I have a 120 liter water heater powered by a 4kW heating element. I own two 350W solar panels totalling ~700W out of which I expect to reliably draw ~500W of power. I am looking for a way to take the 500W at 24V from the solar panels, raise the voltage with a DC-to-DC converter, store the energy in a capacitor and discharge at the rated power on the heating element. So for every 8 seconds of charging, the heating element would work for 1 second.

Is anyone aware of any off-the-shelf module that would do this sort of thing? What are my options to achieving this?

  • 4
    \$\begingroup\$ Why do you want to have an 8/1 second cycle? Or is that just an example? You could just power the heater with a DC to DC converter without storing power for so long. \$\endgroup\$ Commented Aug 26, 2021 at 17:51
  • 3
    \$\begingroup\$ A capacitor large enough to do any good will be much bigger and much more expensive than a battery. \$\endgroup\$
    – JRE
    Commented Aug 26, 2021 at 18:06
  • 10
    \$\begingroup\$ If all you want is hot water, look into a solar collector instead of photovoltaic cells. \$\endgroup\$
    – JRE
    Commented Aug 26, 2021 at 18:08
  • 3
    \$\begingroup\$ Solar water heating is more effective than pv to heat when you want hot water. \$\endgroup\$
    – Solar Mike
    Commented Aug 26, 2021 at 18:19
  • 3
    \$\begingroup\$ @PaulB. You don't need to modify your heating element. See Neil_UK's answer. But if you want to have a duty cycle you could use 40 uSec cycle with 5 uSec on time, and get the same result but with a capacitor 1/200,000 the size. \$\endgroup\$ Commented Aug 26, 2021 at 18:47

4 Answers 4


You have a heating element of 4 kW at 230 V:

$$R=\dfrac{230^2}{4000}=13.2 \Omega$$ This heater would draw 500 W of power when voltage is \$V=\sqrt{P*R}\approx\sqrt{500*13.2}=81V\$

So you do need a boost converter. But you could also wire two panels in series.

Example panel

Table of example panel specifications

Then you have a MPPT voltage at 38.5 X 2 = 77V, which is very close to your needs.

MPPT Voltage VS. Irradiance

MPPT curve of example panel

You can see that MPPT voltage doesn't change drastically. If your panels have similar characteristics, then you can simply connect them on the heater and you'll get the max. performance, because no extra converter losses would be involved.

  • \$\begingroup\$ Thanks. This is what I have been looking for. \$\endgroup\$
    – pbount
    Commented Aug 26, 2021 at 21:22
  • \$\begingroup\$ This correlates very well with what this guy did, maximum power point but without tracking. YouTube link: youtu.be/KVeGummoXS0 \$\endgroup\$
    – winny
    Commented Aug 27, 2021 at 7:13
  • \$\begingroup\$ But also be aware that most heating elements have a positive temperature coefficient of resistance. This means that, for a lower voltage than would produce 4kW, the element will have a lower resistance than you expect. So, your voltage across the heater will be a bit lower than you expect. \$\endgroup\$ Commented Aug 27, 2021 at 19:57

A 4 kW heating element doesn't have to work at 4 kW. It will produce 500 W average heat output whether it's driven at 100% of the time with 500 W, or for 12.5% of the time with 4 kW.

Use a DC to DC converter to produce the right voltage to drive the element at 500 W. You do not need to store energy in batteries or capacitors.

Better, if your panel controller MPPT will work with a resistive load, configure that to drive the heater.

Better still, absorb the sunlight directly into water heating tubes, you will get much higher water-heating efficiency than going via electricity.


Frame Challenge

You are working on a complicated way to:

  • Convert solar radiation (aka sunlight) to electricity, typically at a rate of around 15% efficiency
  • Use the electricity to heat water at a low rate (i.e., much lower than the design capacity of the water heater)

Now if all this was essentially "free", it would be well worth it. But we're talking inverters, converters, batteries, capacitors, etc. Not free by any stretch. Plus the electricity generated is not available to run other things that require electricity - computers and other consumer electronics, fans, lights, etc.

On the other hand, if you use the solar radiation to directly heat the water, it is far more efficient. This does require some additional equipment - the solar collector and piping from/to the water heater. But it works. In fact, Israel (and I think some other countries too) use this for residential hot water on a regular basis. Then your solar panels are free to provide electricity for everything else.

This problem is not unique to solar energy. For example, "ordinary" electric heat (aka "giant toaster") is 100% efficient at converting electricity to heat, but the electricity is, in many places, generated very inefficiently from fossil fuels. In that situation, using natural gas to directly generate heat is much more efficient. (Leaving out the "heat pump" issue for the moment - that changes the equation in other ways.) This is especially the case if you use a generator for backup power - a home generator will generally be less efficient overall than a utility-scale generator, so using a generator to run electric heat will go through your gas a lot faster than the same gas burned to generate heat (with electricity just used for fan, controls, etc.).

  • \$\begingroup\$ Solar hydronic heating is also common where I live in Colorado, USA. Although, PV powered heat pumps seem to be more popular for new construction. \$\endgroup\$
    – Phil Frost
    Commented Aug 27, 2021 at 19:43
  • \$\begingroup\$ You are stating the obvious here. A solar thermal collector is rated at 1.7-2.5kw for the same area that is rated 250-300w for a solar panel. Obviously, a ~2/10 collection rate is expected. Selling points are, cost is low, installation is there, the system is hybrid so on days with little-no sunlight i can still swap to electric with the flip of a switch, the boiler is inside where its already warm so no extra losses on cold days. \$\endgroup\$
    – pbount
    Commented Aug 29, 2021 at 9:05

PSA: High voltage DC is dangerous

OK first, 240V >>DC<< is nothing to fool around with. The nature of DC power changes dramatically at such high voltages. It is much more dangerous than the equivalent AC power. Just look at power handling equipment - you will see the DC rating is much lower than the AC rating. 240 volt AC equipment is found at any home store. 240V DC equipment is either expensive industrial/specialty, or cheap dangerous Chinesium.

So you should be handling any power above about 50V as AC, not DC. And you should still use all the precautions required by your country's or modern electrical codes for that AC power.

So how about this: use a 120V AC inverter. And a diode.

Now, water heaters are simple resistors. With resistor heaters, if you halve the voltage, you quarter the power. It's designed for 4000W @ 240VAC, but at 120VAC, it's only 1000W.

120VAC inverters are readily available.

Now, what happens if you run 120VAC through a diode? You get half power, since only half the sinewave is actually coming through. Now we are at 500W @ 120V.

"That was easy" :)

Now all you need is a 120V inverter that responds to low voltage by shutting off for 8 seconds.

  • \$\begingroup\$ How does a typical inverter respond to being half-wave-rectified at rated power? Low power is probably fine, but I would guess that it's still more stressful than simply a lower resistance. Do that at rated power, and I don't think it would like you at all! \$\endgroup\$
    – AaronD
    Commented Aug 27, 2021 at 16:22
  • \$\begingroup\$ A diode to a resistor is the most non-inductive load imaginable, so everything with it happens in real time. So how would the inverter even know? It doesn't have a memory, not even a 10ms one. It's simply seeing the load change, which an inverter better be able to handle, but again, no memory, so no concept of "change". @AaronD As for sizing, that's a product selection issue. I wouldn't expect anything but magic smoke from a Chinesium inverter run at rated power to any load. \$\endgroup\$ Commented Aug 27, 2021 at 17:36
  • 1
    \$\begingroup\$ Having many years ago taken a foolish shock from a beefy 200VDC lab supply, I can concur that serious DC electrocution is scary as hell - a totally different beast than AC! \$\endgroup\$
    – J...
    Commented Aug 27, 2021 at 19:08
  • 1
    \$\begingroup\$ what really throws people is how docile and well-behaved 5V and 12VDC is, it's like a labrador retriever of voltages. They don't appreciate how sharply DC turns into a rabid wolverine at higher voltages. Once a DC arc starts, it cannot be bargained with, it cannot be reasoned with, it does not feel pity or remorse or fear... \$\endgroup\$ Commented Aug 27, 2021 at 19:19
  • \$\begingroup\$ @Harper-ReinstateMonica Output transformer??? A (very) naive design might use a 12V, 50 or 60Hz oscillator to drive a 1:10 60Hz tranny. That has memory, in the form of a massive DC offset on the secondary. And who knows what someone else might decide to do inside of their "black box". Some would be just fine, others probably not. \$\endgroup\$
    – AaronD
    Commented Aug 27, 2021 at 20:26

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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