Self-regulating Water Heater


Most everyone is looking to live a little greener these days, with motivating factors typically being the preservation of the environment or financial considerations. [Fabien] fit into the latter category after realizing that about 25% of his monthly gas bill went to heating the water he and his family use each day. After a few calculations, he found that they only required hot water 68 of the 168 hours per week that the water heater was typically running. He figured the best way to save a few dollars was to rig the water heater to turn itself down when it wasn’t being used.

He connected a servo to the temperature control knob on his water heater, allowing it to be adjusted by a microcontroller. Having a rough idea as to the schedule his family keeps during an average week, he wrote an application for his Netduino that would actuate the servo when needed. A DS1307 real-time clock was wired to the Netduino for accurate timekeeping, so as to ensure [Fabien’s] wife never had to endure a cold shower.

It’s a shame that most water heaters don’t ship with some sort of programmable thermostat like you see with newer HVAC systems, but this hack is definitely a step in the right direction.

Continue reading to see his power-saving water heater in action.



74 thoughts on “Self-regulating Water Heater

  1. Not to discourage him i love this hack and have thought about doing it my self….until i lost power durring a recent snow storm for 2 days… and did not loose hot water until the end of day 2… so no power for 48 hours… and still had hot water 28 hours in and warm water until hours 40-48 so with the high r value of the insulation in these vessels i doubt that its actually running all the time.. i think he might expend more energy raising the temp back up in the mornings/evenings that letting well enough alone. theres a reason why this has never been done…

  2. 70 to 80 °C (158 to 176 °F): Disinfection range
    At 66 °C (151 °F): Legionellae die within 2 minutes
    At 60 °C (140 °F): Legionellae die within 32 minutes
    At 55 °C (131 °F): Legionellae die within 5 to 6 hours
    Above 50 °C (122 °F): They can survive but do not multiply
    35 to 46 °C (95 to 115 °F): Ideal growth range
    20 to 50 °C (68 to 122 °F): Legionellae growth range
    Below 20 °C (68 °F): Legionellae can survive but are dormant

  3. I don’t know about a 10/10 on the “green hack” scale, because the better solution is to just install a tankless “on demand” water heater which will result in even greater savings, especially when factoring in the nearly $100 in parts used, and electrical draw from the hardware. I doubt that the thermostat is designed to withstand constant adjustment as well. I give it a 9/10 on the “neat hack” scale though…

  4. Clever idea, but I’m with FILESPACE on this one. After taking hot showers during the NW icestorm of ’96 3 days after the onset, I think the actual savings is different than a conventional programmable furnace stat.

  5. I just looked into doing this yesterday. I was going to turn the heater off/down during night and working day hours. A little googling proves this project worthless. A guy tested his 1980 something gas 40gal water heater and it only ran about once every 35 hours when not used. This is because water heaters are very well insulated.

  6. I don’t remember the equations and am too lazy to google it. As mentioned above, the high r value of a modern tank means not much heat is lost.

    Likely the system is turning the dial often, but saving little. The thermostat itself is a control mechanism to save energy. The mod lowers the ON limit but I suspect the burner isn’t actually firing very often without it.

    I left my boiler running on its regular temperature setting for a few months while trying to sell my vacant condo. Over 3 months my energy bill was ~30 dollars total during the winter months of ’10 in NJ.

    Also, I’d worry about metal fatigue due to large frequent swings in temperature of a boiler designed to be at a constant temperature. I’ve come home to angry, wet, downstairs neighbors once and never want a repeat the experience.

  7. I agree – the lag time between turning off the dial and how long it takes to actually cool the water in the tank is pretty long. The energy MOSTLY goes into heating the water up (which takes a lot of energy – 1 calorie per gram) vs maintaining heat once it is heated (which is vastly mitigated by insulation). While obviously “instant” hot water devices are likely to be more efficient (in theory) – I sincerely doubt the savings he expects to receive materialize here. Would love a follow up if the opposite winds up being true.

  8. we have hot water on demand (no boiler) it only uses gas when the tap is turned open it ignites electricly , it allso serves as central heating

    the newest types of these machines have a heatmotor inside making electricity making it even more eficient

    anyone tried to build a solarboiler yet ?
    basicly an isolated storage vessel (old boiler)with a heatexchanger inside (copper coil)and a big solarcollector on the roof (big aluminium lid painted black with glas top ) catching the heat intoo pipes (copper pipe painted black)and pumped through the heatexchanger inside the boiler with an old central heatingpump it gets crazy hot or at least hot enough to safe gas on heating it up more , afraid of legionella run the good water through a second heatexchanger inside the boiler so its just a heatstore

    get creative and safe a forest…… or youre wallet though with the current price of copper it might get a bit expensive anyway

  9. @CutThroughStuffGuy etc

    The amount of energy lost to the surroundings is going to mainly vary according to the the temperature difference between the water in the tank and the surroundings, a greater temperature difference = greater loss of energy. The energy will ALWAYS be lost whether the heater is on OR off. Turning off the heater means that no energy is put back to replace the energy leaving so the temperature of the water drops. As the temperature of the water drops the amount of energy leaving drops (smaller temperature difference). Thus to heat the water back up will always be less than the energy that would be used to maintain the water at a set temperature. And could be a lot depending on the thermal capacity of the tank.

    My rough calculations give about 2 hours for a 50 gallon steel tank at 90C in a 20C environment to cool to about 28C. In other words this could save a lot of energy.

  10. I haven’t looked at my hot water heater recently but I did rebuild a hot tub some years ago, and the temperature sensor for the thermostat was a bulb slid into a bit of tubing that projected into the water. I’d be surprised if the HWH was different since this was partly to protect the sensor from the water pressure.

    It would be much cleaner to remove the original thermostat, slide an appropriate temperature sensor into the tube, and use an appropriately fat relay to completely replace the original thermostat.

    That said, turning the knob every day probably won’t wear it out too quickly; the knob turns a cam which alters the position of the switch contact that makes to start the heating element. But it’s a pretty clunky and expensive solution compared to just building your own electronic thermostat with an electronic temperature sensor.

  11. Note my calculations were for an uninsulated tank. I can’t remember off the top of my head how to do the transient heat conduction/convection with the insulation so yeah it may or may not have a tremendous impact. Although I can say with certainty that this hack will save energy.

  12. “a greater temperature difference = greater loss of energy.” Of course. But unless the r-value is stupid low and there is tons of drafts (there are not) and other methods of heat loss, the concept of not heating your hot water heater up because you fear greater losses doesn’t really hold water. Because you still have to heat the water up and that is where most of the energy is going. Not maintaining it once it gets hot.

    “Thus to heat the water back up will always be less than the energy that would be used to maintain the water at a set temperature.”

    Which is entirely incorrect. Heating it up takes a HUGE amount of energy. We have tanks that are hundreds of gallons and the cost and amount of energy needed to take them from 70 F to 140 F is HUGE. But once you hold them there, a simple PID can maintain them for almost no energy – only minor losses. This assumes a retarded flow of heat out of the system (insulation, etc).

    Think about this. If you heat up soup to boiling on the stove and put it in a vacuum thermos. It will stay very hot for literally an entire day or more. You spent a large amount of energy to heat it up and then it sits and coasts along for a long time without losing much energy. If you wanted to maintain the heat, you would just need to add a small amount of heat every so often. Obviously if you left it out, exposed to air currents and with no insulation, it would cool rapidly. So the biggest question is – how well insulated is the hot water system and that dictates how quickly heat is being lost. If it is insulated enough as I suspect it is, then only a small amount of heat is lost and therefore this hack, while intelligent in its theory, falls completely flat in the real world.

  13. If the tank were uninsulated then OF COURSE this would help because the lag between heating the water and then losing that heat to the surrounding environment would be very short. But that isn’t the case here.

    “Although I can say with certainty that this hack will save energy.”

    I can’t. It MIGHT help ever so slightly. But I sincerely doubt it will have a significant impact. You still have to heat the water. I would suggest focusing your energy on better payouts – like better building insulation, a white roof, energy efficient HVAC equipment and other large energy consumers (like dishwashers or clothing dryers).

  14. “But it’s a pretty clunky and expensive solution compared to just building your own electronic thermostat with an electronic temperature sensor.”

    Or buying a PID and integrating it in or better still, buying a preengineered drop in solution.

  15. @CutThroughStuffGuy

    *Palm to face*

    I don’t think you understood what I was saying.

    Lets assume that we have a water heater brought up to temperature. In order to maintain that temperature we will have to replace the the energy lost to the surroundings. The energy lost to the surroundings will be driven by the temperature difference between the tank and the room, impeded by the thermal resistance of the tank and convection to the air.

    Assuming we hold the water and room temperature constant we will have a constant rate of energy ( power) loss. If we turn off the power input via the heater the rate that energy is leaving will drop with the temperature of the water. So over time less and less power is being lost to the environment. Sure if the insulation in the tank is great the temperature could take a long time to drop, but regardless, as the temperature decreases, the total rate of energy loss decreases. When the heater is turned back on the energy that left will have to be replaced to get the water back up to temperature, however because the rate of energy loss decreases with temperature the total amount of energy that must be replaced will be LESS than the amount of energy that would have been put in to maintain a fixed temperature.

    For example:

    Say the rate we lose energy maintaining the heater at 140F is 10Btu/h. Then the total energy lost in one hour 10Btu/h * 1h = 10Btu.

    The rate of energy lost when the heater is off is a function of time but as stated above decreases with water temperature, in this case from 10Btu/h to 0 Btu/h when the water reaches room temperature. The total energy lost would be the integral of the function describing the rate of energy lost over one hour, and since the function is decreasing, the total energy loss would be less than holding it at a constant temperature. If the temperature is dropping very slowly when the heater is off, and the heater is never off long enough for the temperature to drop appreciably then there wouldn’t be much difference between the energy used when keeping the water at a fixed temperature and the energy used when turning off the heater then heating the water back up, however the energy used will be less than maintaining a fixed temperature, otherwise your breaking the first law of thermodynamics.

    The amount of energy savings is highly dependent upon the time constant of the temperature of the tank and how long it is off. However energy will be saved by turning the heater off even if it is a tiny bit.

  16. I think that the people mentioning having hot water 3 days after a power outage are forgetting that gas water heaters have a mechanical thermostat and don’t need any electric power to continue to function.
    Unless that person has an electric heater; they didn’t seem to specify.

  17. Um for the people talking about R-Values and showers 3 days after loosing power, Keep in mind that he is probably using warm water during the time he doesn’t need hot water.

    Meaning that he is replacing it with water that needs to be warmed up. If the water doesn’t need to be warmed as much then he will save money.

  18. I think some of you do not realize that this is a gas water heater. There is no electricity involved in it’s normal operation. The good news is that it works fine through power failures, the bad news is it is difficult to automate it’s operation.
    I think this is an ingenious solution!

  19. The USA seems far behind in energy efficiency for the home compared to Europe. In Europe for the home you would pretty much always get a combination boiler which gives tankless on demand hot water and central heating in one unit. Usually with a fully programmable digital timer for the central heating.

    I have noticed similar efficiency and design differences with washers and dryers.

  20. Another tip to save money on heating hot water. Drain the tank once a year. All heaters have a drain valve on the bottom where you can connect a regular garden hose. Flushing out all the gunk from a years worth of water heating can save you cost in energy from removing sediment that will deposit on the elements and from buildup destroying the tank itself.

    Takes 10-15 minutes. Turn off power/gas to water heater. Connect hose, turn valve on and let it run till the water is clear.

  21. switghton, that’s great and all, but it’s not like it would save you much overall.

    Unless :
    A) The room where the water heater is stored is really cold
    B) You like your hot water to be boiling
    C) You don’t factor the cost of the parts used in this project
    D) Your tank isn’t well insulated
    E) You think burning gas is efficient and thus no energy is lost burning it
    F) Your time is free

    Besides, you get to live with inconveniences. You don’t have a great supply of hot water unless you’re in the fixed time bracket. You put yourself in a situation where certain bacterias could proliferate.

    I really don’t think it’s worth the hassle. Numbers don’t lie!

  22. @polobunny

    Those are decent points although I disagree with E): the efficiencies of burning the gas would be present whether you are heating the water back up or using the burner every now and again to maintain a fixed temp so the efficiencies essentially factor out.

    I’m not making an argument for the use of the control necessarily (in fact the bacterial scare me enough that I wouldn’t attempt it), I’m simply arguing against the fallacy that the thermal capacitance of the tank somehow makes maintaining a fixed temperature use less energy that turning the burner off then back on.

  23. @polobunny

    Also none of us can state whether it would save much overall – you are correct that numbers do not lie, but without running the numbers we can draw no conclusions as to what the savings actually are, other than hand waving and intuition that is sometimes very right, but also sometimes very wrong.

    We can however use scientific principals to discuss the overall behavior of the system.

  24. @swighton – I was just going to tell him to take actual measurements (empirical data beats the hell out of theoretical argument), but your explanation was beautifully logical. My physics prof would have cried at the elegance of it. :)

  25. I am near the gas powered water heater in my basement most days and sometimes late into the night. The only time I ever hear it turn on is when someone is showering or drawing a bath.

  26. @swighton, you are right about E), my mistake there. It’s only logical that it cancels itself out.

    Otherwise you are also right about everything being only speculation. Nevertheless, water tanks are quite efficient at what they do and have been for many years. Let’s just say it’s not an ever evolving market. :P

    I would like to see some numbers just for kicks and I’m sure others would enjoy too.

  27. “However energy will be saved by turning the heater off even if it is a tiny bit.”

    My point was basically that the savings was so minor that it wasn’t worth bothering with because other things that could be done would result in much better savings.

  28. “The good news is that it works fine through power failures, the bad news is it is difficult to automate it’s operation.”

    It also means that it is less efficient but also cheaper at the same time because natural gas is cheaper than electricity for the same BTU (as far as I know, everywhere in the USA). That probably does not hold true in every other part of the world. Propane is close to a wash with electricity in the USA.

  29. “Then the total energy lost in one hour 10Btu/h * 1h = 10Btu.”

    10 BTU is *nothing*. It takes 12,000 BTU to equal one ton and most houses have 2 – 6 tons of cooling or heating capacity. It would take 7,200 hours to lose that heat. That is 300 days!

  30. “The amount of energy savings is highly dependent upon the time constant of the temperature of the tank and how long it is off.”

    The savings depends on quite a few factors. But if we assume the tank to be fairly well insulated (many are only insulated to a modest degree) and that the room the heater is in to be a reasonable temperature (not 40 degrees F or something) then all else equal, the variation in temperatures of the tank *should* help save a small, small amount. But much of the savings is negated by the fact that you still have to heat the water. It isn’t like you heat the water and then presto – it cools instantly so if you turn it off you can “avoid” heating water that you aren’t actually using. The very nature of a tank means you essentially have a capacitor from which to draw water from. That capacitor leaks slowly of course (self discharges) but for the most part – it stays pretty constant. So again – this hack *MIGHT* save a few pennies at the expense and hassle of colder hot water and other aggravations (what if you want a hot shower at 3 am every so often? – what about the wear and tear on the components? – what happens if the popsicle stick contraption breaks and causes the tank to only output scalding hot water – what happens if it turns off and fails off? You get the idea).

  31. “10 Btu/hr”

    “It takes 12,000 BTU to equal 1 ton…”

    In case you were wondering, that’s 9,338,031.15 foot-pounds.

    We have a perfectly good unit for energy. It is the joule (J). We also have one for power, called the watt (W). Use them.

  32. “We have a perfectly good unit for energy. It is the joule (J). We also have one for power, called the watt (W). Use them.”

    Yeah… you’re in the wrong industry. Feel free to play with wikipedia and come back to pontificate as an expert.

  33. do any of you actually know how a hot water tank works? mine is gas and set to about 20% past “warm” towards “hot”. it is still hot. been like that for the past 10 years and i dont have legionnaires yet and zero problems with heat in a shower, unless the washer is running at the same time it cant keep up for long. all he is doing is turning it from the “hot” to “warm” setting, not off. meaning hes pumping more gas in when the extra volume is needed but keeping it low when its not. there is probably no realistic gains in this set up but it could be worth it and it is not going to overheat or underheat if his system fails.

  34. to extrapolate on my previous comment (yes i know that word doesn’t exactly fit but i try to use it as much as i can) when running hot water there is rarely a time you run hot only full bore and need it. meaning when my hot water tank in set on “warm” i have to use more of the hot and less of the cold for a comfortable shower or whatever im doing. it is still hot when running only the hot tap. when it’s set on the “hot” setting it is hotter then it has to be and requires less of a hot:cold ratio to be comfortable and therefore you can do more things per tank of water at your comfortable level. im not saying this will save him anything, just if he doesn’t need that added volume why keep the temp up?

  35. @swighton,

    i’m sorry, but the recommended best practices for efficiency in hot water heat systems is static temp maintenance. in those systems you bleed heat via radiators into the house, and in this you bleed heat via hot water down the drain. both of those are intermittent as exterior temps change and water demands shift.

    it takes more energy to REheat the system back up to operational levels than it does to maintain the level. there is very little heat loss as modern tank hot water heaters are quite well insulated. (yes, on demand or combo heaters are more efficient for most residential uses)

    this is NOT the same as forced air furnaces, in which the entire house becomes the thermal mass.

    the reason to adjust a tank water heater is if you have occasional limited need for more than one tank of hotter water (you run a commercial dishwasher at church once a week, and only need hot water to for the toilets most of the time) otherwise, while counter intuitive, it’s more efficient to keep the temp stable.

  36. Oh a hot topic!
    1 insulate
    2 insulate
    3 insulate all hot water pipes to taps.
    In the 70’s jackets were common as add on’s. R ratings have improved some, but a 4 foot wide heater ‘won’t sell’, so we can’t have them. Gas heaters should have a auto damper, as draught is cooling the heater during off time
    Be thankful for hot water. Use it wisely, and thanks be that you don’t have a thankless water heater that can not provide a trickle of conserving hot water. If demand heaters had a very small tank to buffer the `trickle in the shower when shampooing’ problem ,that would pass. When you resume it’s severe temp swings, ouch.
    Hands free faucets would save much more, but insulate pipes first right up to tap. It ought to be law-code.
    I lived in a custom made home with 30 feet of run in a cold crawl space. 2 gallons to run for hot, 10 minutes later do it again, after a wall batt jacket hot water was on tap an hour later. That’s savings! Better, it’s convenient. This should drive design in green motives.

  37. Pete – yep, I too am rather amazed that some places are still so far behind in heating technology/wasting energy. Combi boilers (hot water/heating on demand) have been around for the best part of 20 years in the UK, condensing boilers for 10+ IIRC at 92+% efficiency. I dread to think how much energy is being wasted in system like this.

  38. First interesting hack. Personally I don’t see much in the way of savings to be had here, but interesting solution to the perceived problem.


    I am not going to say who is more efficient, but I would like to point out that climate has a lot to do with the type(s) of HVAC used. For instance, here in the states boilers are frequently used in the North-East, While forced air systems are used elsewhere to reduce the need for separate Heating and Cooling Systems. In the South many homes have no heating systems at all.

    As for the Gas water heater, I would not waste my efforts, they already very efficient, in the summer my gas water heater costs me less than $20/Month and my gas dryer costs me about $5 – 10 a month. It is VERY rare for my gas water heater to come on unless I draw some hot water (thus adding cold water to the tank). I would add additional insulation if I were concerned about heat loss, passive and very effective. Oh and guess what there is a minimum $20/Month on my gas bill anyway, so I may as well use at least $20 /Mo.

    Those tankless heaters, won’t even pay for themselves in their lifetimes when you factor in additional costs, maintenance… A tankless heater costs 2 – 3 times as much as a standard water heater and in order to achieve their stated efficiency you have to use water at the rate they were designed for. If you use more than their rate you get cooler water; Less and the energy is wasted. Better not try to take 2 showers at once (unless it is designed for that rate of draw) both may not get the hot water they want. Many of the tankless heaters now in fact include a small tank for retaining some hot water, Hmmmm…

    Of all the waste in the house, a water heater (especially a gas water heater) is probably about the last thing I would look at. Turning off an unused computer will save you $20 a month on your electric bill. Unplugging wall warts when not in use will save a lot too. What about that TV no one is watching, TVs are HUGE power wasters, some even use nearly as much power when ‘off’ as when on (mostly CRT models), this is due to our ‘instant on society’. In fact for most electronics, off is not off. If you are interested in energy conservation start where is counts.

    For me:
    Summer gas bill $180/Mo
    Winter Gas Bill >$150/Mo
    Winter Electric Bill < $90

    I pay more for my cable TV and internet than I pay for electric or gas. Same for my cell phones bill.

  39. Well those bills got screwed up somehow


    Gas less then $25 a month
    Electric more than $180 a month

    Gas more than $150 a month
    Electric less than $90 a month.

    So I will concentrate on electricity conservation.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.