASHEET

I decided to look at the Eltec Pyroelectric detector IR-Eye 442-3 Integrated Sensor from Acroname. It detects infrared radiation (IR) which is electromagnetic radiation with a wavelength is longer than that of visible light. IR is generated by objects that generate heat. The sensor is able to detect people or objects and be can be used for purposes such as intrusion detection, lighting control, robotics, motion sensing, automatic door control, or safety warning.

Pyroelectric charge is the electric charge that is created in response to infrared radiation. This particular sensor uses lithium tantalate, which Wikopedia defines as a “material similar to lithium niobate, possesses unique optical, piezoelectric and pyroelectric properties which make it valuable for motion detectors, cell phones and possibly nuclear fusion”.

If I understand it correctly, the sensor looks not for a certain temperature, but rather for a difference or jump in temperature. [Is this what “Parallel Opposed Dual IR Detector” means?] This becomes a problem when the person or object you are trying to detect remains motionless- they essentially “disappear” or fade into the background and the sensor is unable to see them. This can be addressed by having the sensor move back and forth in a sweeping motion.

This sensor is small & circular. It measures .36″ in diameter.
It has 4 pins: power, output, reference voltage, and ground.
It operates with a range of 5 to 15 volts with a maximum current of 2.2 milliAmps.
The field of view ranges between 30-60 degrees, making it extremely directional.
It has a temperature range of -40 to +70 degrees Celcius.
An internal OpAmp provides 100x signal amplification.

Examples of use:
http://www.seattlerobotics.org/encoder/200012/pyro.htm
http://www.robotics.com/robomenu/pogo.html

Questions:
What is…
parallel opposed
integrated analog signal processing
high gain
burn-in test
common mode rejection

For this assignment, I originally used Spectra Symbol’s linear potentiometer. Though I was able to complete the assignment, it was a frustrating re-entry into phys comp ridden with silly problems that made me feel like I had taken three giant steps backwards over the break. I also managed to screw up my sensor before I was able to take any screen shots. So in an effort to clear the air and give a better start to my work this semester, I decided redo it on a better day with a flex sensor and ended up having a much more satisfying experience.

This was actually my first time using a flex sensor so it was great to work through some incorrect assumptions I had about the way they work. I thought that the flex was measured when the sensor was sort of bent in half, but rather it’s when you stablize the base and the bend it just above the base.

The other assumption that I had was that it worked when bend in either direction, which is completely untrue. When bend the wrong way, values simply jump to zero.

My data was surprisingly clean, so for my Pic code, I just sent it straight to serial out without having to eliminate noise. For the Processing end, I just did a basic variation of Tom’s Datalogger code.

The first graph illustrates bend the flex sensor in the right direction- slowly at first and then several quick bends. The second shows the value drop when you bend it in the wrong direction.

cell phone alarm
elevator door
magnetic card swipe
bar code scanner
heart rate monitor
laptop touch pad
remote control
ipod wheel/buttons