PIR (outdoor) false triggering
dblaine
Posts: 6
Does anyone have experience using Parallax PIR sensors outside, especially at night? Are there known incompatibilities with using these sensors outdoors?
I'm trying to develop a sensor to detect someone approaching my car in the driveway.
To test out sensor placements, I have a Parallax PIR powered with 5V from a LDO regulator supplied by a 9V battery (supply is as quiet as you can get). The output connects to the gate of an NFET driving an LED+resistor directly from the 9v battery (the PIR's load is as light as you can get, and the switched current is from the unregulated side of the LDO).
Indoors, day or dark, it detects motion reliably. But outside, day or night, it false triggers every several seconds. The PIR's is typically looking at places like a shaded garden or wall, with no obvious IR radiation sources or air convection, although the night sky and the city beyond is sometimes in the field of view.
Any advice would be appreciated.
I'm trying to develop a sensor to detect someone approaching my car in the driveway.
To test out sensor placements, I have a Parallax PIR powered with 5V from a LDO regulator supplied by a 9V battery (supply is as quiet as you can get). The output connects to the gate of an NFET driving an LED+resistor directly from the 9v battery (the PIR's load is as light as you can get, and the switched current is from the unregulated side of the LDO).
Indoors, day or dark, it detects motion reliably. But outside, day or night, it false triggers every several seconds. The PIR's is typically looking at places like a shaded garden or wall, with no obvious IR radiation sources or air convection, although the night sky and the city beyond is sometimes in the field of view.
Any advice would be appreciated.
Comments
Generally speaking, these types of sensors are watching for relatively rapid changes in infrared light. Sometimes those changes can be caused by light being "blinked" on and off by the wind moving leaves or tree branches, etc back and forth across a warm source. They can also be triggered by sunlight, campfires, or distant IR light sources, etc. playing off of ripples on water. So look for things in the environment that might be moving and periodically blocking possible IR sources, even distant sources. Also note that walls, etc. can reflect infrared fairly well or might even store heat during the day and behave like IR emitters at night. If there are ponds or other water sources in the PIR's field of view, try to move the PIR so it's not going to see reflections from those ponds, etc.
http://www.parallax.com/StoreSearchResults/tabid/768/txtSearch/x-band/List/0/SortField/4/ProductID/606/Default.aspx
Bean
Thank you kwinn. I am in the process of hacking one of those. My only concern is modifying the circuit for battery power. These 120V based detectors are not designed for current efficiency. Still, your idea is showing a lot of promise, especially if I can modify it to use low current amps and larger resistors, and to implement sleeping.
Thank your ElectricAye. I really appreciate your thoughtful response. I had tried to mitigate the variables you mentioned before posting my question, but hadn't considered the water reflection effect. There may have been a few puddles in the field of view on some placements. I did test on a windless day, though.
There is an impressive group of beams and arrays about 1/2 KM away (2 meter, vhf, uhf, maybe a 20 meter dipole). But the false triggering only happens outdoors, so the culprit may be something very local, maybe the cell phone in my pocket, my car's RFID or one of the other gadgets I have around. I'll check that out and post the findings.
Got squirrels?
I once used non-parallax PIRs for bear detection around a campsite and I noticed that squirrels would set if off if they got within about 8 feet or so, but those PIRs were mounted only about 2 feet off the ground. Of course, deer also set it off. And so did distant campfires whose light flickered off the surface of a nearby lake.
I've often wondered if a person could reduce false alarms by using two or more PIRs aiming in from two or more locations so the lines of sight are nearly perpendicular, then write the alarming logic to require that signals from at least two PIRs had to be detected simultaneously. I've never tried it, but it's a thought.
No squirrels.
Does anyone know the Parallax PIR output drive spec with supplies at 5v and at 3.3v?
' File....... ' Purpose.... ' Author.....W. Stefan ' E-mail..... ' Started.... ' Updated.... ' {$STAMP BS1} ' {$PBASIC 1.0} 'INPUTS '-------------------------------------------------------------------- SYMBOL PIR = PIN6 'OUTPUTS '-------------------------------------------------------------------- 'RC-4/AP-8 SERIAL I/O '-------------------------------------------------------------------- 'CONSTANTS '-------------------------------------------------------------------- 'VARIABLES '-------------------------------------------------------------------- SYMBOL timer = B2 'EEPROM DATA '-------------------------------------------------------------------- 'INITIALIZATION '-------------------------------------------------------------------- Reset: PAUSE 5000 ' let PIR warmup PINS = %00000000 ' clear all outputs DIRS = %00111111 ' make P0-P5 outputs 'PROGRAM CODE '-------------------------------------------------------------------- Main: timer = 0 ' reset timer DEBUG "Waiting for PIR...", CR Check_Trigger: PAUSE 5 ' loop pad timer = timer + 5 * PIR ' update timer IF timer < 100 THEN Check_Trigger ' wait for 0.1 sec input DEBUG "Triggered!", CR, CRThe unusually cold air has move east and I finally tested the PIR outdoors at night in 25F to 40F temperatures. As soon as the temps got above freezing, the false triggering stopped. Although the humidity was never above 70% and I didn't see any condensation when the weather was cold, condensation may also have played into it.
Mystery solved. A possible solution for use below freezing may be to use multiple sensors followed by combinatorial logic before transmitting from a single RF node, as mentioned earlier in the thread. We will see.
Thank you all for your help and advice.