The Bird Blaster v1.0
I keep a pair of peafowl as pets. Peafowl (the male in particular) like to look at themselves in anything that’s reflective. They also like to eat potted plants. This leads to a lot of peacock poop and damaged plants on our porch, which leads to me being in trouble with my wife :). Being an engineer, I figured the only reasonable solution would be to make a motion activated sprinkler to discourage them from hanging out near the house.
This was my first real electronics project, starting in the fall of 2011. My plan was to use a motion activated security light, a sprinkler valve from an automatic watering system, and a custom circuit to connect the two.
I started with the motion detector. I purchased the cheapest motion activated security light at the local Home Depot, a Zenith SL-5408. Searching around on the web yielded a few examples of people hacking the motion detector and others like it for other purposes. Similar to those projects, the first step was to separate the PIR motion detector unit from the light fixture. With this complete, it was clear that the motion detector was fed 120VAC (black and white wires), and returned a switched 120VAC leg (red wire), which powered the lights.
I then attempted to figure out how to run the motion detector circuit off of a 5V or 12V DC supply. While I was able to follow the circuit well enough to locate the 12VDC and 5VDC locations, providing power at the appropriate locations didn’t result in successful operation of the motion detector (I’ve since figured out why, but I’ll cover that in a future post :)). I decided to give up on DC power, opting instead to use 120VAC wall power that the detector was designed for. This made the setup less convenient, but at least it worked.
With the power supply sorted out (given up on…), I needed to convert the output of the motion detector to something usable by my external circuit. In its stock form, the output of the detector is a relay, switching 120VAC for the lighting assembly. An NPN surface mount transistor is used to pull the relay coil to ground when motion is detected, closing the relay contacts. I could have used the relay contacts as the output of my circuit, but that was a bit overkill for my application. Instead, I desoldered the relay, and used the coil drive transistor as an open collector output.
There are two user adjustable settings on the detector, an ON time duration switch (20 min, 5 min, 1 min, TEST), and a sensitivity adjustment. For the purposes of the bird blaster, the detector is always used in the TEST mode. This makes the detector work during the day (it disables the photocell), as well as making the ON time for each motion event relatively short (~5 seconds). I was able to adjust the sensitivity to reliable detect the peacocks, without too many false positives.
The sprinkler valve I picked up was also the cheapest I found at Home Depot, an Orbit 62035. Apparently, this valve is no longer sold. This one appears to be similar, but the connector is proprietary :(, and I don’t know if they’re electrically compatible. I found some good references on driving the 62035 valve here and here. According to those sites, the valve latches ON and OFF, requiring a separate 22 millisecond pulse to open and close the valve. The stock valve controller runs off 3 AA batteries, which led me to believe I would be able to actuate the valve with a 5V signal. Unfortunately, that wasn’t the case. As referenced in the previous links, the controller has a voltage boost circuit that charges a capacitor, which is then dumped into the solenoid coil to change the valve’s state. I did a few tests, and concluded that 12V was sufficient to operate the valve, so my circuit would operate at 12V.
For the circuit connecting the motion dectector to the sprinkler valve, I ended up using a combination of three 555 timer circuits. Each 555 circuit was in the “monostable” configuration, providing a single timed pulse each time the input is triggered. The first circuit controls the ON coil of the sprinkler valve, sending a 22ms pulse when the NPN transistor of the motion detector gets pulled low. The second 555 circuit is triggered at the same time as the first, and is used for a 2 second delay, controlling the ON time of the sprinkler. When the 2 second delay ends, it triggers the third 555 circuit, another 22ms pulse, this time for the OFF coil of the valve.
The monostable 555 circuit alone doesn’t work in this application, because it requires the input pulse to be shorter than the output pulse. Because the motion detector output is pulled low until motion is no longer detected or a minimum of around 5 seconds, the short pulses required wouldn’t be achievable. I found a solution for this in a “trigger” circuit shown on doctronics.co.uk. This circuit is essentially a falling edge detector that creates a very short 0V pulse whenever the input transitions from high to low. Each 555 circuit was provided with its own trigger circuit at the input.
Putting it all together, the circuit looks like this:
Putting it all together resulted in a perfectly functional peacock deterrent. However, the requirement for both 120VAC and 12VDC power supply was somewhat inconvenient. I also never moved the circuit from the breadboard to something more permanent, so I didn’t end up using the setup long term. I’ve since revisited the idea, in an attempt to make it more convenient and functional. Check back later for an update with all the improvements :).