At our off-grid cabin, I want to power a 240V 3 kW residential hot water cylinder more or less directly from a 5 kW solar array (when the sun is shining). The panels are not required to do anything other than heat the water. If the sun is not shining then we have no hot water - that's OK.

I want to use as few as possible parts, except for the panels themselves and the hot water cylinder.

Looking for design recommendations.

Here are some solutions I have rejected.

Solution 1 (the "chuck the kitchen sink at it" solution):

Add to the above some kind of circuitry to turn off the inverter when the sun is not shining (we never want to actually dip into the ultra light-weight battery bank).

I believe this would work, but would be hideously expensive. Also when the power from the sun drops below the 3 kW drawn by the water heater, the inverter needs to be shut down - whatever power is available is wasted.

Solution 2 (the "AC? who needs stinking AC?" solution):

  • Connect the panels in a series/parallel configuration, with a blocking diode in series with each string, such that the peak voltage from the array is < 240V (the rating of the water heater).
  • Connect the resulting DC directly into the heater element of the water heater.

I like this solution from the point of simplicity, and virtually zero parts cost. Also it makes best use of the available solar energy, and will work even when the sun is low, by just pumping a lower DC voltage into the heater.

However I am concerned about damage to the heater's element due to electrolysis, from running it on DC instead of the AC that it is designed for.

Solution 3 (the "dunno quite how" solution):

  • Some kind of DC to AC chopper that can turn the DC from the panels into AC, at <= 240V.

I don't believe the heater will suffer if it is driven from a lower voltage than it is designed for. This solution also makes maximum use of the available solar energy.

However I'm not familiar with building this kind of circuitry, or whether this is in fact a good idea or not.


closed as too broad by Brian Carlton, PeterJ, Voltage Spike, Dmitry Grigoryev, Dave Tweed Sep 30 '17 at 11:54

Please edit the question to limit it to a specific problem with enough detail to identify an adequate answer. Avoid asking multiple distinct questions at once. See the How to Ask page for help clarifying this question. If this question can be reworded to fit the rules in the help center, please edit the question.

  • 13
    \$\begingroup\$ Why not use the sun directly to warm your water, that would be much more efficient and you'd be surprised about the water temperature. en.wikipedia.org/wiki/Solar_thermal_collector Cheaper and more robust. \$\endgroup\$ – jippie Sep 15 '15 at 6:14
  • \$\begingroup\$ presolarnet.com/products/liberty_box.htm \$\endgroup\$ – Bruce Abbott Sep 15 '15 at 6:49
  • 4
    \$\begingroup\$ Agree with above comment: it will definitely be much, much more efficient to turn sun directly into heat rather than sun to electricity and electricity into heat. You can build a solar hot water heater super cheap and they work amazingly well. \$\endgroup\$ – bigjosh Sep 15 '15 at 14:10
  • 5
    \$\begingroup\$ Also, solar water heater panels will be perhaps 3-5 times smaller area and many times cheaper than the same power of PV panels. And if the water tank is installed above the panels, the circulation is driven by convection so there are no moving parts. If you must use PV panels, use them to power a heat pump to heat the water, not a resistive heater. \$\endgroup\$ – tomnexus Sep 15 '15 at 21:27
  • \$\begingroup\$ Thanks for the comments above re using solar water heating, but I have access to the panels, hence the electrical approach. \$\endgroup\$ – abraae Sep 15 '15 at 22:27

Thermal panels would be preferale, but since you already have the panels...

If you're looking at a DIY solution, the simplest would be this:

Since the peak of a 240VAC sine is 336V, the heating element's insulation should withstand 336V. Thus you have some flexibility in wiring, as long as you keep the panels' output voltage between, say 200V and 336V.

Next, we need a MPPT, because wiring the resistor directly would produce almost no power in overcast weather, due to the solar panels' I-V characteristic which is more or less a voltage-limited current source, whose current depends on light.

So, I would simply put a capacitor at the output of the panels. It should have a ripple current rating high enough for the power considered here, so something like 15A should do the trick, for example a bunch of big 450V caps in parallel, maybe 1000µF capacitance, value isn't critical.

This is obviously a "kill yourself" project so you have been warned. The panels will also kill you.

Now we have the panels charging the caps, it is a simple matter to do a crude MPPT with a dumb hysteresis comparator: when the voltage on the caps exceeds, say, 250V, turn on the heating with an IGBT or a big MOSFET. When the voltage drops below say, 230V, turn it off.

The idea here is to PWM the load to make it draw an average current that allows the max amount of power to be extracted from the panels. Too much current, and the voltage would drop. The capacitors smooth the current and do the averaging.

Of course you can do a fancy MPPT with a micro if you want to have fun, but a simple bang-bang comparator should work just fine and you'll get hot water on overcast days, it'll just take longer.


You've got me a little worried here:

Hopefully the boiler's heating element should look like that:

enter image description here

This is a resistor, inside a high-temperature insulator, inside a metal pipe which is then earthed. Thus no part of the resistor which is at mains voltage ever comes into contact with the water. Thus no electrolysis takes place...

If the pipe is punctured by corrosion then water will get in and current will flow to Protection Earth, tripping the RCD.

If the mains carrying bits of your heating element are in direct contact with water, this is extremely dangerous. This shouldn't be the case in a boiler, it would be totally insane...

Which reminds me I forgot to recommend tell you to add a RCD to your installation. A quick bit of googling revealed that I have no idea if you can find a RCD which would work on DC.

So... you have another problem now, which is how to make it safe. Without some kind of protection, if the heating resistor gets corroded you'll have a dangerous situation. Since I'm a landlord as a side job, over the years I've seen this happen once, although the most likely scenario is that the tank itself gets corroded and punctured, and then the water leaks onto the wires, thermostat, etc. In this case the RCD also trips and the tenant calls wondering why they have no power...

I'm not familiar with solar, so I guess the inverters must incorporate such a feature... but if you want to go without an inverter then I guess you should ask another question about how to make this safe :)

  • \$\begingroup\$ Very insightful thank you. Do you have any opinion on potential for damage to the heater's element due to electrolysis, from running it on DC instead of the AC that it is designed for? I have to imagine this is unlikely in any reasonable timeframe, and even if it happens then just replacing the element should get things going again. \$\endgroup\$ – abraae Sep 28 '17 at 22:50
  • \$\begingroup\$ @abraae see my edit, you just made me realize the thing is actually rather risky... \$\endgroup\$ – peufeu Sep 29 '17 at 0:31
  • \$\begingroup\$ Thanks for the update! Regarding the RCD, my thinking was that the whole system would be unearthed, so no RCD needed for safety, just a DC fuse. Unless I am missing something (very possible). \$\endgroup\$ – abraae Oct 1 '17 at 20:30
  • \$\begingroup\$ I should add that since I posted this question other demands on my time have mounted and I doubt I'll get the time to work on this any time soon, so this discussion is really just for interest. \$\endgroup\$ – abraae Oct 1 '17 at 20:49

Not the answer you're looking for? Browse other questions tagged or ask your own question.