diy solar

diy solar

Using NiChrome wire for PV diversion water heating

Sverige

A Brit in Sweden
Joined
Oct 8, 2020
Messages
733
Location
59.5N, 15.5E
I have a run of about 25m of copper water pipe which draws from my hot water tank and has a loop back (U connection) at the destination end (close to the taps in my house), also a circulation pump which will drive water through this circuit and, if the water returns warmer than it left the tank, the heat will accumulate into my hot water tank.

What I am pondering is the feasibility of using nickel chrome wire wrapped around the outside of the pipe as a resistive heating element to dump heat energy from my PV panels into the water, once my batteries are fully charged each afternoon. If the circulating pump runs at the same time and if the heat can be delivered to the outside of the pipes at a suitable temperature (below 100C!), I would imagine this might work and has the benefit of avoiding breaking into the pipework and doing any plumbing (which I suck at!)

But.. I have no idea how to calculate what length or gauge of NiChrome to use, how many turns around the pipework, what kind of insulator to put on the copper pipe first (so the NiChrome doesn’t short circuit), or in fact if the idea has any merit or will be a big waste of time and money :LOL:

The cabling from my PV panels arrives in the vicinity of this loop back connection, so I could wrap both the send and receive leg of the pipework and heat the water twice before it makes it back to the accumulating tank. I have around 2-3m of exposed pipework inside the house close to my PV system available for NiChrome to ne wrapped around.

2 of my PV circuits are from 3x 100W panels in parallel, with an open circuit voltage of 18-20V and short circuit current of around 18A once disconnected from my batteries. The other two circuits are 2S3P arrays with an open circuit voltage of around 54-58V and short circuit current of about 18A.

I’m guessing there won’t be many people using NiChrome for PV diversion, but is there anyone good with sums who fancies having a go figuring out what arrangement of wire gauge, turns and covered pipe length might work to get the water warmed enough but not boiled before it goes back to the tank? The flow rate of the circulating pump is “up to 1.5 cubic meter per hour” according to the spec sheet.

ps… I mentioned NiChrome as I’m familiar with it being a resistive heating product , but maybe there are off the shelf “heating mat” type products in a long thin form factor which could just be wrapped around the pipes? Let me know if you know or anything suitable ?
 
----------
EDIT

I went the wrong direction with my second calculation, so ignore everything about 36awg nichrome. See my later post for the correction.
----------
Given that you're not using an MPPT tracker, you'll lose quite a bit of energy, but you're already losing that energy so I assume it's not a concern for your application.

Nichrome heaters are resistance based, so you can calculate the length needed based on the voltage [1] and current input, knowing the wire resistance.

Typical 12V panels have an optimum operating voltage around 19v, and with 300W of panels the current should be around 16A.

We'll randomly choose a nichrome wire gauge - 22awg - just to get started, but will look at the results and optimize later.

22awg nichrome wire has a resistance of 1.015 ohms per foot [2]. If our optimal setting is 19V and 16A, then, via ohms law Voltage / Current = Resistance we find we need 19/16 = 1.2 feet of wire. If you heat 1.2 feet of nichrome wire with 300 watts of power it'll get very hot, and the temperature transfer into the water over such a small section won't be very efficient at all.

The equation is linear, so quick calculations based on wire resistance are possible. 36 awg nichrome wire has a resistance of 26 ohms/foot, so you'd need about 26x the length of the 22awg, or 31 feet - about 10 meters. So if you spiral 10 meters of 36 awg nichrome wire along your copper pipe, then surround it with fiberglass or ceramic insulation (to reduce the chance of fire if the wire does overheat), then almost all the heat will go into the water in the pipe. It will heat about 1 watt per centimeter, which is more than enough length to safely heat the water. Copper is conductive, so provide electrical insulation between the nichrome and copper - a few layers of kapton tape would work well for this.

As always, make sure you're protected against temperature and pressure in this loop - even if there's a temperature and pressure relief valve elsewhere in this closed system consider putting one at the 'exit' of this heater - each heater should have it's own T&P valve so future owners who modify the system don't accidentally create a closed loop containing this heater without appropriate protection. The pump is a one-way check valve, so it only takes one failure or valve closure to create a closed system with only this heater.

Note that this is suitable for the OP's situation, with a copper pipe and 300W of power input. Anyone coming along later should be aware that plastic or other pipe materials may cause this solution to be unsafe - do your due diligence when making any water heating system, steam explosions are just as deadly as conventional explosives, and electrical shorts, fires, and electrocution hazards can also occur if the design is inappropriate.


[1] Nichrome resistance varies with temperature, the hotter the wire, the higher its resistance. This is useful because it protects itself as it heats up and even prevents hotspots from forming. However, in this application the temperature range is so low and small (2C to 98C) compared to nichrome's maximum temperature of 1,177C that the resistance change can be ignored.

[2] The conversion to metric is left as an exercise for the reader. ;). Or just find the resistance in ohms per meter and perform the calculations again.
 
Last edited:
Why not just use pipe wrap heat tape (AC)? You could add a timer to only function during max solar time.
 
Given that you're not using an MPPT tracker, you'll lose quite a bit of energy, but you're already losing that energy so I assume it's not a concern for your application.

Nichrome heaters are resistance based, so you can calculate the length needed based on the voltage [1] and current input, knowing the wire resistance.

Typical 12V panels have an optimum operating voltage around 19v, and with 300W of panels the current should be around 16A.

We'll randomly choose a nichrome wire gauge - 22awg - just to get started, but will look at the results and optimize later.

22awg nichrome wire has a resistance of 1.015 ohms per foot [2]. If our optimal setting is 19V and 16A, then, via ohms law Voltage / Current = Resistance we find we need 19/16 = 1.2 feet of wire. If you heat 1.2 feet of nichrome wire with 300 watts of power it'll get very hot, and the temperature transfer into the water over such a small section won't be very efficient at all.

The equation is linear, so quick calculations based on wire resistance are possible. 36 awg nichrome wire has a resistance of 26 ohms, so you'd need about 26x the length of the 22awg, or 31 feet - about 10 meters. So if you spiral 10 meters of 36 awg nichrome wire along your copper pipe, then surround it with fiberglass or ceramic insulation (to reduce the chance of fire if the wire does overheat), then almost all the heat will go into the water in the pipe. It will heat about 1 watt per centimeter, which is more than enough length to safely heat the water. Copper is conductive, so provide electrical insulation between the nichrome and copper - a few layers of kapton tape would work well for this.

As always, make sure you're protected against temperature and pressure in this loop - even if there's a temperature and pressure relief valve elsewhere in this closed system consider putting one at the 'exit' of this heater - each heater should have it's own T&P valve so future owners who modify the system don't accidentally create a closed loop containing this heater without appropriate protection. The pump is a one-way check valve, so it only takes one failure or valve closure to create a closed system with only this heater.

Note that this is suitable for the OP's situation, with a copper pipe and 300W of power input. Anyone coming along later should be aware that plastic or other pipe materials may cause this solution to be unsafe - do your due diligence when making any water heating system, steam explosions are just as deadly as conventional explosives, and electrical shorts, fires, and electrocution hazards can also occur if the design is inappropriate.


[1] Nichrome resistance varies with temperature, the hotter the wire, the higher its resistance. This is useful because it protects itself as it heats up and even prevents hotspots from forming. However, in this application the temperature range is so low and small (2C to 98C) compared to nichrome's maximum temperature of 1,177C that the resistance change can be ignored.

[2] The conversion to metric is left as an exercise for the reader. ;). Or just find the resistance in ohms per meter and perform the calculations again.
This is awesome info, thanks, and it gives me a good start. A few questions spring to mind:

I wonder what the thermal conductivity of kapton is like, and whether it’s the best choice for this application or if there’s something which conducts heat well but is an electrical insulator?

With only 300W of heating, I don’t know if I will make a useful temperature rise in my 180 litre tank. I have 1920Wp of PV panels, although with cable losses maybe 1600W makes it to the head end location where this would be installed. If it takes until 1pm each day for my batteries to charge, I might have 4-5 hours available for PV diversion heating of water.

If the available length of 2-3m for heating the pipe is the limiting factor I could maybe install a further stretch of NiChrome in the garage, which is the space the water pipes transit through on their way from the tank.

Your point about needing safety measures is a good one, I had not thought about what happens if the pump stops. I’m not keen to break into the water circuit but I wonder if I could install some temperature monitoring on the outside of the pipe, as there would be a spike in temp if the water stops flowing. With some home automation products that would even be set up to disconnect the power from the nichrome.
 
Why not just use pipe wrap heat tape (AC)? You could add a timer to only function during max solar time.
Honest answer, because I’ve never heard of it and didn’t know it existed! If there’s a product which can be fitted around pipes to provide heating then it would definitely be of interest, even if it’s AC. I could either pull apart the input side and power the element direct with my DC or more likely just use an inverter to drive it.

I‘m not too concerned about efficiency as I’m just looking to burn off excess PV power which will otherwise go to waste.
 
Common water heater elements fit inside standard pipe fittings. Why reinvent the wheel. There would be poor heat transfer with the nichrome method.

"I‘m not too concerned about efficiency as I’m just looking to burn off excess PV power which will otherwise go to waste."
With everyone here water heating is just an afterthought, just cheap and easy.
 
Common water heater elements fit inside standard pipe fittings. Why reinvent the wheel.

Well because I’m not installing a new system of pipework, I’m working with what’s already installed as a sealed system I don’t want to break into. The main reason is that I’m truly incompetent at plumbing - it’s taken me literally 9 months to get my fridge connected to a water supply under the sink and this involved taking the connection to the cold tap at the sink apart 10 or more times, trying and failing to get the various different parts and adapters I’d bought to fit together without leaking.

I understand electricity. If something doesn’t work at first I can find a way to fix it so it’s right. Plumbing is a mystery and every part which should just screw together and seal evidently does not in my hands. A secondary reason, as if I need more, is I know that hot water system airlocks after it’s drained down and refilled. It did last time when the installers put it in and I had months of slow flowing water before it got solved.

There would be poor heat transfer with the nichrome method.

I don’t doubt you’re correct, but as efficiency is not a primary concern, I’m hoping that limitation can be overcome. Nevertheless the point of starting the thread was to get opinions on whether this can work so I’m grateful for your contribution ?
 
The resistance of 31 feet of awg36 nichrome is 806 ohms. At 19 volts, 23.5 mA will flow, giving you about half a watt of heat. I guess you can ignore the resistance change with heat.
Thanks for pointing that out. I see there are a few factors which need to be considered to find the sweet spot which might work.

I suppose there’s also the option of using step up or step down DC/DC converters if I really need to put a different voltage on the wire, or multiple circuits of wire in parallel.

Probably wrapping a super long wire around the pipe gets tedious, so maybe multiple shorter runs in parallel is advantageous anyway.
 
I've been saving coffee pot water heater elements (have three so far 17 ohms each) to make a hot water dispenser. They have a resistor bonded to the tube. In the same vein, this might work. There are power resistors that are designed to be mounted to a heat sink. I could imagine several wrapped around a pipe with a hose clamp to attach them. Wouldn't go over third of the rated power of each one due to lower contact surface. Use white heat sink paste to improve heat transfer. This method would give better electrical insulation.
 
I use NiChrome for my Lifepo4 battery heaters. There are two things to keep in mind- Insulation and watts per foot.

Insulation is important here, just wrapping the wire around a pipe will basically short it out and you'll burn something out. Always measure it with an ohm meter after installation and before powering it up to make sure you haven't got it touching any conductors.

Watts per foot is incredibly important. I have an electric crucible that uses NiChrome heating elements and it melts metal allowing me to cast. This wire can get *extremely* hot if you let it. It can literally melt the pipe you wrap it around as long as it's insulated enough to not short out and kill itself prior to the copper liquifying.

For my battery heaters I aim for closer to a half a watt per foot. This allows the battery to heat but it doesn't create hot spots that melt through the insulation.

 
I made a mistake with my previous post - ignore everything about the 36awg nichrome. I'll re-do the equations below. That's not the only mistake I made, so read on...

I wonder what the thermal conductivity of kapton is like, and whether it’s the best choice for this application or if there’s something which conducts heat well but is an electrical insulator?
Kapton tape is very thin, and it's not a terrible thermal conductor, but it's not a good one either. The thing is, though, that it doesn't matter - as long as the insulation outside the nichrome and pipe is great, then even if the heat flow through the kapton isn't perfect, it still has nowhere else to go and the majority will end up in the water and pipe. The only difference is that the wire space between the kapton and insulation will be a few degrees warmer than it would be if you thermally bonded the nichrome to the copper pipe.

If you're looking for efficiency, though, this isn't a great solution. You'll do far better putting in an MPPT controller than worrying about the kapton heat transfer.
With only 300W of heating, I don’t know if I will make a useful temperature rise in my 180 litre tank. I have 1920Wp of PV panels, although with cable losses maybe 1600W makes it to the head end location where this would be installed. If it takes until 1pm each day for my batteries to charge, I might have 4-5 hours available for PV diversion heating of water.
Ah, I misunderstood how much solar power you have available.
2 of my PV circuits are from 3x 100W panels in parallel, with an open circuit voltage of 18-20V and short circuit current of around 18A once disconnected from my batteries. The other two circuits are 2S3P arrays with an open circuit voltage of around 54-58V and short circuit current of about 18A.
This isn't perfectly clear to me given your statement of 1920w of panels, but the main point appears to be you have a maximum voltage and current of 20V and 18A on two wires, and 58V and 18A on two separate wires.
Your point about needing safety measures is a good one, I had not thought about what happens if the pump stops. I’m not keen to break into the water circuit but I wonder if I could install some temperature monitoring on the outside of the pipe, as there would be a spike in temp if the water stops flowing. With some home automation products that would even be set up to disconnect the power from the nichrome.
You can increase the safety of the system with thermal cutoff switches or thermal fuses, but it's still pretty risky to heat water without a pressure relief valve at the heater.

With two circuits of different voltages you'll need to calculate the resistance separately. I'm just working with the maximum voltage and current, more power could be extracted, but this should get you close.

On the lower voltage circuit you need a resistance of 20v/18A = 1.1 ohms to push up to 360W into the pipe.

On the higher voltage circuit you need 58v/18a = 3.22 ohms to push up to 1kw into the pipe.

18awg nichrome has a resistance of 1.4ohms per meter, and you can get 5 meters of it for about $10.

Use 1 meter for the low voltage lines, and 2.5 meters for the higher voltage lines. Wrapping them around 3 meters of pipe will be easy.
 
I made a mistake with my previous post - ignore everything about the 36awg nichrome. I'll re-do the equations below. That's not the only mistake I made, so read on...


Kapton tape is very thin, and it's not a terrible thermal conductor, but it's not a good one either. The thing is, though, that it doesn't matter - as long as the insulation outside the nichrome and pipe is great, then even if the heat flow through the kapton isn't perfect, it still has nowhere else to go and the majority will end up in the water and pipe. The only difference is that the wire space between the kapton and insulation will be a few degrees warmer than it would be if you thermally bonded the nichrome to the copper pipe.

If you're looking for efficiency, though, this isn't a great solution. You'll do far better putting in an MPPT controller than worrying about the kapton heat transfer.

Ah, I misunderstood how much solar power you have available.

This isn't perfectly clear to me given your statement of 1920w of panels, but the main point appears to be you have a maximum voltage and current of 20V and 18A on two wires, and 58V and 18A on two separate wires.

You can increase the safety of the system with thermal cutoff switches or thermal fuses, but it's still pretty risky to heat water without a pressure relief valve at the heater.

With two circuits of different voltages you'll need to calculate the resistance separately. I'm just working with the maximum voltage and current, more power could be extracted, but this should get you close.

On the lower voltage circuit you need a resistance of 20v/18A = 1.1 ohms to push up to 360W into the pipe.

On the higher voltage circuit you need 58v/18a = 3.22 ohms to push up to 1kw into the pipe.

18awg nichrome has a resistance of 1.4ohms per meter, and you can get 5 meters of it for about $10.

Use 1 meter for the low voltage lines, and 2.5 meters for the higher voltage lines. Wrapping them around 3 meters of pipe will be easy.

Thank you for this! Do i need to try to calculate what the increase in the temp of the water is, as it transits through that 3m heated section? I wouldn’t want to be turning it to steam before it makes it back to the homerun, returning to the tank…

The circulating pump spec says it can provide up to 1.5m3/hr but this may of course not be very accurate. The amount of useful power reaching the water might be hard to estimate given the inefficiencies involved. Perhaps the water will absorb all that energy easily and the temp increase with each pass will be just a few degrees and it’s nothing to worry about, but feels like I should make sure.
 
. However, in this application the temperature range is so low and small (2C to 98C) compared to nichrome's maximum temperature of 1,177C that the resistance change can be ignored.
Is there a way of calculating what temperature the nichrome reaches when 1kW+ is put through the 3.5m or so we are talking about? It’s just that above you’ve mentioned 98C as an upper limit (I’m in fact looking to heat the water in my tank to just 40C or so, from the 15C ambient it starts at), but I’m a bit worried about hot spotting where the nichrome might melt thru the kapton tape, which melts at around 280C.

I’d like to imagine the copper pipe full of water drags the heat away quickly enough to limit that temperature but if I’m dumping energy into the nichrome too quickly and the conduction of heat away from it is limited by the small surface area the wire presents then perhaps the wire can get much hotter than I imagine?

I do have another 10m or so of pipework in the garage into which I could be sinking the heat, if the practical rate at which the pipework can absorb heat is limited and running a lower level of watts per metre is desirable.
 
Using the calculator here:


50 gallons of water is 190 liters. With a start temperature of 20C, an end temperature of 21C, and a full 2kW input heat, it'll take 7 minutes. So if you total storage is 190 liters of water, it'll heat up approximately 8.6C every hour with a full 2kW of heat input.

I don't know the size of your reservoir, so you'll have to re-do the calculation based on that. A standard water heater in the US is 50 gallons, or 190 liters, so that's what I used.

Kapton is considered an insulator. It has a thermal conductivity coefficient of 0.12W/mK (Watts per meter-kelvin). Using this calculator:


And assuming a 1/2" (12.5mm" diameter pipe 3 meters long, enter 0.12 k, 0.112 A (area in meters square), 155 t1 (155C - max temp kapton), 100 t2 (100C max temperature of water), 0.000025 s (thickness - 0.025mm or 1mil kapton tape):

30kW is the heat transfer to maintain that temperature difference over that area of that thickness of kapton.

The equation is linear, so assuming less than 3kW of heat moving across the 0.11m^2 of tape, you should see that the nichrome wire space is less than 5.5C above the temperature of the water.

Hot spots in the kapton are self regulating - the higher the temperature, the higher the resistance, so the entire wire heats very evenly - this is one of the properties that makes it so useful, it wears very evenly. So the worst case scenario is if all the water boils away, the copper pipe starts heating above 155C, the kapton tape fails, and the nichrome makes contact with the copper pipe. As long as the pipe is well grounded, it can handle all 40 amps without issue, but since it's a good conductor with low resistance, it'll simply reduce the voltage drop, which will reduce the power output of the solar panels, assuming the panel ground is connected to the house and water pipe ground.

As long as you have a 90C thermal fuse thermally coupled to the copper pipe, you'll never see that temperature, and the water will never boil. Make sure you have one fuse per nichrome wire.
 
Back
Top