Flooding inside your home can have devastating effects. Protect yourself by building a simple flood alarm.
I've heard many stories of colleagues and neighbours that had to deal with flooding in their homes. Often it's an old pipe that bursts or a plastic coupling that breaks somewhere. The consequences can be dire, involving replacing all your carpeting or wood floors and dealing with the moisture/smells afterwords. Fortunately, I haven't experienced something quite this bad. The other day however, one coupling under my kitchen sink was leaking very slowly and that went undetected for a day or two. By the time I detected it, it had messed-up the wood under the cabinet pretty bad. This experience got me thinking about moisture sensors and flooding alarms and hence this article. I plan to eventually build something more sophisticated, but in the meantime, this simple approach should be enough to sound a loud alarm if water is sensed in the floor where this sensor is installed and it took me less than an hour to build.
What to use?
If you have been reading my articles, you know I'm a fan of "re-purposing" gadgets I already own towards something useful. I found in my garage this item I had bough at a Dollar Store (See Figure 1). It's marketed as a "LED Light with Panic Alarm". (Here's a similar one from Amazon.com, though a bit more expensive).
The operation of this device is quite simple. Press the top button to activate the LED light and pull the pin on the left to activate the "alarm". See Figure 2.
Figure 2 - Operation
So it seems this little gadget already implements the alarm portion (which I can use) and also has a switch that enables it. As a bonus, I can keep the LED and button as a periodic "battery health" monitor. Press it once in a while to see if the battery still has some juice left... We are almost there! Time to open-up the thing to see how it works:
Figue 3 - The Insides
The circuit is remarkably simple. There's a push-button and current limiting resistor for the LED at the bottom. The top portion contains a small PCB with a small IC (presumably generating the sound) a transistor and a transformer driving the piezo buzzer (shiny disk toward the bottom half).
Power comes from three 1.5V button cells connected in series (4.5V total). The standby current consumption when nothing is activated (no LED nor buzzer) is so low, I wasn't even able to get a reading using the most sensitive 200 uA scale in my multimeter (Figure 4). This is a good thing, since you want the flood alarm to be active for many months without battery replacement.
Figure 4 - Almost zero Current
After some probing with the meter, it became clear that when the pin is pulled out, the contact points marked as "+" and "-" in Figure 5 are connected (i.e., this is the normally-open portion of the switch).
After removing the input jack, my first thought and experiement was to simply connect some wire to the "+" and "-" terminals and hope that, when water was present between these two "electrodes" it would be sufficient to activate the alarm. Unfortunately, this didn't work. Water is a conductor but not a very good one and this circuit evidently needs a pretty low resistance to activate. The solution to this issue proved quite simple though. Since I had a BS170 N-Channel MOSFET available, I ended-up implementing the circuit shown in Figure 6.
Figure 6 - Final Circuit
Note that the "IN" pint in the schematic above corresponds to the "-" in Figure 5. Resistor R1 pulls down the Gate of the MOSFET so it is normally in the Off state. When humidity/water is detected between the electrodes, the gate voltage increases beyond the MOSFET threshold and the 4.5V is connected to "IN" through the Drain to Source channel. The larger the resistor, the more sensitive the circuit will be though I found 3M3 to provide more than adequate sensitivity. (If you have a P-channel MOSFET available instead (say a BS250), then the circuit would be similar but with the resistor between Gate and the +4.5V supply. the electrodes would connect to the Gate and Ground. This is probably a better implementation than Figure 6 actually:).
Figure 7 shows a detail view of the BS170 installed in the Pcb (replacing the input jack).
Figure 8 shows the completed modification with the 3M3 resistor and the electrodes made of thin solid copper wire:
And that's it! Figure 9 shows the completed product (am I the only one noticing a remarkable spermatozoid resemblance ?). To install it, just make sure you position the sensor electrodes where they would be in contact with water should the flood happen. In my particular installation, I glued it to a corner wall in the cabinet with the copper electrodes touching the floor. Hope never to hear this little fellow, but if I ever do, I'll know my time in this project was well spent:)
Figure 9 - Final Product
Comments, questions, suggestions? You can reach me at: contact (at sign) paulorenato (dot) com