PV Hot Water Heating--Why Not Skip The Charge Controller?

[aaron_c]

I'm wondering about hooking up a PV panel directly to a water heating element. I'm 99.9% sure that this is a dumb idea and I've been told that it will often be inefficient, however I'd like to understand why that is.

PV and Direct Heating Of Hot Water

I've seen some folks talk about using PV panels to heat hot water semi-directly, without the need for batteries, inverters, or any kind of grid tie. What DOES seem to be required is a specialized solar charge controller that knows the voltage the resistance element prefers and will give the resistance element exactly that. However charge controllers seem to crap out sooner than panels do (I've seen people say anywhere between 2-8+ years but yall have more experience than me so...maybe I'm wrong). I don't know what the lifespan or cost of these specialized charge controllers are, but the epever 3000 W MPPT controller I looked at was maybe $240 plus labor to install. I'd need maybe 3.5 of those for my theoretical 9.4 kW array. If it craps out every 5 years that's maybe $900 every five years to replace it. Given that the panels themselves might cost all of $9,500, the 30-year cost of the MPPT controllers looks non-negligible.

PV Direct Water Heating Without MPPT

This has me wondering about how efficient PV panels would be if I hooked them directly to a heating element without the charge controller sitting in the middle. In the post I linked above the author provides a good explanation of what happens if you size the resistance element for the load the panel would provide with full sunshine vs what happens if you size the load for half sunshine. Basically, the lower sunshine situation calls for a higher resistance circuit to fully utilize the lower power and slightly lower voltage.

The Dumb Question

So my question is...what happens if increase the size of the resistor so that the element makes full use of the element on a cloudy day? The resistor would be larger (have a larger impedance) than would be ideal when the panel was in full sun. My question is: What would the effect of this be?

The total power going through the resistor doesn't go up, so I don't *think* it would burn the resistor up. Right? Would it just throttle the current in a way that would reduce the amount of power it would convert to heat? My simple understanding of these things is that any power I put into a heating element should come out as heat, and the limiting factor should be whether the current and voltage wind up melting the element. So I look at this and think that "well it won't melt the element if I keep the power under the listed threshold, so what's the problem, how does it lose efficiency?"

Thanks for the help

 

[Bobert]

I

I'm wondering about hooking up a PV panel directly to a water heating element. I'm 99.9% sure that this is a dumb idea and I've been told that it will often be inefficient, however I'd like to understand why that is.

PV and Direct Heating Of Hot Water

I've seen some folks talk about using PV panels to heat hot water semi-directly, without the need for batteries, inverters, or any kind of grid tie. What DOES seem to be required is a specialized solar charge controller that knows the voltage the resistance element prefers and will give the resistance element exactly that. However charge controllers seem to crap out sooner than panels do (I've seen people say anywhere between 2-8+ years but yall have more experience than me so...maybe I'm wrong). I don't know what the lifespan or cost of these specialized charge controllers are, but the epever 3000 W MPPT controller I looked at was maybe $240 plus labor to install. I'd need maybe 3.5 of those for my theoretical 9.4 kW array. If it craps out every 5 years that's maybe $900 every five years to replace it. Given that the panels themselves might cost all of $9,500, the 30-year cost of the MPPT controllers looks non-negligible.

PV Direct Water Heating Without MPPT

This has me wondering about how efficient PV panels would be if I hooked them directly to a heating element without the charge controller sitting in the middle. In the post I linked above the author provides a good explanation of what happens if you size the resistance element for the load the panel would provide with full sunshine vs what happens if you size the load for half sunshine. Basically, the lower sunshine situation calls for a higher resistance circuit to fully utilize the lower power and slightly lower voltage.

The Dumb Question

So my question is...what happens if increase the size of the resistor so that the element makes full use of the element on a cloudy day? The resistor would be larger (have a larger impedance) than would be ideal when the panel was in full sun. My question is: What would the effect of this be?

The total power going through the resistor doesn't go up, so I don't *think* it would burn the resistor up. Right? Would it just throttle the current in a way that would reduce the amount of power it would convert to heat? My simple understanding of these things is that any power I put into a heating element should come out as heat, and the limiting factor should be whether the current and voltage wind up melting the element. So I look at this and think that "well it won't melt the element if I keep the power under the listed threshold, so what's the problem, how does it lose efficiency?"

Thanks for the help

believe that the lack of efficiency issue is the fact that if the voltage at the panels is not reasonably close to their maximum power point they will only produce a tiny fraction of their potential output.

 

[rhino]

David Poz did create an excel spreadsheet that does calculate based on the panels you are using what element resistance you should use if you scroll down a bit on this page: https://www.davidpoz.com/. There is also some videos on YouTube showing some people doing PV direct water heating. I assume you already know you have to figure out how to turn the DC power off based on temp since you can't use the built-in thermostat controller of the water heater that only works with AC.

 

[aaron_c]

I

believe that the lack of efficiency issue is the fact that if the voltage at the panels is not reasonably close to their maximum power point they will only produce a tiny fraction of their potential output.

Huh. I thought that the maximum power point of the panels was kind of in-built. In other words....I thought the panels put out what they put out. They're set to put out a certain voltage but that voltage is, in practice, modified by the temperature of the panel. And the maximum power point is modified by...I would guess both the temperature and the irradiance available. But I thought that just meant that in different conditions the panel might put out a slightly different voltage and a very different current, with the result that the panels function best in certain conditions and function suboptimally in others. And I thought that the MPPT charge controller was only important insofar as the electricity consumer required a specific voltage input--like with a battery--because the charge controller took the slightly variable voltage the panel supplies and transformes it to exactly the voltage the consumer requires (provided everything is sized and wired properly, anyhow).

Does the charge controller actually impact the charge of the panels themselves, allowing the panels to always function at maximum efficiency? Is that where the efficiency is lost without one?

 

[Quattrohead]

Solar panels are a poor power generator and will sulk if they do not like the connected load. An MPPT controller will feel out the panels and figure out how best to treat them to make sure they put out the goods !!!!
So for a purely resistive dumb load, you need to match the resistance to the panel for best output. I think there is an in depth discussion either here or on the tube.
As this will be a DC current, you cannot use a standard thermostat as it can weld the contacts closed...they are not quick or open wide enough for DC.

 

[Bobert]

Huh. I thought that the maximum power point of the panels was kind of in-built. In other words....I thought the panels put out what they put out. They're set to put out a certain voltage but that voltage is, in practice, modified by the temperature of the panel. And the maximum power point is modified by...I would guess both the temperature and the irradiance available. But I thought that just meant that in different conditions the panel might put out a slightly different voltage and a very different current, with the result that the panels function best in certain conditions and function suboptimally in others. And I thought that the MPPT charge controller was only important insofar as the electricity consumer required a specific voltage input--like with a battery--because the charge controller took the slightly variable voltage the panel supplies and transformes it to exactly the voltage the consumer requires (provided everything is sized and wired properly, anyhow).

Does the charge controller actually impact the charge of the panels themselves, allowing the panels to always function at maximum efficiency? Is that where the efficiency is lost without one?

Solar panels don’t produce as much power when they operate outside of their maximum power point. The maximum power point is a moving number depending primarily upon the amount of sun the panel is receiving. Temperature and resistance in the wires does affect power output but unless these factors are extreme they have little impact on the output of the panel compared to the voltage the panel is being held at to produce power. For instance this morning my MPPT controller stuck at 62v and with early morning sun I was pulling in about 300watts of energy. I reset the controller and it successfully found maximum power point at 92 volts bringing in 500watts..Presently I am in partial shade an only get half of my potential output under ideal conditions. Past experience has proven to me that to low of a voltage can reduce your output to 25% of your potential output. As long as your voltage is near 0 you will produce almost no energy. Vice versa if the voltage goes to high you will also produce almost no energy. It seems to me that getting the proper resistance to operate at various conditions efficiency would be nearly impossible.

 

[aaron_c]

Thanks @Quattrohead and @Bobert

OK, so if I'm understanding yall correctly--and I hope this isn't gibberish--the issue is that increasing the resistance of the solar panel's circuit, which is necessary to use that power, changes the...current?...put out by the panel. And the change in current, in turn, changes the voltage as per the IV curve. So unless the total resistance is just right--which I guess is what charge controllers do, MPPT more precisely than PWM(?)--you can wind up with, as Bobert said, 25% of the power you'd have hoped for under more ideal circumstances.

[This is a generic IV curve. So the way I'm understanding this is that the resistance lowers the current, and that then changes--for better or for worse--the total power output. So if in the example above the current was dropped to 4 amps the voltage might be raised to something like 19 volts. However that would drop the total power output to to something like 60% of the MPP.]

I *think* this is why PWM controllers don't do as well in low-light conditions as MPPT? Because in low light the current drops but the voltage increases (slightly--it's still a net loss in efficiency for the panels). Since the PWM controller can't do anything with the extra voltage, it just cuts the voltage down to the set point, whereas an MPPT controllers can convert the extra voltage to current, making the panels more efficient in low-light conditions.

Let me know if I've got this right or wrong.

 

[aaron_c]

Part II:

Here's where I may go off the rails even more. If I hooked a solar panel up directly to heating elements--no charge controller, no battery, no nothing--then the resistance I chose would impact the current. And the current would, in turn, impact the voltage and, thought the IV curve, the power output of the panel. Right?

But what I'm trying to understand is what happens to the efficiency of the panel in this situation. I get that PWM charge controllers are less efficient than MPPT because of the inability to turn excess voltage into usable current. And I *think* I get that changing the resistance of the circuit will alter the current and thus also the voltage.

What I don't understand is what happens to panel efficiency when irradiance drops and there's no charge controller at all. Should I imagine the red line in the above graph being basically squashed down vertically so that it intersects with the Y axis at about 3.25? That wouldn't seem to change the voltage of the MPP all that far, I don't think? So why was I told by the person from Missouri Wind And Solar that I'd lose 40% panel efficiency doing it this way vs with a charge controller? Is it...something with a massive efficiency drop when irradiance causes the current to drop and I have no way to change the total resistance of the circuit?

 

[efficientPV]

Solar panels are a current source and with direct connect fixed resistance, power is a function of the square of the voltage. So, small drops in current result in considerable drops in power. The chart below shows the increase in power produced with a Power Point Controller over direct connect.

% Rated Increase over
Panel Amps Direct Connect
With Power Point
100% 0%
90% 10%
80% 25%
70% 50%
60% 67%
50% 100%
40% 250%
30% 333%
20% 500%
10% 1100%

The 80-40% current region provides a critical power gain. Below that is just gravy providing quicker recovery. I use a controller in parallel with charge controller for efficient diversion. Even though I have a small system I have enough excess energy for a second hot water tank just for the laundry which uses hot water for all cycles. Below is a typical system which is fairly simple with dew parts and shouldn't be a maintenance item.

This power graph is just a half hour of diversion on a day with scattered clouds showing how the voltage remains constant and power is quickly diverted.

Third is a study which shows David Poz is wrong in his assumptions about ideal resistance. His ideal resistance can almost be doubled and average daily production increases in direct connect. This is because maximum current is rarely seen for any period of time and designing for lower currents is more realistic.

 

[aaron_c]

@efficientPV thank you, that is very helpful!

Actually, I think the most useful bit may be the chart I found in the paper where your center image came from:


So if I understand this correctly, with the panels hooked up directly (without charge controller/battery/inverter) to a resistance element, I will either be optimizing the system to work with a particular amount of light. As shown in the diagram, if I optimizing it to work with 200 W/m^2 irradiance then it should produce almost the same amount of heat whether it's sunny or extremely cloudy. That amount of heat would be about 1/5 the rated output. If I'm aiming for a 90% solar fraction then maybe it's ok to get suboptimal returns on the cloudiest days, perhaps I should optimize the array for days with 400 W/m^2? Then, just to eyeball it, I might get a fairly reliable 250*3=750 watts on a nominal 2000 watt array.

Figure maybe 9 hours of light in the winter, so for every 1000 watt array I'd get 375 W * 9 h = 3,375 Wh/winter day. I'm figuring I'd need about 56 kWh of winter heating capacity (home and water) to reach about a 90% solar fraction. So I'd need maybe 16.6 kW of rated panel capacity for that. That's more than the 9.4 kW I estimate I'd need with a charge controller.

Folks might say "why don't you use an MPPT controller?" and aside from the issues with finding one that will work without a battery, there's the cost. The one I found cost $330 for 1200 watts of capacity. I don't know how long this off-brand controller would last, but even if it lasts 8 years that's still maybe 10 $340 charge controllers (I added $10 for installation) every 8 years, so the thirty year cost is about $340 * 30 / 8 * 10 = $12,750. If the panels cost about $1.60 per watt (install included) that's comparable to the $11,520 that the additional panels would cost. And the benefit would be to sunny day efficiency, which isn't where I need the added efficiency in this scenario.

Compared to solar thermal this might actually be about the same price but with somewhat better efficiency in low-light, low temperature conditions (and worse efficiency in high temp, high light conditions but as I can do nothing with the excess heat this doesn't concern me). And no huge storage tank for low temperature heat storage.

I do have a follow-up question: Does anyone know what safety precautions people take when wiring panels directly to DC heating elements? I don't love the idea of having heating elements that automatically come on when the sun is up, regardless of the tank's temperature and/or water level. It seems like a fire risk if somehow the water in the tank dropped to low. Is there some way to ensure that the elements would turn off if that happened? A circuit breaker wouldn't work, I imagine. I mean...probably it's a good idea to protect from short circuits? But it wouldn't fix the fire danger from leaving a heating element on in an empty tank, right?

With solar thermal drainback it seems like the systems are set up with heat dumps, and then if the heat dump fails there's a pressure release valve.

Anyhow, since it seems like people have installed these systems I'm hoping folks have thought this through already and I'm curious if anyone knows what the solutions were.

 

[Leeds]

To offer a simpler explanation- You need to match the impedance of the source and the load to maximize power.

We older folks are well versed in this stuff because back when we all had hi-fi stereos this came up a lot.

With sound systems you need to match the impedance (resistance) of your speakers (load) to that of your amplifier (source) so that they you get the most volume. Generally that was 8 ohms for home audio and I think about 4 ohms for car audio.

But because the impedance of a solar panel changes with the clouds an MPPT charge controller changes its impedance to always match.

 

[efficientPV]

I did label that chart 28.4

so for every 1000 watt array I'd get 375 W * 9 h = 3,375 Wh/winter day

I think that could be overly optimistic. I like to call any PV heating supplemental PV heating because you can never depend on solar. I use PV water heating only in the summer months at camp where it provides all the water heating till the last week of September. Then hot water for laundry disappears. A lot of people have more panels than they know what to do with and for them direct connect seems reasonable. It might even be the only solution for those who do not have an advanced knowledge of solar electronics. The products out there now are either too expensive or not well designed for long term service. I think it is nuts to have panels just for hot water. Any successful hot water system should turn off and then what happens to that power. My wife doesn't want to hear excuses for why there isn't hot water when she wants to run the dishwasher. I only have limited space for panels and 70% of the panels I have are in the shade at any one time. From a southern google earth view of my property not one solar panel can be seen. But, I have hot water and lots of it. I use strategies beyond just having a power point controller to be successful.

Easiest thing is to just tell me what you are thinking of doing so I can say what is wrong with the idea.