Water damage can quickly make even the nicest buildings unliveable. [Andres Leon] suffered a small flood from an air conditioning unit, and wanted to avoid such issues in future. Thus, he built a wireless monitor to solve the problem.
The device is based on the ESP8266, allowing it to wirelessly communicate with Home Assistant. Thus, if it detects water via its rust-proof probes, it can notify Home Assistant via an MQTT message. From there, Home Assistant can advise the home owner remotely via phone and email. Plus, just for completeness, there’s a loud buzzer in the unit that goes off when water is detected, too. Thanks to a 2500 mAh lithium-polymer battery on board, the device can run for up to 5 months between recharges.
Integrating warning systems into one’s smart home system can be particularly useful when one is away for long periods. Things like water leaks tend to do damage over time when we’re not paying attention, so any IoT device that can assist in this regard is helpful. If you want to investigate the cause of a difficult leak, though, this other project may help. Video after the break.
Good plan… But battery power means someone need to remember to charge it. A small battery giving 8-24 hrs runtime is a great idea in case of power outage. 5 months is long enough for someone to forget to charge it. This would be far better powered from a small wall wart.
The HA interface include a low battery warning – no need to remember to charge it, it’ll report when it’s needing a charge.
Ok, but then someone gets busy and dismisses the notification, or forgets it on the charger so it misses a leak (it’s high and dry on it’s charger)
Battery powered means opportunity for human failure.
How much handholding should a person expect? I use commercial water detectors (Shelly and Grohe). They periodically wake and say hi to the system. If they don’t check in for a while I am informed that they’re offline. If I ignore that, I wouldn’t blame the sensor.
“Battery level warnings – Notifies me via Home Assistant when battery level is low.”
That’s probably good enough, but a low battery beep (like a smoke detector) would be a nice addition as well.
My primary suggestion would be to put the water sensors on a breakout board connected to the box via a decently long pair of wires. Then you could mount the electronics box safely out of the way as well as being able to put the sensors as close to the water concern as possible.
I’m the creator of this project. I built this back in 2019 and have been running these sensors since then with very little problems. My biggest concern was the battery and how long it would last. However they have turned out to last a great deal longer than I expected. They are programmed to wake up once a day to report their battery status and with that frequency each battery has lasted an average of 9 months (about 250 days). I can’t really explain why, but that’s the reality of it.
They have also saved my bacon at least three times already. the buzzer is not very loud but the phone notification has been very reliable. my water heater in the garage begun to leak and that saved me once. Then the AC condenser started leaking and that was another save. So overall this has been a very successful project!
The typical problem I have is that if the sump is overflowing, it’s because the power has been out longer than the UPS can run my server, modem, and router… so no email or phone call. Often, if the power is out, the local C*mc*st broadband is out anyway, so again, no notification. Too bad 1G and 2G are dead — does anyone sell a really really cheap 3G Pay-as-you-go data service?
If you already have more than one mobile device on your account it’s usually really cheap to add one more.
3G is dead too, but you can get cheap ultra low data service from Twilio IIRC – 4G/LTE
I’ve used Ting in the past
Not all water situations are related to Sump backup due to power outage. These can be used to detect when your hot water heater gives out, if a clothes or dishwasher springs a leak, etc.
If your on Google Fi you can get several data only sims.
Great project, I’ve been needing something like this for a long time.
I bought a bunch of Insteon Water Sensors with a similar design. They worked and save us some pain, however the design is basically one time use. The water intrudes into the case and corrodes the electronics.
The newer water sensors I’ve seen have a replaceable probe allowing you to place the sensitive electronics higher up and out of harms way.
Nice idea. Rather than the 13 dollar water sensor I would suggest 2 stainless steel bolts. 4mm probably fine
Nice project, for years I’m running the Xiaomi water sensor with zigbee. Battery life is great more then one year
I’m a huge believer in the Hacker ethos. If I can make it, fix it, or modify it myself I do. That said, I’ve also learned that you have to value your own resources and time and make sure it’s worth doing.
I’ve been replacing my Wink based smart home system with a Home Assistant based one. In that vein, I’ve been building most of my own sensors and other smart home devices. Many based on ESP8266 or ESP32 boards using ESPHome. One device I very much wanted was a flood sensor. After looking at my options, I finally decided I was better off just buying one. In my case an Aqara model SJCGQ11LM which works perfectly with Home Assistant, is about the size of a juice bottle cap, and has battery life measured in months.
Here are the issues with sensors like this.
One, you have to keep the main components away from what’s being sensed. In this case water. This leaves you with the choice of either trying to waterproof an integrated unit, or use a remote sensor on a cable.
An integrated unit must be small enough to fit in the low spot in the area you want to monitor, often a small hole in a concrete floor for a drain or clean-out port. A remote unit must run the cable to the main unit without getting in the way or be at risk of being cut, tripped on, or caught up in something.
Another issue is power. These types of sensors often sit in hidden areas of the house where they are easily forgotten. If you build an integrated device small enough to fit into the intended location, battery space will be limited. That means either short battery life and frequent charging or battery replacement. Something you can forget about or get frustrated with and skip, eliminating the point of having the sensor. To make it work you have to get into some real electrical and software engineering to use low power components and deep sleep modes. With a two part design you can use larger batteries or even a wall plug, but you’re back to the cable run to the sensor issue. Also, an AC powered unit has the issue that flooding often happens during a power failure.
Like I said, I’m all about hacking and making, but like anything, you have to evaluate each use case for yourself and decide if it’s worth it for _your_ use case. That said, it’s still important that some people do experiment with their own solutions. I have absolute confidence that this community is capable of eventually coming up with a DIY, Open Source design capable of competing with units like the Aqara on size, price, and battery life.
Absolutely spot on. When I was younger, with fewer responsibilities in life, I loved putting lots of time into making things that didn’t work as well as commercial products, because I enjoyed the process. Now, I have more demands on my time, and have to be more selective with how I spend it.
Engineer1000, I get it it, but don’t undervalue those peojects from your younger days. Sure if you measure your time in dollars per hour, it doesn’t make sense… But you were solving real-world problems (albeit not always perfectly) within real-world constraints, and probably a very restrictive budget. Imagine if your were building those projects as part of a hackathon, or an internship, or a univeristy course – you’d have to PAY for the privilege and still probably wouldn’t learn as much. So, when we see “yet another flood detector/weather station/power bank” on here, remember that the author is one of today’s 10000 [1] and their hacking together something that could be bought off the shelf is often MUCH more economical than racking up the equivalent semester of student dwbt and time :-)
[1] https://xkcd.com/1053/