Clip-On Thermoelectric Shower Booster

The most common type of shower is a head fed by a simple mixer-tap that variably combines hot & cold water. The hot water is usually supplied by the main hot water heating system in the building. These can be gas, solar, geothermal etc. but in the US as least the most common is a bloody great (slightly) insulated tank heated by a resistive electrical element. Heating things is energy intensive, water especially since it has a high specific heat capacity. This leads to big electricity bills, which are annoying.

Thermoelectric elements, or Peltier devices are small solid state chip-like things that get hot on one side and cool on the other when you apply the right kind of electricity. You can use them to make horrendously inefficient mini- fridges, they can be quite useful in temperature controlling small electronics etc. But they're still a sort of solution looking for a problem. Their simple construction and the wonders of mass production mean that these things are now dirt cheap - ~$5 or so for a 40-50W device.

Instead of using them as coolers, when used as heaters they get quite interesting. As an example, I considered replacing my aquarium heater with a Peltier. The tank in question needs about 30W to stay at 77F/25C. To achieve this, my 50W heater is active ~3/5 of the time, consuming 30W. Because the temp gradient between the room and tank is only ~5C, calculations suggest a Peltier could pump heat from the room with only a net electricity cost of 2- 5W*. That's a big relative saving. For 25W, it's probably not worth it, but it illustrates the point that Peltier elements are good at moving heat across modest gradients.

To boost a shower with Peltier element, you could heat the hot water. But, the delta of 50C or so would mean they're operating <link> so inefficiently, that there's no real gain. However, under the bath/shower tray/behind the wall, there is also the cold feed to the shower. Temperatures vary, but a range of 8-18C seems reasonable. No-one in their right mind uses a shower in pure cold mode, nor does anyone pop into the shower for a fresh glass of water, so we're free to manipulate the temperature of the cold side, where the temperature gradient is perfect.

So, we build our device. The market is the casual sunday hardware store browser. They've been looking at heat pump water heaters, but the $1000's more than a conventional unit stings. Our target price is the $100-200 range. We start with an aluminium block, with a radius relief cut/cast into it to allow good contact with standard copper pipe. Then, ~10 or so 45mm Peltier devices attached in-line, simple screw clamps heatsinks and fans. Pg9 of <link> suggests we can pump 55W using only 20W, but we get to use that 20W as well.

So, our 10-element design gets us 750 extra Watts of heating, good for ~3.0C warming at 1 gpm. This assumes the cold is 1/2 the total flow. All for 200W input, it's 370% efficient, the remainder of the heat is extracted from the underbath environment, heat is reclaimed from the draining shower water. The other components are a non- contact flow meter for switching, Hall effect or even sound will work. A controller, some heatsinks/fans and a fairly beefy DC supply.

The result to the user will be that they simply turn the mixer tap slightly toward cold since the average temperature of the input water is now higher. This will reduce flow from the water heater, reducing workload. There will also be lower losses since less heat will be lost through the hot pipes to the air. In summer, the pumped heat ~550W can be subtracted from any air conditioning load, saving electricity there.

Cleverly, the power supply can route through the floor/wall so that only low voltage DC is in the bathroom environment.

*online calculators exist, but power supply/control losses also do.