Lots of detail provided in case it's needed.
I'm continuing to tinker with remote proximity sensors driven by an Arduino (and eventually just an ATMEGA* chip and its safety buddies) and have a pretty good prototype so far. It picks up motion just like the PIR module and signals the same, which the Arduino turns into a mechanical knocking via 5V solenoid. The PIR module does not work behind insulated glass (predictably) and making it an outside unit requires more work (and power management for batteries) than I'm willing to invest at the moment.
To simplify things, I decided to try an x-band radar module so I can keep the units indoors and avoid weatherproofing. The allure of pointing a microwave device out through a brick wall to detect motion was quite powerful. :-) I'm using this Parallax module (data sheet).
The problem is the unit is picking up movement behind it in addition to anything antenna-side, even with the sensitivity pot set all the way down. Looking at the data sheet (and seeing the graphics representing its field - don't know the terminology), it does seem to be expected behavior. HOWEVER, I want to focus it into a forward-facing beam (relative to the device) so I can aim it outside.
I know enough to figure I need some kind of waveguide but every attempt at blocking its back with metal (aluminum foil, a cut Coke can, and a conical piece of 1950s pendant lamp I replaced with a ceiling fan) has failed. The module keeps picking up movement behind it. I even encased the whole thing in an aluminum project enclosure just to verify that I can in fact blind it. It seems somehow to go straight through the enclosure and pick up movement anyway. This was unexpected. I suspected interference but the detection ONLY coincides with actual movement.
So: what makes a good waveguide? How would you guys approach the problem of focusing a pulse doppler radar into a reasonably narrow beam for motion detection? Thanks for any insight you can offer. I'm happy to get more specific or clarify any of the above.
Note: I'm sure the subject could probably be more specific but this is my first electronics project and I'd rather be vague than confusingly incorrect. :-}
I'm still not convinced there isn't some sort of interference occurring. As I have no oscilloscope (yet), I'm not sure how to test this theory.
This thread seems to suggest my approach is flawed: apparently the module's state should be read from an analog pin with noise and time thresholds. I've been reading it on a digital pin HIGH/LOW.