Direct to water heating.

TLJester said:

Im not tied to the grid (boat system).
water heater is AC

Click to expand...

I am guessing it is a smaller tank with only one element..... This is not conducive to direct-to-water like I drew before.

TLJester said:

not tied to the grid (boat system).
water heater is AC

Click to expand...

Then the inverter would power the water heater load along with the other loads. I would not call that a dump load in the sense that it is not wired like the diagram @FilterGuy posted. Depending how the inverter was programed it would get the power it needed from available solar supplemented by the batteries. If there is a relay involved it may be programmable.

Ampster said:

Then the inverter would power the water heater load along with the other loads. I would not call that a dump load in the sense that it is not wired like the diagram @FilterGuy posted. Depending how the inverter was programed it would get the power it needed from available solar supplemented by the batteries. If there is a relay involved it may be programmable.

Click to expand...

I think of dump load as different than direct-to-water.


There are other ways of doing dump load type arrangements, but this is the traditional way of doing a dump load.

The term 'dump load' comes from Turbine based systems like wind and hydro. Turbines need to be under load at all times so if there is no call for the energy, the controller would 'dump' the load to some kind of a big resister that heated up and dispersed the energy. People realized that this was wasted energy and started using it to do things like heat water. However, they still need a traditional dump load for when the water reaches temp. In the diagram below, the thermostat at the heater element will re-rout the power to the dump load when the water reaches the set temperature.

I wanted to throw in another update on what I have done on my end and future plans.

I have changed my 40gal electric water heater to 24v Dernord heaters, hooked in parallel through an 80A motor contactor. The contactor currently has two overrides, a manual on off and a thermostat. At 26+V on my off-grid system nominally that pushes the elements to over 1kw apiece. Having about 12kw useful battery life I can get the water warm enough for a comfortable shower in most conditions, but keeping it set at 71C the tank will stay hot for several days without any additional power if I thermostat out on a good sunny day.

Future plan is to have the ability to take either half or all of my array and send it directly into the water heater as energy becomes excess. Would require switching from parallel elements to series, and redirecting the panels from the SCC to the elements. Arduino program in the works.

falgsc-al said:

I have changed my 40gal electric water heater to 24v Dernord heaters, hooked in parallel through an 80A motor contactor. .....

Click to expand...

Hopefully the contacts are DC rated.

falgsc-al said:

Having about 12kw useful battery life I can get the water warm enough for a comfortable shower in most conditions, but keeping it set at 71C the tank will stay hot for several days without any additional power if I thermostat out on a good sunny day.

Click to expand...

Do you use a tempering valve? I use one on my heat pump water heater to effectively increase its storage capacity. I only set mine to 54 C but the tempering valve mixes the outgoing water so there is no risk of scalding.

Ampster said:

Hopefully the contacts are DC rated.


Do you use a tempering valve? I use one on my heat pump water heater to effectively increase its storage capacity. I only set mine to 54 C but the tempering valve mixes the outgoing water so there is no risk of scalding.

Click to expand...

Spoke to an EE doctorate and he advised I derate the contactor and sent me a chart. At lower DC voltages it said derate by .5-1 (half to none) hence the 80A contactor... which is twice the current passing through and using two of the poles to connect the battery instead of 1. The new dpdt relays are DC rated and overrated.

As for a temper valve, thats a no. I live alone and I like disinfected dishes so I use it with caution, and the hot water is plentiful. If I had family, most definitely solely on safety reasons. I do have a temper valve on my bidet for obvious reasons.

I opted to keep it as simple as possible by using my inverters 240v output. During the middle of the day a timer turns on the water heater for about 3 hours. I also switched out the 4500 watt elements for 2500 watt ones. The house I bought has two 80 gallon water heaters plumbed in series with only one hooked up to power.

My next plan is to also swap out the elements on the other one and feed 120 volts to both which would to give me about 500 watts of heating to each one and I will extend the timer on time for both. The water won't start out getting as hot but I don't use a lot and over several days I believe it will get hotter.

efficientPV said:

There is successful and then there is really successful. With enough space and panels heating water isn't a problem. I do diversion heating directly from array voltage and not using a battery or charge controller by diverting any increment of unused power not needed for charging. I have a 40 gallon tank in the garage just used for the clothes washer. All laundry cycles use hot water, it's free. This is excess diversion from diversion system that heats the house water. This is just an example of what can be done. I know there are lots of used panels, but think of how that separate hot water array could be used for household needs.

12V panels with a 12V heating element is always a bad match. A heating element closer to 18V or higher should be used. 24V panels will match closer to 36V elements. You don't want to over voltage elements. Dual element heaters can be placed in series instead of parallel shorting bars. This increases your resistance options. I use 2,000W 120V elements, these are 500W at 60V. There is an ideal resistance, but that is only useful if you live where there are no clouds. If the panel can produce 10A max, 7.5A would seem pretty good. Fact is the power has dropped to 50%. At 5A it id down to 25%. Some studies have shown increasing ideal resistance 50% does not affect daily power production that much and is a big help on those partly cloudy days.

I use power point control to always get what the panels can produce. I have very limited space. A number of panels are not even on my property. I keep them at a low angle almost flat so they can't be seen. All the panels suffer from extreme shading and I have to clean off twigs every morning because they are under trees. It is a site that shouldn't work art all. And I have hot water.

Click to expand...

Hello,
I'm thinking of many solutions to heat my water, for some month now, I end up with the idea of putting a bunch of resistors (elements) of different internal resistance in parallel with mppt/inverter (Chinese all in one 5kw) connected to the panels. I had the idea of using a pilot panel to guess the available power (using Arduino to control all that) and connect those resistors depending on the guessed power available.
My main problem is that my strings are 200V/10A, I could use (I'm in Europe) 230V element BUT...there is no way to secure the electrical circuit with a differential cutter ...cause it's DC. And that is a problem, I do not want to die of electrocution in 15 years when the element will be old and maybe contact with the heater chassis (ground may not be enough to save me)
So....the idea of using a buck converter to convert the 200V/10A To something "harmless", a dirty alternative current or a lower voltage dc (maybe 48V) arised in my head.
I then searched forum...and find this discussion.
So, did I understand correctly, you use a buck converter paralleled with your mppt/inverter and panels, how the buck converter only take unused power.and do not "steal" from the inverter.
Is the mppt still able to work and put the string in the mppt mode...?

Is the solution you actually use a mature one or does it still have main caveat..?
Do you sell board, schematics... or do you give schematics for free...?
I'm sorry for so many questions...?

My other solution is to use the AC from the inverter, guess the available PV power from the pilot panel, a sensor on the AC output to guess the used power, a shunt on the batteries and voila...
I was thinking of making a nice algorithm that would feed a database and improve itself Everytime the shunt detect a use of the batteries. Current flow.in the wrong direction in the shunt --> cut heater immediately.

Now...this idea got a big problem....if the inverter die...I got no hot water (well...I got a wood boiler, but let's 0ush it aside) and it annoy me...

I talked a lot with "Electrodacus" and his system is just to specific to it's own case, he got a house with concrete slab encapsulated in polystyrene, in this slab there are many resistive conductors...of different lenght and resistance. The DMPPT then connect the the best option it got on hand to pair panels with resistances.
I would like to see what happens when some resistive wire died in the concrete..... I was very excited by this project like 2 years ago...then I saw all problems that go with it.
Yes it's beautiful, yes it's kind of efficient and durable...but ...drawbacks are not compatible with my needs and those of most people.

curiouscarbon said:

This paper reminded me of this thread: Performance of a Solar-Assisted Air Source Heat Pump in Heating Condition [zhou2010] it's about heating indoor air but applies to water heating just as much.

Reviewing options:

1. Joule Heating: resistive element in water tank
2. Air Source Heat Pump: pump heat from outdoor air into water
3. Solar Thermal Source Heat Pump: pump heat from solar concentrator into water
4. combination of all the above

Solar Thermal Source Heat Pump can use either dedicated solar concentrator or be integrated into the back of photovoltaic panel.

My ideal solution would be a photovoltaic panel with heat exchanger on back. During winter pump supplemental heat out of them to harvest otherwise wasted solar thermal energy. Bonus: photovoltaic production will increase due to temperature coefficient of most panels...

This would allow the air heat harvesting device ("condenser") to be smaller

View attachment 59488

Click to expand...

It exist in my country, PV with water flowing behind them, it's not efficient and the cost make them only usable when you lack space for your PV.

As for the one with air circulating behind them...they also exist but are pretty bad, expensive....not efficient at all.

Honuz said:

Hello,
I'm thinking of many solutions to heat my water, for some month now, I end up with the idea of putting a bunch of resistors (elements) of different internal resistance in parallel with mppt/inverter (Chinese all in one 5kw) connected to the panels. I had the idea of using a pilot panel to guess the available power (using Arduino to control all that) and connect those resistors depending on the guessed power available.
My main problem is that my strings are 200V/10A, I could use (I'm in Europe) 230V element BUT...there is no way to secure the electrical circuit with a differential cutter ...cause it's DC. And that is a problem, I do not want to die of electrocution in 15 years when the element will be old and maybe contact with the heater chassis (ground may not be enough to save me)
So....the idea of using a buck converter to convert the 200V/10A To something "harmless", a dirty alternative current or a lower voltage dc (maybe 48V) arised in my head.
I then searched forum...and find this discussion.
So, did I understand correctly, you use a buck converter paralleled with your mppt/inverter and panels, how the buck converter only take unused power.and do not "steal" from the inverter.
Is the mppt still able to work and put the string in the mppt mode...?

Is the solution you actually use a mature one or does it still have main caveat..?
Do you sell board, schematics... or do you give schematics for free...?
I'm sorry for so many questions...?

My other solution is to use the AC from the inverter, guess the available PV power from the pilot panel, a sensor on the AC output to guess the used power, a shunt on the batteries and voila...
I was thinking of making a nice algorithm that would feed a database and improve itself Everytime the shunt detect a use of the batteries. Current flow.in the wrong direction in the shunt --> cut heater immediately.

Now...this idea got a big problem....if the inverter die...I got no hot water (well...I got a wood boiler, but let's 0ush it aside) and it annoy me...

Click to expand...

Make a differential cutter (GFCI?) built for DC. A DC current shunt in each leg with their outputs connected to cancel eachother out. If one side draws more current than the other, you'll get the differential voltage to trigger your safety disconnect with. The circuit details so far as output isolation are more involved, but that's the basic idea. Hall effect sensors might work too, but I'd trust current shunts more.

K W said:

Make a differential cutter (GFCI?) built for DC. A DC current shunt in each leg with their outputs connected to cancel eachother out. If one side draws more current than the other, you'll get the differential voltage to trigger your safety disconnect with. The circuit details so far as output isolation are more involved, but that's the basic idea. Hall effect sensors might work too, but I'd trust current shunts more.

Click to expand...

Hall effect seems to have a large offset. I have to keep zeroing the DC clamp scope probe I use.

Shunt - how much error in each?
If carrying 10A DC, want to detect and shut off for about 10 mA or above, because something in that range or slightly higher causes involuntary muscle contraction (can't let go), and above that death. At least for AC, we use 5 mA in the U.S. for GFCI. In Europe, 30 mA for whole-house GFCI (but that's a level where muscles contract.) Might be a different physiological limit for DC.

Probably difficult to ensure < 0.1% difference in measurement error between the two DC shunts.

For AC, the two wires go through a magnetic core, fields cancelling each other, and only common-mode is coupled to sense line. I think that does a better job of detecting small differences. We have 30A, maybe 50A GFCI designed to detect 5 mA leakage so 0.01%

TLJester said:

ahh never quite understood that. Have not found DC elements for the hot water tank/heater i have. so would assume in that case its a relay off the AC inverter ?

Click to expand...

You can feed the solar directly into the DC element, just disconnect a number of panels from the battery to feed the element, you can do that any way you like. You can also do that via the inverter, but why tax the inverter and incur the losses if you do not need to.

K W said:

Make a differential cutter (GFCI?) built for DC. A DC current shunt in each leg with their outputs connected to cancel eachother out. If one side draws more current than the other, you'll get the differential voltage to trigger your safety disconnect with. The circuit details so far as output isolation are more involved, but that's the basic idea. Hall effect sensors might work too, but I'd trust current shunts more.

Click to expand...

Hello,
There are differential cutters for DC...but they cost a shitload money.... 300€ in my country. They are called PFIM if I'm not mistaken.
Here is one : https://www.habitamat.com/interrupt...eur-differentiel-30-ma-type-b-40-a-dx-id.html

Hedges said:

Hall effect seems to have a large offset. I have to keep zeroing the DC clamp scope probe I use.

Shunt - how much error in each?
If carrying 10A DC, want to detect and shut off for about 10 mA or above, because something in that range or slightly higher causes involuntary muscle contraction (can't let go), and above that death. At least for AC, we use 5 mA in the U.S. for GFCI. In Europe, 30 mA for whole-house GFCI (but that's a level where muscles contract.) Might be a different physiological limit for DC.

Probably difficult to ensure < 0.1% difference in measurement error between the two DC shunts.

For AC, the two wires go through a magnetic core, fields cancelling each other, and only common-mode is coupled to sense line. I think that does a better job of detecting small differences. We have 30A, maybe 50A GFCI designed to detect 5 mA leakage so 0.01%

Click to expand...

This 10mA things depends of course on the voltage. In my country, in France, we use 30mA, it's the code, it's everywhere in the house, for all "lines" of conductors. We use 230 VAC/50Hz.
If there is a current leak, higher then 30 mA which is something like 7 watts, this part of the circuit is automatically cut off from the system.

What voltage is safe? People used to die with 32V Delco generators if they were sweaty. I admit I was a little anxious when I climbed into a shower in Europe and there was a 230V instant heater there with me. There has to be a current path. Ground makes a local safe area. Even with low voltage elements you have the potential to creating a current path with the panels. I use 60V which is relatively safe in most cases. You are already courting death with 200V so you must have some comfort with it. It sounds like you are currently using this as a system and are interested in paralleling that with the water heater.

A 2,000W power supply is not cheap. You likely only need 600W so four small power supplies are a cheap option. There are dozens of ways to do this. An arduino is an easy way to control these power supplies, multiple elements or pulse the element. All can be used to keep the array at near power point. If a current MPPT system exists, you can not put another in parallel that calculates power point with current and voltage calculations. My system uses a calculated NATURAL power point using temperature to predict that voltage. If a panel is rated at 60V standard conditions, if it reads 62V in the sun it is obvious it has power not being used by a charging system. My array can easily climb to 66V during the day when power point is at 56V. I just draw off enough heater power to bring the voltage down to something reasonable. I have multiple parallel arrays due to intense shading. I don't bother using temperature tracking as this can be adjusted seasonably. I have no problem heating all the water I need with my size array.

I do have a board and my stock of parts does not allow 200V. Shipping to Europe is also too expensive. Of all the boards I have sold in the last year, not one person has even gotten around to putting them in operation. I am no longer interested in selling boards. Many around the world have built these systems from scratch. None are interested in making a video due to controversy surrounding this method. Do it yourself apperars to be too technical for most in solar. I would suggest looking at the ACTii AC7391 which is made in Poland as an off the shelf solution.

Here is a picture of a 1300W system someone made in Lithuania. He had no technical background and even made his own circuit board. He followed one of my designs and heats water for floor heating. A little overkill because of the parts that were available. I was extremely impressed with the build for so someone with no electronics experience. The next is a guy in Florida who designed his own board. last is mine.

Honuz said:

This 10mA things depends of course on the voltage. In my country, in France, we use 30mA, it's the code, it's everywhere in the house, for all "lines" of conductors. We use 230 VAC/50Hz.
If there is a current leak, higher then 30 mA which is something like 7 watts, this part of the circuit is automatically cut off from the system.

Click to expand...

Yes, and they are available for AC.
The link you provided for one was also AC. Do you have a link for DC?

"
"

We had been concerned about switching DC for a water heater, because thermostat and over-temperature cutout are made for AC, would fail at moderately high voltage DC. You brought up the issue of ground fault, which is a valid concern too. Your 30 mA whole-house GFCI in Europe would shut off if water heater element started leaking current to ground. In the U.S., we generally only have our 5 mA GFCI on outlets outside, kitchen and bath, and now garage. Water heater would not trip the the GFCI, so only its ground wire would prevent a shock hazard.

You mentioned Arduino. Water heater as a dump load would ideally consume exactly the surplus power available. It could be toggled on and off, like PWM. On AC side this would cycle battery as inverter drew more power than PV provided. DC from battery would be similar. PWM from PV to heating element would draw no more than the power PV could supply into its resistance, so not has large a load during times of less sun.

A dimmer circuit on the AC side could be used to adjust power from zero to full rating of element. The waveform might be difficult for inverter to supply, with a multi-kW load as compared to small lights/motors. A separate square-wave or MSW inverter might be good for this.

The more ideal on AC side is a high-frequency switcher with variable width pulses to draw current in proportion to voltage, synthesizing sine wave current.

As an alternate to power electronics, on AC side, a Variac with its shaft servo-controlled would produce variable amplitude AC to the heater.

On battery or PV DC side, high frequency PWM could draw reduced average current. With an inductor (filter), current draw could be smoothed, reducing high frequencies that might confuse SCC.

In the DC PWM, the voltage and current from the panel remain relatively constant as the heating element is pulsed. Energy is stored in a capacitor bank. Again, the current to the heating element remains relatively the same while on. Only the duty cycle of the on time changes as in a light dimmer. One advantage is at lower frequencies a minimum off time can be used as an arc interrupt so that AC rated mechanical switches can still be used in series with the heating element. No inductor is needed.

Honuz said:

This 10mA things depends of course on the voltage. In my country, in France, we use 30mA, it's the code, it's everywhere in the house, for all "lines" of conductors. We use 230 VAC/50Hz.
If there is a current leak, higher then 30 mA which is something like 7 watts, this part of the circuit is automatically cut off from the system.

Click to expand...

Same system in South Africa.

efficientPV said:

I just added this separate 40 gallon tank in the garage just to wash clothes.

Click to expand...

I want to do the same thing, but also get a circulation pump and some sort of radiator to preheat the drier air so it doesn't use as much power to reach the temperature it is looking for.

Even on top of something like this: http://www.alkeng.com/AL/dryer Although, I cant find those cores for cheap.