PV to DC Water Heater Elements

[OnIsland.OffGrid]

Good Morning, new forum member but I've done several off-grid solar systems on campers and RV's.

I'm getting mixed answers both in researching other threads on here and from solar panel experts. Setting up a hot water heating system in Hawaii, where freezing temperatures are not a factor but the degradation of a thermal solar tank on the roof doesn't last is, and is too much weight for the area we want to install it. We've resolved to use a regular Rheem hot water heater but swapping out the AC elements with Dernord 12v 600w DC elements on Amazon. Two of these powered directly by two 400w panels on the roof (800w total to each element, 4 panels total) with 50a breakers on each element circuit.

We only want water to heat when the sun is out and are trying to avoid using charge controllers, inverters, or storage.

My question is two parts:
a). Are we adequately matching panels to the resistance of the elements?
b). How can we thermostatically switch these without arcing?

My first thought would be the use of heavy duty solenoids wired to a thermostat, maybe with an intermediate relay if the switching power of the thermostat is too low.
Two of these: https://a.co/d/6I7WmmL
And one of these: https://a.co/d/iC9VX6m

Any thoughts or advice are much appreciated.

 

[efficientPV]

There is lots of bad advice. Avoid low voltage elements because they lose too much power in the wires. Put your panels in series for the highest voltage. Use a solid state relay, but use two in parallel of at least 40A even though you will be nowhere near that. Match your heating element to real world conditions. There is a big penalty for going under resistance. Very little for going much over in resistance. For HI I would calculate resistance at 70% of panels Mpp current. You can likely use the existing heating elements, possibly in parallel.

 

[OnIsland.OffGrid]

There is lots of bad advice. Avoid low voltage elements because they lose too much power in the wires. Put your panels in series for the highest voltage. Use a solid state relay, but use two in parallel of at least 40A even though you will be nowhere near that. Match your heating element to real world conditions. There is a big penalty for going under resistance. Very little for going much over in resistance. For HI I would calculate resistance at 70% of panels Mpp current. You can likely use the existing heating elements, possibly in parallel.

Thanks, luckily the wire run is very short (less than 10ft). I'll be using heavy gauge wiring as well, with two 50A (600w ÷ 12v) breakers for each element circuit. Do you have any recommendations for a solid state relay? Here on island Amazon is my best bet for timely shipping unless we can find it at a local store. Also, it's my understanding (perhaps the bad advice you were referring to) that AC elements suffer from electrolysis when used with DC.

 

[yabert]

Good advices from efficientPV.
Except that SSR in parallel isn't a good idea. When one open, the second one will have to open 2 time de load fraction of second later.
If you stick with 120V ac heater and run them at 80-90V DC from panel in series that can be way better than those 12V 600W heater who see to only give 1/4 of their rated power.
Also, why do you think that water will reach boiling temperature from two 400W panels? Maybe you don't need thermostat.

 

[AntronX]

a). Are we adequately matching panels to the resistance of the elements?

No you are way off if panels are 60 cell equivalent.

600w / 12v = 50A
12V / 50A = 0.24 ohm
(400W * 0.85) / 32V = 10.6A
32V / 10.6A = 3 ohm
3 ohm / 2 panels in parallel = 1.5 ohm versus 0.24 ohm heater resistance. Huge mismatch.

Assuming you have 2 panels in parallel that's 21 amps total. So if your panels are connected to one 600w 12V element, then they are only making roughly 21A * 0.24 = 5 volts * 21A = 105 watts vs. about 680 watts they should be making in full sun.

 

[efficientPV]

I test lots of amazon relays. You never quite know what you will get even from the same seller. Mostly this relates to on/off speed which isn't a factor for you. Typical 40A 200V SSR will be a FET and two in parallel will reduce the total amount they heat up to 1/4 of a single SSR. Heat is the real killer of these relays. These SSR will fail if the control voltage drops below the minimum voltage. They should have a solid power source. Many heating elements are totally encased with a metal shield which is connected to the case, so no electrolysis. Some, the dark grey ones, are electrically isolated from the mounting base. UL doesn't seem to think there is a leakage problem with these. I suppose it could be possible for some. If the panels are floating and not grounded there isn't an electrical path. Provide the actual nameplate data of the panels and how many will be put in series.

 

[AntronX]

4600 watt 240V heater element has 12.5 ohm resistance. Each of your 400w panels has 3 ohm ideal load resistance. Connecting 4 panels in series to one heater makes perfect resistance match: 12.5 vs 12 ohm. You should put back 240V elements and wire panels via a PWM controller like @efficientPV or I have made to keep panels at optimal voltage regardless of illumination level.

 

[efficientPV]

In direct connect, everyone calculates for the ideal resistance from the Impp of the panels. This is true for about one hour of the day in ideal conditions. It discounts production for the rest of the day and less than ideal cloud conditions. The chart below is daily production comparing resistance. For this array calculated ideal resistance was 28.8 ohms. Daily production actually increased when the heater resistance almost doubled. That is because power is a factor of the current squared into a resistance. Small drops in current have a dramatic effect on power. Most of the day the panels will be off angle and producing lower current.

 

[OnIsland.OffGrid]

Glad I consulted you guys first. I haven't purchased the panels yet and thanks to Amazon I can return the DC elements if needed. So electrolysis on AC elements isn't an issue? It sounds like wiring the panels in series to a single AC element would be more efficient. I don't think we have the surface area on the roof for more than four panels...

Still trying to do this in the simplest way without any charge controllers. We were going to use Rheem's heat pump unit (lots of ambient heat availabe) but I can't get an answer from Rheem as to the energy demand in just Eco (heat pump) mode.

 

[AntronX]

For this array calculated ideal resistance was 28.8 ohms.

It's easy to do constant voltage PWM and pretty much get MPPT level of solar harvest. But if doing direct attach i guess connecting 2 x 240v elements in series to 4 x panels in series would be most efficient.

 

[AntronX]

So electrolysis on AC elements isn't an issue?

I don't think galvanic or electrolytic corrosion has anything to do with AC or DC going through the heater element.

 

[OnIsland.OffGrid]

OK, so let's say I can get a hold of four 400-watt panels. I'm still waiting to hear back from the local island company on their STC, Voc, etc and I'll post them as soon as I get the specs.

You're saying I should be ok to run them in series, through an appropriately-rated circuit breaker, and then through the elements in series? Why do companies like Dernord make those DC-specific elements?

I've done quite a few setups using charge controllers and batteries, including sizing them for the demands of off-grid setups, but the resistance of a heating element is a new ballgame.

 

[AntronX]

Why do companies like Dernord make those DC-specific elements?

Probably to run from low voltage battery.

 

[v_green57]

Somewhere on the forum recently was a DIY project for PWM'ing for resistive water heating.

Antron, was that you?

It was pretty comprehensive.

 

[AntronX]

Antron, was that you?

From what I know it was efficientPV and recently my own attempt at direct PV to water heater element via PWM. It works great but special care must be taken to tame massive RFI it can generate (if you do ham radio).

 

[efficientPV]

The majority of the PWM designs use high frequency > 10KHZ. I use a maximum on time of 10ms which matches the frequency of 50 Hz compatible with mechanical thermostat switches. In this way there is an arc interrupt which turns off the current for 0.6ms. This allows heating elements to be daisy chained as in the typical water heater. Since the rate is so low, The FET can be slowed down and reduce EMI without having any transition heating in the FET. Doing this requires a larger capacitor bank which is slightly more expensive.

It is surprising that the power point controllers aren't more common. The added performance is amazing outperforming direct connect designs by two or three times in less ideal sun conditions. My domestic water heaters turn off by 10am so I can run the dishwasher. Having panels only for a water heater can be a great waste of energy. A power point controller can run in parallel with a conventional MPPT charge controller to harvest excess energy. Extra panels often aren't needed because any successful PV system has to have excess capacity.

 

[CanadianRube]

Great thread, thx guys. I have https://www.aimscorp.net/product/10...36-48-vdc-mppt-etl-listed-to-ul-458-csa-22-2/ as charge controller on a 3200W system.

It has a dump load output. Would connecting this to one or two 48V DC water heater element(s) work well or fall short in context of this discussion? What would I need to add if anything? I was thinking to have two typical 40gal tank style water heaters in series. First tank has the solar heater dump load to preheat with any extra solar collected. Second tank runs like normal on fossil fuels as needed but at a reduced cost based on amount of preheating accomplished so I’m never without hot water or having to think about it.

Things I’m not 100% sure of.
I’d get the advantages of the mppt even on the dump output?
The controller only exports power via the dump when batteries are fully charged, so they should get time to balance?
I need a thermostat to be sure I don’t overheat the water?
I want a set it and forget it. Don’t need anything fancy.

Hopefully this is relevant and helpful for anyone reading the thread in the future. I had to google a number of acronyms just to follow this.

 

[OnIsland.OffGrid]

OK, I have four of these panels and one 4500w water heater. How can I make this work in the simplest way possible? The upper and lower elements of this heater are in parallel. Should I tie four panels in series and swap the lower element out for a 1600 watt element, if I can find one? I wish it was that simple but the resistance is coming into play.

 

[OnIsland.OffGrid]

I'm thinking I need to swap in one of these 48v 1500w elements to better match the current.

https://a.co/d/i0wXJFH

 

[TM48]

It's not so simple. The AC thermostats cannot be used with DC. Arcing and possible fusing the contacts make it hazardous. Only one element is on at any time so most people just reduce the element wattage and run them off an inverter. I changed my 5500w elements for 3500w.

 

[efficientPV]

It is that simple.

Simple seldom works that good.

You were already given the answer.

Show some calculations and a diagram.

NO math = NO answer.

In HI even a brain dead design should work somewhat.

Do what everyone else does here. Throw a bunch of stuff together and hope for the best.

In solar the only performance goal is. Works better than nothing.

 

[AntronX]

You were told what to do, why do you cling to this 48v idea? Connect 2 existing 240v heating elements in series directly to 4x panels in series. That should give you 1200 watts of heating power in direct sun. Make sure to disconnect or bypass thermostat switches or they will arc and burn up when they open the circuit and that could start a fire.

 

[OnIsland.OffGrid]

Pardon me, I'm trying to understand the impedance (including an element in the mix) of using the panels in series vs running the panels in parallel. I'm worried about the safety and efficiency of the system. But I think you're right- I'm thinking too much about this. My math took me down this rabbit hole:

Vmpp is 42.1v
Impp is 9.51 amps
With 4 panels in series I have 168.4v total Vmpp. That makes my total Impp still 9.51amps, right? So MPP resistance should be around 17.7 ohms. And so my guess (and per your suggestion) is that a lower wattage element (like 2500w) to series-connected panels would be better than running in paralell with a 48v 1500w element, like the one I posted from Amazon.

I'm bypassing the AC thermostats because of the risk of arcing. Correct me if I'm wrong, but that poses a safety risk if the system works TOO well (fat chance, it seems). So all I can think of is to make sure the pressure relief valve is working and piped outdoors. I'm also toying with the idea of using a local 120v outlet to power the thermostat which would then control a DC relay to switch the PV supply on and off, but maybe that's just another rabbit hole. Either way, thanks for your help. I'll tie 4 panels in series to 2 4500w elements in series and see how it goes.

 

[AntronX]

You can use DC contactor controlled by the thermostats. You may need thermostatic mixing valve to regulate hot water outlet temperature to prevent scalding.

 

[efficientPV]

It should work much better than you expect as it is a pretty ideal match. With dramatic temperature swings of the tank an expansion tank is necessary. Should likely trip the T&P valve from pressure without a thermostat or blow thru valves in the house. A solid state relay i in your future.