EDIT: I'm settling for a simpler system: no solar thermal panels; just a small 50 liter water heater, AC powered, run through the inverter. (So, practically no hot water in the winter time.)

I'm planning to build an off-grid summer house in southern Finland.

The plan includes:

  • electricity: PV panels + inverters + batteries, enough for a 2000 W induction stove (single burner).

  • water: a well for drinking/kitchen/shower water, an electric pump

  • heating: a body of water (or a non-toxic propylene glycol mixture) is heated in the heat tank. The water from the well would be heated in a coil of the tank, thus the water we use will never stay in the heater. The tank would be heated by solar thermal panels and/or a water heating fireplace. A small pump is needed for the thermal panels and another one the fireplace.

The cabin would mostly be used during the summertime (June to August), when the building doesn't need to be heated. For May and September-October, some heating will be required, and for the winter months, a lot of heating. The cabin is 70 m² (750 sq ft).

Since there is no external electricity and the outside temperature gets below 0°C (32°F) for several months, some measures must be taken. The lowest temperatures are -30°C (-22°F), which would mean 50 % glycol.

Batteries will survive only if they are fully charged for the winter.

The tank (and all pipes) must either be emptied for the winter, or it must contain non-freezing glycol mixture.

I think I now need to decide between these two options:

1) Fill the tank with water, and empty it for the wintertime. Heat it with solar heating panels.

  • Pros:

    • simple system
    • cheaper
    • the fireplace is not connected to the tank, so it’s easier to use it for heating the cabin in the winter (without the risk of overheating the tank)
  • Cons:

    • no possibility for water in the wintertime
    • obligatory autumn maintenance

2) Fill the tank with propylene glycol. Heat it with a water heating fireplace, and possibly solar thermal panels.

  • Pros:

    • less maintenance?
    • Possibility for occasional hot water even in the wintertime? (Must empty the water pipes every time, though)
  • Cons:

    • more expensive
    • have to add radiators (using the same glycol) to be able to heat the cabin in the wintertime, distributing the heat from the fireplace

The second option raises some important questions:

  • Is 50/50 glycol too thick for the pumps? Can 50/50 transfer enough heat?

  • How often must the glycol be changed? I need 500 litres (130 gallons), so the expense is non-trivial.

  • Most important: is this feasible? Does anyone have this kind of setup? I found very little information on the subject.

  • 1
    Solar water heating in cold climates usually employ a drain back system so that when the sun goes away, the entire system drains and no freeze damage can occur. The downside to that is the need for a circulation pump to run while heating is taking place. In an off grid scenario, constant electric loads are death. But if there is enough sun to heat water, pv panel can run the pump. Commented Jun 5, 2016 at 6:01
  • Check with your local regulations, maybe you need extra safety features when you use glycol because leaks can be an environmental hazard.
    – mart
    Commented Jul 7, 2016 at 9:09

2 Answers 2


It's quite common, it can be pumped, it's a bit less efficient to pump particularly when cold and somewhat poorer at heat transfer than pure water.

Primary impact on service life is minimizing oxygen exposure (sealed system) and choosing a PPG product with a suitable pH/corrosion control additive package.


Since this is a new-build, I would focus really hard on passive solar design. This is a new concept of building, with different materials and practices - it's not a glue-on afterthought to a conventionally built house). This type of design is likely to be earth sheltered, heavily insulated, have huge thermal mass inside the insulation envelope, windows with southern exposure slurping up every bit of sunlight. Snow and drifting have to be thought about. This is hard, but extremely worth it, because the house literally heats itself.

Don't even try to use photovoltaic power (solar panels) to make electricity to make heat. That is a complete thermodynamic "net lose" - you can't make enough heat with PV electricity to make a difference. Go ahead and spec it out, and you will see. If there's any way you can get natural-gas or propane to the site, even in portable containers, that's the way to go for as-needed use such as cooking or drying.

Wind is not a great way to get electricity for heat either, but it's sure better than solar PV.

To get heat from the sun, go solar-thermal. Obviously that's not much use for heating or clothes drying.

There is one heating method that's not so bad: heat pumps. Those can transfer 2-3 times as much heat as the electricity they consume.

To store heat with solar-thermal, you need lots of thermal mass. The heat-transfer fluid (glycol) will obviously be some of that mass. But it doesn't need to be all of it! You can fill the storage tank with rocks, (note a misconception I'll describe at the bottom) as long as the glycol can still circulate through it. If your home has high thermal mass as part of a passive solar design, that mass counts too. There's nothing wrong with pumping extra heat into the house while the sun is out, especially if you're absent.

You will also need ways to keep snow off the collectors during your extended absences. Given the northern clime it may be worth putting the panels straight vertical - think about that option when laying out the site - but blowing and sticking snow may still be a concern.

And then, you will need to think about coping with potential long periods of no useful sun owing to weather. It may be worth thinking about a fuel "backup heater" of some kind but you will need a very substantial tank if you won't be around to resupply it.

There is an interesting misconception about thermal mass, which is the term-of-art used within the industry. The term is actually wrong. Mass doesn't store temperature - atoms do, at roughly the same amount per atom. And in solids, roughly the same number of atoms fit in the same space... so most solids (volumewise) are roughly equivalent in heat storage. Ranging from 1.8 to 3.0 J/cc/degK of volumetric heat capacity. It's best to work in volume, because people design homes in terms of volume (dimensions) not mass. Given the narrow spread, material X vs material Y won't perform that differently, but cost certainly varies. And that's always a factor in the capitalist real world.

They say the best things in life are free. That'd be plain water at at 4.1 J/cc/°K. It is literally the perfect thermal "mass" if you can engineer away the freezing problem (i.e. drain the thermal-storage calandria when you're away). If cost is no object, use 50/50 glycol at 3.8, not only dense but pumpable. If cost is an object, well, that's where common materials come in, even if their performance is as little as half of water. It's worth minding their values, but unit conversions throw a lot of people (and crash space probes) - stick with volume-based units: joules per cubic centimeter per °K and direct SI equivalents (MJ/m3/°C etc.) Here are some sources.

  • Stones are a bad idea. True, stones have more mass per unit volume. But you are not taking into account the specific heat of each substance. Water stores roughly 20 times more heat per kg than stone. True, stone is more dense. But only about 2.5 times more dense. So, water holds about 8 times more heat per unit volume than stone. Thus, adding stones to your tank loses you thermal mass. Commented Jun 5, 2016 at 5:50
  • 1
    Your numbers don't sound right to me. Mass doesn't enter into it. Scientifically speaking, I think the relevant unit is "volumetric heat capacity", 2-3 J/cm3degC for most solids. True, water wins at 4.1, but now we march across campus to where they study the Dismal Science. OP needs Glycol (3.8) and that loses horribly on cost, while only being better than rocks by a factor of 1.4 to 1.9. Rocks are cheaper. en.m.wikipedia.org/wiki/Volumetric_heat_capacity Commented Jun 6, 2016 at 19:20
  • You expressly state that the OP "need(s) lots of thermal mass." Yet even by your numbers, he loses thermal mass by adding rocks (I guess that explains your shift to cost as a justification). But your numbers are wrong. Look at everything that could fairly be considered a "rock" in this table: bit.ly/214aQmh You'll see that common values hover around .2. Water is 4.18. So, water will hold ~20 times more heat per unit mass. You get to "volumetric heat capacity" by properly accounting for the density difference which I did and which should have been clear. Commented Jun 6, 2016 at 23:57
  • 1
    Stone is about 150lb/ft^3. That is 150lbs/7.5 gallons or 20 lbs per gallon. Water is about 8 pounds per gallon. 20/8=2.5 So, just assume the rock has 2.5 times more heat capacity (per unit mass) than it does, and you can directly compare that number with 4.19 of water. So, .2*2.5=.5. That means that per unit volume, water has 8.36 times more heat holding capacity than most rocks. True, you could pick an "exotic" rock that is better than .2. But not much. Indeed, displacing water with rubber balls would work better (rubber~=.5)!! Commented Jun 7, 2016 at 0:06

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.