OT: Not a smoke alarm

Hey SNIPPERMAN, READ thy manuals rather than blaming the plumber (assuming that you CAN read!). Also, have a drain pan or equivalent under your hot water heater - they can, and DO, leak (fundamental basic home smarts).

At least your batteries hadn't leaked and destroyed the water hear electronics...

John ;-#)#

Happily they were dead boring zinc-alkaline Duracells.

It's a Rheem Stellar - but you didn't read carefully enough. It wasn't the water tank or it's heater that created the problem, but a completely separate Evolve battery-powered water sensor and cut-off valve. What I'm pissed off about was that I wasn't told that it was completely separate. If I'd looked hard enough, I would have been able to work it out for myself, but who wants to contort themselves into the positions that let you look closely at what a plumber has squeezed into a cupboard? As it was I had to crawl around on the floor to work out what was going on, and replace the batteries.

They could have used one of those batteries which are stored with dry electrolyte, and only get activated when they get wet. I found one once in a surplus radiosonde that I bought on eBay out of curiosity. The shelf life would be very long if it stays dry. Obviously you would have to install it in the pan, so that it gets wet when it is required to work. Finding a replacement might be difficult for consumers, but they likely require a new tank anyway at that point. There would be a problem if the user decides to tip in a jug of water to test the shut-off valve though, as those batteries fail quite soon after they are activated.

They could have also use a normally-closed solenoid valve running from mains power. The coil power would be a significant waste though, which could be mitigated a bit if they could put it inside the thermal insulation so that the waste heat goes into the tank. Still it would mostly be a loss if the water would otherwise have been heated with a heat pump using a sixth of the electricity per unit of heat. The coil power could be reduced even more if it were only activated for a time after the tank pressure drops significantly, but the sensor for this might be expensive and the clicking could get annoying.

One thing I have seen that might use a lot less electricity than a normally-closed solenoid valve is to use a spring-loaded valve opened using a gear train and small motor. I have some air conditioning damper servos that do this. An enormous spring closes the damper on power loss, and a tiny motor with very high ratio gearbox winds up the spring to allow the damper to move freely when the power is on. A second motor then controls the damper position when the spring is wound up and out of action.