How do you heat a passive house?

Some people may think the answer is "you don't," but Passive houses do need some heating. 

The real answer is that you don't need very much, so it's not so important how you heat it. 

A comprehensive discussion can be found here from Zehnder Passive House, listing the pros and cons of each approach.

My personal favourites at the moment are underfloor heating and air-source heat pumps, athough I have not always thought so, and cannot guarantee that as a final answer. 

I wrote about underfloor heating before, highlighting the advantages: saving space, evenly distributing heat and increasing thermal mass; and also warning of the problems that can come from incorrect design and installation. One disadvantage that I forgot to mention is that it's very difficult to get to the underfloor pipes if something goes wrong with them, although they are just pipes and it's very unlikely that anything will go wrong.

Air conditioning units are becoming standard installations in Japanese houses, and their COP is getting better all the time. As well as cooling, they can reverse the circuit to heat air. In a regular house using these for heating can be uncomfortable since they are only heating the air, while the building itself stays cold and the temperature is not balanced. In addition, the hot air can rise giving you cold feet and a hot head when you stand up. This may also be an expensive way of heating your house, and it may even be both uncomfortable and expensive. 

Since a passive house has such low heating demands, modest air conditioning units can easily provide the heating needs of a building. The peak heating load of a Passive house should be around around 10 Watts per square metre, so for a 100-square-metre house, you need 1 kW. Units are typically rated at several times this, so one air conditioner can produce all the heating or cooling you will need for a whole house, although you may have to think carefully about where to put it. It can go on a wall or ceiling so, similar to underfloor heating, it won't take up any floor space. 

An additional advantage of an air conditioning unit is that it can also cool the house, and may also have a dehumidifier. The units are stand-alone, rather than incorporated into the ventilation system, so they may be less complicated. And depending on where you are, you may get something that comes with a guarantee, and can be easily be maintained or replaced.  

And by the way, if you're in Japan trying to work out how to read the symbols on the remote control so you can use the air conditioner this post from Surviving Japan may be useful.

Heating on and on heating

We turned the heating on 29th November last year, a couple of weeks earlier than last winter, but close to the normal time. It had snowed the week before, and there had been rather too many of the kind of November days you get in England.

The underfloor heating goes on from 8 to 8:30 then from 9 to 9:30, and a week later, according to the T and D thermometers we have in the slab, it's now about a degree warmer at the top than the bottom.

Underfloor heating, called radiant heating in the US, seems to elicit strong emotions, both for and against. In the building pantheon, it is a god for some, and a devil for others. 

There seems to be a strong sentiment among Passivhaus proponents against underfloor heating. I don't think it's completely logical. 

Certainly badly implemented underfloor heating has problems. Typical mistakes are: 

1. Only heating part of the floor

2. Not insulating the other sides

3. Setting the temperature too high

4. Using the wrong floor materials

5. Not having a thick enough floor

6. Not distributing the heating elements evenly

7. Incorrect dimensioning

8. Using electricity

If only part of the floor is heated, the heat will obey the second law of thermodynamics and go to the part that is not heated. Our neighbour has underfloor heating in the living room, but it is in the slab which also runs under their garage. A significant amount of the heat they put in will be trying to heating their garage.

In a Passive House the first two problems won't happen, as long as the heated floor is within the thermal envelope. Since the heating demand is low, there should never be a need to set the temperature high. 

The advantages of underfloor heating are:

Underfloor heating visible before the walls went up

1. There is no need to add radiators or any other heating devices that will take up floor space and wall space, attract dust, produce noise and may leak.

2. The heat is evenly distributed, and will radiate through the whole building.

3. Underfloor heating may increasing the thermal mass of the building and store heat.

Here's a picture of our underfloor heating, but it had to be taken before the screed floor was poured. There is usually nothing to see, unless you have a thermograph. You can see in the picture below with a temperature difference less than 5 degrees. 

26 degrees enough to heat a Passive House

So why do the Passivhaus people seem to be so against underfloor heating? It may be partly sticking to the Passivhaus ideal that the house can be heated only by warming the incoming air. In that case why is there no similar disparaging of radiators? All you need for underfloor heating is a boiler, and in fact most houses have one of those for the domestic hot water. Where do you think we get our hot water from... the kettle?

Since the temperature for underfloor heating doesn't really need to be above 40 degrees centigrade, there are lots of other options for a heat source, such as ground source heat pumps or solar thermal. With the low heat requirement of Passivhaus, a slightly oversized domestic hot water supply should be able to cope with underfloor heating in its stride. Typical boilers use power in kilowatts. The heating requirement of a passive house is of the order of 10 Watts per square metre.

There may be more technical arguments against underfloor heating, regarding the efficiency of the heat transfer. All heat losses are within the thermal envelope and will end up eventually heating the house anyway. If you're in a poorly insulated house where you want to quickly make a room warm while you're there, and not waste too much energy keeping it warm when you leave, then you may not want underfloor heating and its inherent thermal mass. Passive houses are designed to stay at an optimum temperature, and heating is not a quick fix for comfort on demand, but keeping the thermal envelope topped up.

Compared to radiators, underfloor heating probably requires a more powerful pump to circulate the heating fluid since there is a greater resistance. But pumps draw negligible power compared to the heat they deliver. 

This article from Buildinggreen.com gives four reasons for underfloor heating being the wrong choice in low energy homes. If you're ripping up the floors of an old house to put underfloor heating in, it's going to be expensive. But if you're putting some pipes in while you pour a screed for a tile floor it may be no more expensive than radiators. The piping may even be cheaper than for radiators since all the pipes can go into the floor in the same place, while each radiator will need its own pipe. 

The argument against overheating in houses with high solar gain may be more pertinent. However, passive houses don't need the heating on for twenty-four hours, so the heating could be switched on after the sun goes down. There should be no problem with solar gain then. And it can still be switched on for those winter days when there is no solar gain.  

Perhaps some powerful character within the Passivhaus movement had a bad experience with underfloor heating. Maybe he tried it once and got burnt, sitting on an overheated floor with his pants off. 

Healthyheating.com: Radiant floor heating myths

Connectingindustry.com: Underfloor heating vs radiators; everything you need to know

When did we turn the heating on last year?

Ask the slab. 

We have ten thermometers in the slab, recording and logging the temperature every few minutes. There are five buried at the bottom of the foundation, one near each corner and one in the middle, and another five in the screed, hopefully a safe distance from the underfloor heating pipes so that they are measuring the temperature of the floor and not just the pipes. They should give us an idea of when the heating went on though.

The best one to look at is probably the thermometer in the middle of the floor. Before the heating went on, it was nicely cycling day to day with the temperature at its lowest mid-morning, then rising from 10 or 11 am, as the heat of the sun found its way into the slab. The heat either does this directly by hitting the floor and conducting through the tiles and concrete, or indirectly by heating the air in the house, and the air heating the floor. My instinct is that solar radiation is going to have more effect heating the floor than ambient air, probably because I've been indoctrinated into the mantra of hot air rising, and the consequence that not much heat will be going down from the air into the floor. 

The data is a bit grainy, since we only have precision to a tenth of a degree while  James Joule reckoned he could measure the temperature of his beer vats to 1/200 of a degree Farenheit. Making the best of our 21st century tools, the floor seems to stop warming around 3.30, which is about when the sun stops reaching it directly. The room temperature is still a couple of degrees warmer at this point, but can be five or six degrees warmer around noon. Hopefully you can see this in the first chart, where the green line is the room temperature, the red line the temperature just under the floor, and the blue line the temperature at the bottom of the slab. Of course a bigger temperature difference means more heat would be conducted from air to floor, and in fact the ambient thermometer is half way up the wall on the south side of the house, so it's possible that the air temperature at the floor is only above the floor temperature until 3:30. Anyway, this is not strictly relevant to my question. 

On 5th December, we must have switched the heating on from around 5:30 am because the slab started to heat up then. From then, depending on the weather, there's a double peak effect when the morning injection of heat starts wearing off and the solar gain hasn't kicked in yet. 

On some days there was obviously no solar gain, and the temperature just falls after the morning boost. This happened on December 8th, 9th and 10th when there was a fair bit of snow. 

Then from 12 December we get a triple peak effect when the heating was also on from 7:30 or 8 pm for half an hour. This only lasted for a couple of days, perhaps until my Yorkshire genes got the better of me. 

The triple peak starts up again from 19th December, this time the heating going on from 10 pm for half an hour. 

We went away from 22nd December for 2 weeks, leaving the heating off. The screed went back to the diurnal cycle, with the temperature at the bottom of the foundation plumetting to an all-time low of 19.2 degrees centigrade. The lowest trough in the screed was 18.7 degrees. 

The heating went back on again in the mornings from 3rd January, and I can't tell exactly when it went on in the evenings. There is a rise in screed temperature around 8pm on 7th, 24th to 26th and 30th Janary, and 2nd, 3rd and 8th February. On other days there is a slight plateauing of temperature around that time, before the fall over the night, so I guess night time heating was on at least until 16th February. There are a couple of days when the heat went on around 10:30 pm.

We turned the morning heating off on 8th March, and just turned it on again 12th December. Instinct once again tugs at my coat tails, urging me to switch on the heating so the house doesn't lose too much heat making it more difficult later. Knowledge of thermodynamics suggests that making the house warmer is just going to mean losing more heat, so if we can survive the temperature, we should be OK. Experience also shows that it's not going to get that cold. It was still above 18 degrees in the middle of the slab with no heating on for two weeks at the end of December. Also, experience of the underfloor heating is that the response is not so bad, and while it doesn't give the instant blast of hot air you get from a fan heater or air conditioner, it feels warm within ten or fifteen minutes of switching it on. And if things get really desparate we can switch on the air conditioner, or do something really drastic like put socks on.

Whether to turn it on in the mornings or evenings is another question for another day.

Too bloody hot

December and January could be the hottest months in the house. At least, somewhat counterintuitively, they are the months with the highest solar gain. It's not that the sun is hotter in December and January. In fact, the sun is more or less the same temperature all the time, and cares little whether it is winter or summer in the northern hemisphere on Earth, but of course there is a difference in how much of that heat reaches the surface of our planet.

In terms of the radiation from the sun, there is more in the summer than in the winter. There are two reasons for this. First, the days are longer, so there are more hours of sunlight. More hours of sunlight mean more heat. Second, the angle of the sun is higher. This has two benefits. First, more sunlight is going to hit a given area of the earth. If the sun is directly above, a square metre of sunlight is going to hit a square metre of the earth. If the sun is 60 degrees below vertical, 30 degrees above the horizon, a square metre of sunlight will be elongated over two square metres of the earth so the incident radiation is halved. Also, the higher the sun is, the less atmosphere it has to get through, so the rays are stronger when they reach the ground.

The point with a house is that the windows are on the walls, so we aren't really interested in how much sunlight reaches a square metre of the ground. We want to know how much reaches a square metre of window. And this, almost by some divine intervention, means that in the winter, when we may expect it to be coldest outside, we get the most heat coming in through the windows. And when it gets warmer in the summer, less heat comes in. If we are careful with balconies and eaves, then we can try to keep this radiation to a minimum. Reflection is another thing that may lead one to believe that God invented windows, or at least that God was a double glazing salesman. The smaller the angle between solar rays and glass, the more is reflected and the less heat comes in. This means that more of the low winter sun will get through, and more of the high summer sun will be reflected.

So this is why it got up to 28 degrees centigrade in the living room at lunch time on 12th January, even though it was only one degree above freezing outside. The bottom line on this graph of temperatures over the first few weeks of our residence shows outside temperature (green - averaging more than one degree below zero). The highest temperature is inside temperature south (red at the top), and inside temperature upstairs north is pinkish below that, but dancing to the same tune. The others are slab temperatures. The big leap in inside ambient temperature was when we closed the windows and switched on the ventilation system on 23rd December, but you can see the jump in the temperature at the middle of the floor (light blue) as the underfloor heating started working on 26th December three days later. The effect at the bottom of the foundation slab (middle - dark blue) is slower, with about a three-day delay. At the North West corner of the foundation, the temperature change is much slower.   

Obviously it would be churlish to complain about the house being too hot, when all around are pouring gallons of oil into theirs and still freezing, and of course there are a few things that we can do before resorting to opening windows and letting the heat out. According to the thermometer in the upstairs north room, it is significantly cooler there, so if we open the inside windows from the atrium into the bedroom, the heat should go in there. Also we can open the door into the genkan and washitsu, which are to the north and significantly cooler. 

Part of the reason the north side is cooler is that the slab is much cooler there. This is by design. Kind of. The underfloor heating passes from the boiler to the south side of the floor, then to the north side of the floor, then back to the boiler, so the south side is being heated more effectively. Eventually the slab will probably have a constant temperature, but it actually seems like a good idea to have some temperature difference in the house. It would be nice to be able to control it a bit better, and I'm sure there is something we could do with the ventilation system. At the moment we are using a fan to blow air from the cooler northern parts of the house. 

But, going back to emissivity, I can't help feeling that it may have been a good idea to have had a higher emissivity for the floor and the walls so that they would have been absorbing more heat. What I guess is happening is that the radiation is just bouncing around the floor and the walls and getting the air really hot. The white terrace outside is probably helping by reflecting more sun into the house.

We're going to get some blinds soon anyway. I'd really like Venetian blinds with white on one side and black on the other, but I'm not sure if they are available or aesthetically pleasing. 

The Crookes radiometer shows the difference between black and white, invented by the eponymous Victorian chemist, William Crookes, who was pleased with himself for being able to make vacuum tubes. It was supposed to work as a kind of light mill, the white sides of each panel reflecting the sunlight, the black sides not reflecting anything, and spinning accordingly. When it started spinning, it went the wrong way; the black panels going away from the sunlight. The simple explanation is that the black sides get hotter than the white sides, and heat up the air molecules next to them, because actually the vacuum was far from perfect, which push the wheel around. A more detailed and accurate definition can be found 

here on wikipedia, unless the US government has shut it down. The difference the vacuum makes is to greatly reduce the resistance, so the effect of the heat becomes more significant. 

Turning up the heat

When we moved in, the underfloor heating wasn't working. If you're going to have no heating, then it's a pretty good house to be in.  The next morning it was -8 outside and 14 at the coldest place inside.

The heating engineers are on a direct contract, so they are not under the skilled and careful control of the builder, which may be why they hadn't got this ready, and it is probably a very busy time for that trade. There seemed to be some problem with the eco cute boiler and a new part needed sending. The architect had brought the heating engineers in a couple of years ago when we were still trying to work out some basic approaches. After this problem, he said he'd get them to throw in the air conditioner for free, telling us that the list price was 300,000 yen on it. For starters we didn't ask for a 300,000 yen air conditioner, and it seems to be massively over spec. It was certainly useful, though, to have the air conditioner to add some heat to the house, but working underfloor heating would have been much better. 

We moved in on 23rd, and the heating engineers came back three days later on Boxing Day to fix it. In the evening I went to play with the controls, which are still dangling from wires sticking out of a hole in the wall under the stairs. We went through lengthy discussions about the system with them last year, before the concrete slab went down and the pipes went into the screed on top of it. The original plan was to put pipes in the bottom layer of the slab too, so that it could be used as a heat store when solar heat was available but we didn't want to use it until later. In the event we just put pipes below the floor. Another item on the "wish I'd" list. 

Basically there is a figure of eight circuit, with the boiler upstairs, the underfloor piping downstairs and a mixing valve in the middle. One pump will circulate the water around the pipes under the floor, and a pump in the boiler will send water through the tank to get heat down to it. The mixing valve will turn the hot water from the boiler to a lower temperature. Each pump has a switch with a timer.   

The system seemed to have been working fine when I got back in the evening on Boxing Day. The floor was a lot warmer, up from 14 degrees in the morning to around 17 degrees. But the hot water tank was empty. In view of the heat capacity of the concrete, a good part of a metre thick and scores of square metres wide, this is not surprising.

Economically, I've been hoping to use cheap night-time electricity for heating the water, at 9 yen per kilowatt hour, and selling as much of the solar-generated stuff in the daytime as possible for 48 yen per kWh. So I set the boiler to start making hot water, and set the timer to start pumping heat around the floor. 

In the morning, the error was back again, and the pump from the boiler wasn't working. I called the heating engineer, who had not been planning to come again until January 13th, and he said they'd get here in the evening. The sun was up and a fair bit of heat gets in during the daytime, and we would certainly survive until then, and were in no danger of freezing to death anyway. 

As it happened the electricians were in that day, and part of their work involved switching the boiler on and off again, which cleared the error. I called the heating engineers, but they were coming anyway, and I did want to talk to them about what was happening. When they came they showed me how to reset the error, by pressing two buttons on the Eco cute control panel at once.

At that point, I hadn't realised there was a timer on the switch for the underfloor pump, which we're going to want to use during the daytime to get some of the heat from the sunny South side to cooler North. It just looked like a normal light switch, but they showed me how you can open it to reveal an LCD and some buttons.

The problem behind the error, it seems, is that the water coming out of the boiler is too hot. There's a thermyster on the pipe, and if it stays above a certain temperature for over 20 minutes, alarm bells start ringing and the boiler switches it off. The problem is a combination of the level of insulation we have making the water coming back from pipes into the boiler too warm, and the slow speed of the pump allowing it to heat up too much. This then gets to the mixing valve, and the hotter it gets the less goes into the floor circuit, because the mixing valve works to a certain temperature, so the slower it gets and the more it will heat up in the boiler upstairs.  

Later they will fix this by moving the thermyster to somewhere less critical. Already we've had to trick the boiler to let us use heat under 60 degrees. In the mean time we have to juggle the timing so that we get heat under the floor but leave enough in the boiler for baths and cleaning.

When we moved in the slab was around 10 degrees, thanks to building work and doors left open over the beginning of winter, rather than keeping doors shut and using the ventilation system, which seemed like a good idea to me.

It was never intended to get that cold. When we pump heat into the floor, a lot of it will be moving down to the bottom of the foundation to heat it up. Once it's up and running, the system should only have to add a little heat when it gets really cold. 
It's like we built a bicycle for climbing mountains and we're using it to start a hike from the beach. We will get there eventually though.

A hot slab of concrete

Actually it was a rather cold slab, and what we really want is a warm slab of concrete, not a hot one. Well, perhaps a luke warm slab. In fact what we really want is a room temperature slab. 

To be honest, the temperature of the slab is not a direct concern, but we want the temperature inside to be warm in the winter and cool in the summer. Because the concepts of warm and cool are relative, we may reasonably get away with cool being hotter than warm. 18 degrees may be warm in the winter when it's below freezing outside, and 25 degrees cool in the summer when it's over 35 outside.

So we want the slab to be slightly above room temperature in winter, and slightly below in Summer. If we can keep it somewhere between 20 and 25 degrees for the whole year, we should be comfortable.

The slab itself, with fifteen centimetres of concrete at the bottom, ten at the top and 70 centimetres of aggregate between, is mostly going to work as thermal mass, maintaining such a steady temperature,as I wrote before.

We have a 460 litre tank of hot water, which will be heated by an atmospheric heat pump, using cheap nighttime electricity to elevate the abundant but cool heat in the nighttime air to piping hot water. Actually, it may make more sense for us to modify this to take heat from the hot air under the solar panels in the day time, but that needs to be dealt with in a whole new blog. 

The heat pump, known as an Eco cute, is capable of controlling four heating circuits as well as providing domestic hot water and reheating the bath. But, the heating circuits send water at around 60 degrees, and keep sending it until the temperature sensor reads something like 50 degrees in the return pipes. As a 25 degree slab is going to be sufficiently warm, the heating circuits, as they are, are not much use to us. Instead they pass through a thermostatic mixing valve, which can be set to some temperature between 30 and 60 degrees, that will mix a suitable amount of hot water from the boiler to the water coming back from the slab. I suspect we will usually set this as low as possible, although as soon as we move in, we may want to get the slab up from 10 degrees as quickly as possible. 

They filled the underslab water pipes with antifreeze. It will be a very cold day when anywhere near these pipes gets anywhere near freezing, but I suppose there are advantages with protection from rusting, and there may be an increase in heat capacity. The water may need changing every couple of years, but we will see.

The temperature is being logged

Some data loggers have just arrived for the thermometers we put into the slab at a couple of levels. The recording all started back in the middle of winter at the bottom of the foundation, when we added sensors into the rebar.

Then we had all these sensor plugs growing like flowers from the wet concrete.

Most of them survived the layer of aggregate, then another sensor was added in the metal grid for the screed floor. I tried to get them to move the sensors as far as possible from the underfloor heating pipes, outlined in red.

Hopefully the sensor in the middle is around this position, although it may have moved when the concrete was poured.

After the screed floor, there's a few centimetres of wire to the plug. This is at the back of the house where the store room floor is a few centimetres lower and the slab a few centimetres thinner.

To avoid another decapitation, the carpenters very quickly made little boxes to cover the protruding sockets.

Tsukanaka-san arrived from T&D, and fixed the one we broke when the aggregate was poured in. I had tried to fix it, and stripped back the cover of the wire only to find three identical wires inside. I tried to fix them together, hoping for the best, but when Tsukanaka-san and the company president Morizumi-san plugged in the sensor, there was no reading. The president pulled my work apart, and reconnected two of the wires. One is apparently a dummy. He got the right two wires first time!

When the data loggers arrived they had to make bigger boxes.

You can see from the readings that there's already over 2 degrees difference between the temperature at the top of the slab to the bottom. There was also a difference from the west side of the slab, which gets some sunlight in the evening, while the other parts are in the shade. We can track the temperature over the next few months and get some idea of its reaction time, which should help when we start operating the underfloor heating. I'm hoping to some extent we can build up heat at the end of summer, and then release it over the winter so the slab is as cool as possible when summer hits. Probably not enough thermal inertia though. 

Each logger has an index number from 1 to 10, so we have started in the middle of the house with number 1 at the bottom of the foundation, inside the insulation, and number 2 in the concrete floor. Each logger can store 16,000 readings, or around 110 days' worth of data. The batteries last six months, so we need to fix the loggers where we can get to them. They have extension cables, but three pairs are positioned in or under cupboards, one is under the stairs and the other is in the storeroom. Maybe the cupboards should have removable floors or something. The data can be collected by a radio collector, which can then put the information into a computer.

There are fifteen channels, so we can record the room temperature and humidity in a few places, possibly outside as well. Also I was wondering about recording the temperature in the air channel under the solar panels, which I think is going to get quite hot and impair the power generation.