Nobody likes to monitor things as much as a hacker, even mundane things like sump pumps. And hackers love clean data too, so when [Felix]’s sump pump water level data was made useless by a new pump controller, he just knew he had to hack the controller to clean up his data.
Monitoring a sump pump might seem extreme, but as a system that often protects against catastrophic damage, the responsible homeowner strives to take care of it. [Felix] goes a bit further than the average homeowner, though, with an ultrasonic sensor to continually measure the water level in the sump and alert him to pending catastrophes. Being a belt and suspenders kind of guy, he also added a float switch to control the pump, but found that the rapid cycle time made his measurements useless. Luckily the unit used a 555 timer to control the pump’s run time after triggering, so a simple calculation of the right RC values and a little solder job let him increase the on time of the pump. The result: a dry basement and clean data.
We recently discussed the evolution of home automation if you want to know more about the systems that sensors and actuators like these can be part of. Or for a more nuts and bolts guide to networking things together, our primer on MQTT might help.
Something’s off in his write-up. He says his controller had a 100k resistor, giving him a 5 second delay. But the picture shows a 200k resistor, which should already give him the desired 10 sec delay. (The instruction manual also states the controller should have come with a 10 sec delay.)
Other than that, the write up was clear and well documented, so I kind of hate to point out the error.
It took me entirely too long to figure out why his graphs were suddenly “messed up”. I’m assuming the pump didn’t run very long each time but ran more often, leaving the basin going back and forth between levels very close to the tripping point and not showing up well in his graph. Since there’s no indication from the level meter of how many pump runs there have been, it’s not easy to see usage at a glance. Running the pump longer pulls more water out which shows a difference in level better in his graph, while running the pump less frequently.
Interesting read. Good find :)
Urg. You take a good quality submersible pump, and a good quality sealed float switch and you have a system that will run for a decade or more. Go and ass electronics to it and you cut the reliability markedly. There are a lot of places where electronics add no value,and there are some where they actually hurt reliability. This is one of those cases.
You must be new here…
Why not just get a water-powered backup pump installed? It would cost more, sure, and isn’t as techy and hacky… but hey it doesn’t require any power. Normal electric one will work like normal, but if it fails and the higher-mounted backup float is triggered, then the water-powered one would take over.
I am on well water… So a water-powered backup pump is not an option.
For my sump pump backup system, I use two 2.2kW APC UPS, with four 12V, (48V system) 75Ah batteries each.
One UPS per sump pump (just encase one fails).
Run time at 100% on time, is around 4 hours… If the power is out longer than that, then I start the generator. ;)
The last flood, I had 3 sump pumps going:
https://youtu.be/bQQu8ugYKos
That’s pretty cool. The windshield wipers make the video sound like something out of a horror movie with its heartbeat-like sound.
I’ve had one week where it flooded that bad and I sure wish we had had a backup system, because it was overwhelming the main system.
Spent the next week with the wet/dry vac in the basement… :/
I’m late to the party, but I have to say I love my water powered backup pump. It was really easy to install and I test it once a year and it runs flawlessly. We had a really weird storm that knocked the power out for over 24 hours, and even neighbors with battery backed pumps got flooded. I have said before that if the power is out AND the water is out, there’s probably no house left to protect.
Only one bit of data is needed. Sump pump failure detected when water level reached X inches above max. We simply use a second sump pump with a float at a higher level to detect this, make an automatic assumption the first sump pump failed, and therefore the secondary sump pump should kick in. This is like… 30+ yr old tech.
Only one bit of data is needed. Sump pump failure detected when water level reached X inches above max.