IR Sensors

[koossa]

Thank you!!

[Jimbo]

If you are planning to use a PIR sensor outdoors, be aware that they tend to
pick up *ANYTHING* moving. I have a neighbor with PIR-activated spotlights
and the things go on every time a field mouse sneezes or a bug flys past and
momentarily blocks some more distant heat source.

PIRs respond to a *change* in temperature, take one of the little plastic lenses,
bend it like in the unit and move past a light source. You'll see that they're
designed to create a parade of little images that slide past the sensor. A moving
object will cause a hot spot to reach the sensor, then there's a null, and then
the next image comes by. The 'sensitivity' setting on some sensors generally
sets how many times the hot spot has to come by in a given time to trigger the
device.

The best way to get around the "Sees everything" problem is to use two or
more sensors set to have somewhat overlapping fields of vision in the zone
you're interested in. Hook the outputs up "AND"-style, so both sensors must
be activated at the same time before you register a positive hit. Likely, a
mouse or whatever will be in view of one sensor but not another. Aiming them
at slightly different heights - say 3 feet and 5 feet - is also useful since all
but midgets will trigger both, but a dog or cat or mouse won't be tall enough
to trigger the 'high' sensor.

An alternative to PIRs or straight thermopiles is to use an IR "distance sensor".
Sharp makes a lot of these, and they're pretty cheap and straightforward. If
you've ever been in a public restroom with the automatic urinals and sinks,
this is what they're using. They work on 'triangulation' - an IR beam goes
straight out and then the reflection hits a photodiode strip behind the receiver
lens. Close objects illuminate just the cells at the edge of the strip wheras a
more distant object is fully "in the picture" and illuminates more. Max range
last time I checked was about four feet - and I've got part of a PBP program
for reading them. In theory an external lens can extend the range somewhat
at the expense of accuracy.

They ARE pretty accurate too, about 1/10th of an inch through most of the
operational range. Won't work if the optics get fogged-up though. Got 'em
from some robot-supply company. Some models supply a 0-3vdc output
you can use with any PIC with an A/D converter. Others source current.
I don't recall if any output a frequency-changing pulse or not.

http://www.mouser.com/search/Product...2-GP2Y0A02YK0F

http://document.sharpsma.com/files/gp2y0a02yk_e.pdf

A related cheap way to sense someones approach is to buy one of those
"ultrasonic rulers" from the hardware store and wire into it in some way.
I've seen some for $19.95 and they sport a range of 12 to as much as
25 feet. HOW to hook 'em up would depend on the exact model and
what kind of info you want from it. They're a fairly narrow beam, so
interference from nature is reduced. Alas, a hard rain might reflect
back enough of the ultrasound beam to create false data. Thermal
sensors aren't so easily fooled by rain.

-Jimbo

[koossa]

Thank you Jimbo!!!!