Hi Oliver,

I haven't actually thought about an automatic bath filler, but I have given some thought to something kind of similar.

In my experience, most of the times that a sewer backup causes a water overflow in one's house, it isn't because of a localized toilet obstruction. Most of the time it happens because roots grow in the sewer line. Instead of a complete blockage, the roots create an obstruction which slows the flow rate of the water. In most cases, the sewer line itself provides enough capacity to accept most discharges (toilet flushes, washing machines, dishwashers, bathtubs, etc.), allowing the waste to trickle through at a slower rate. But then, over time, solids in the waste stream become entangled in the encroaching roots slowing the flow rate even more. Eventually, the obstruction is sufficient that water backs up into the house, and a mess ensues.

Now if this happens when someone is flushing one of the toilets, disaster is usually averted because the person doing the flushing sees the problem and intervenes before any overflow occurs. Unfortunately, this problem is usually first manifest when a large discharge occurs, much larger than a toilet flush, like a washing machine emptying. Water backs up into the house and the first indication is when someone goes to the toilet and finds a wet floor.

I thought about how one could monitor for such an event, and it occurred to me that the first place such an problem could be detected would be in the cleanout drain. By putting a sensor in the cleanout drain, by detecting the rising water level in the sewer line at a much lower level than the lowest floor of the house, one could avert a backup before the unsanitary mess makes it into the house.

So the next time someone is looking for a student project, design a water level sensor/indicator intended to monitor the water level in a building's sewer clean out drain. Do it cheaply enough and you might even sell a few. I can tell you that it's one time that an ounce of prevention really is well worth a pound of cure.

And if you want to make your sensor even more marketable, design it to interface to a standard alarm system panel. In other words, make it run from a +12V DC power source and provide a SPDT dry contact (galvenic contact) output. Even if a homeowner isn't inclined to install such a device himself, an alarm dealer could easily sell it as an inexpensive add-on to their system.

In fact, if you want to get fancy, add a second output to indicate when water flows without rising. If a homeowner has a house unattended there should be no water flowing out. If there is, something is wrong and they would likely want to know about it promptly. Even something as simple as a running toilet can cost several hundred dollars if left running for a couple (or few) weeks. (I found this out, the hard way of course, when selling my house in Dallas after already having moved to Oklahoma.)

Anyway, it's just a thought.