From the very start of this project I have now realised that I have wasted hundreds of hours thinking far too much about things that really don’t need worrying about. In this episode, I’m going to try and focus on a few of them, I’ve no doubt more will come up in later posts.
I mentioned in a reply to Mike in the previous part that I had wasted a lot of time doing things that I thought would be needed for the building control completion inspection, but that it turns out weren’t. The same is true about so much of this build, that I think it’s well worth trying to summarise some of them in a single post.
Firstly, I should point out that as a retired scientist I find it far, far too easy to get sucked in to investigating details and finding out why things behave as they do. Sometimes this is a good thing, but 90% of the time it leads to a great deal of wasted effort. The classic is the initial house design. I was obsessed with how to heat it and making sure it had enough passive solar gain in winter. It didn’t even cross my mind that it might need cooling, neither did it cross my mind during the design stage that having a house deep into a cut out into a south-facing hillside, near the bottom of a fairly sheltered valley, would mean that the micro-climate around the house was significantly warmer than the average for our area, which is on the edge of Salisbury Plain – a notoriously cool place!
One consequence of this is that the local climate data from the Met Office isn’t very representative of the temperatures we get in this valley, and even less representative of the local temperature around our house, set, as it is, 2.5m down into the hillside and sheltered on the North and part of the East by the ground a couple of metres away. This meant that the climate data I used both in my own spreadsheet and in PHPP was in error. My estimate is that in summer our average local temperature is around 2 to 3 deg C warmer than the Met Office data set for the area, and in winter at least 1 deg C warmer, maybe a bit more.
The effect of this is to skew the heating requirements downwards, and create an overheating risk that didn’t show in my spreadsheet, PHPP or SAP.
Curiously, most of the overheating comes from the East, not the South. The sun rises well North of East in summer, so the East elevation has a long period of direct exposure. This shows clearly on the cladding; the larch on the East side has faded a great deal more than that on the South or West faces. The bedroom with a window facing East overheated badly, so this has been mitigated by applying heat reflecting film to the outside of that window and by replacing the very thermally lossy thermal store with the Sunamp PV (the services area is adjacent to the East bedroom).
The kitchen East-facing window also gets a lot of solar gain, as does the South facing one. The dark grey of the stone window internal window cills, and the kitchen worktops, exacerbate this, by heating up and convecting and conducting heat into the room. By contrast, the solar gain from the South facing gable is not that massive. The large roof overhang limits it, but adding solar reflective film has also made a big difference, such that cooling is not often needed now.
I wasted more time thinking about complex heating and cooling control systems than you would believe. I built several versions of rather sophisticated controllers, using internal and external sensors measuring air temperature (inside and outside), floor slab temperature, relative humidity, flow and return temperatures from the floor heating, you name it. I had weather compensation, predictive heat demand from rate of change of outside temperature and very accurate control of the floor slab temperature, which, in a low energy house, is directly related to the heat output at any time and the house internal temperature.
All this was a complete and total waste of time and money. At a guess I spent around 500 hours playing around with different systems, plus a few hundred pounds building ever-more sophisticated controllers. I can’t believe, now, that controlling the temperature of a passive house is really very, very simple. There is no need for any complex controls or expensive valves, sensors, etc. The house will pretty much look after itself when it comes to heating in winter, with just a very simple room thermostat controlling the whole ground floor under floor heating. It’s a complete and total waste of effort, time and money to invest in anything clever or sophisticated to regulate the heating, when a simple and cheap room thermostat works so exceptionally well.
We’re in the heating season now, and have had temperatures down as low as -6 deg C on a couple of nights. The only heating has been from the under floor system, as we don’t ever use the Genvex active MVHR for heating at all – it just isn’t needed and, anyway, it tends to dry the air out a bit too much. The coldest temperature I’ve seen in the house has been about 20 deg C first thing in the morning, the warmest (this December) about 21 deg C. The floor slab sensor now only displays the temperature, it doesn’t control anything, and that shows a steady temperature of around 21.2 to 21.8 deg C. The heating system comes on about once every two to four days, for around an hour, and usually pumps water at about 23 to 24 deg C around the floor slab. Any hotter than this and the house will heat up too quickly, and there will be a noticeable overshoot in the floor, and room, temperature. On very cold nights, the heating will come on in the morning. A couple of weeks ago we had three nights in succession where the heating came on for an hour or so every morning. The room thermostat is on the wall, in the shade, on the ground floor, pretty much in the very centre of the house. It’s currently set to 20.5 deg C, and has a 0.1 deg C switching hysteresis. This means it turns on at 20.5 deg C, and off at 20.6 deg C.
Because there is a lag between the floor slab warming up and the room air warming up, the room temperature will continue to slowly increase through the day as the slab gives up heat, even though the heating is off. This seems ideal, as it means that the temperature tends to peak in the evenings, at around 21 deg C if there’s been no one in the house and no solar gain, then drop very slowly overnight on a very cold day. The rate of change of temperature is so slow that the subjective effect is that the house is at a very constant temperature all day. This isn’t surprising, as the daily room temperature variation rarely exceeds 1 deg C in winter and is often only around ¼ of that, unless the weather is very cold. Our old house, even with double glazing, good loft insulation and cavity wall insulation starts to cool around half an hour after the heating goes off, and loses around 1 deg C every couple of hours if the heating is left off during the day in winter. We often see a 4 to 5 deg room temperature variation in winter in the old house, and it’s very noticeable that the new house just feels far more comfortable even if the heating hasn’t been on for two or three mornings.
So, the moral of this tale is to forget about complex heating controls in the low heat loss house. Forget about zoning, as all the rooms will end up being close to the same temperature, forget about programmers and clever thermostats, as all that is needed is a low temperature heat source into the floor plus a simple high resolution room thermostat situated near the centre of the house.
Had I know this at the start I’d have been amazed. In fact, quite frankly I’d never have believed it. I’ve wasted hundreds of hours trying to solve a problem that turned out to be imaginary!
I’ve just realised that this is all getting a bit too long, and that there are other things I’ve wasted time over-thinking that I’ll need to put in other posts. To finish, these are the diagrams of the floor heating and cooling system. This excludes the Genvex active MVHR, which is separate and only ever used occasionally for summer cooling now; it’s never used for heating in practice, and has it’s thermostat set permanently to 19.5 deg C, and if the floor heating is working then the house never ever gets this cool, so it’s really just a backstop in the event of failure of the main system:
The above diagram shows the ground floor part of the system. The thermostatic valve on the UFH manifold is normally set to around 24 deg C, which is right near the very bottom of its control range. If the weather turns extremely cold (below about -5 deg C or so) then there is a very slight benefit in turning this temperature up by a degree or two, but also a risk that by doing so there will be a temperature overshoot. The system is very sensitive to small changes in flow temperature, and if there was one thing I might, possibly, change it would be to add a very limited weather compensation capability to the flow thermostat. I’m not sufficiently convinced yet that it’s worth the effort, for the handful of days in a year when it might, possibly, give a small benefit. It’s worth noting that a lot of thermostatic mixer vales won’t control down to 23 to 24 deg C or so. Ours is a Wunda manifold, and the mixer valve uses a sensor set into a pocket in the flow manifold, and as far as I know is one of the few that will control well down at such low flow temperatures.
For those that have doubts about the effectiveness of such a simple control system, here is some temperature data from the house logger. The logger collects data every 6 minutes and measures and records lots of different parameters, including room Co2 and room RH, but this plot just shows the room temperature (red) at around 1.2m above floor level in the living room and outside air temperature (blue) measured at the sheltered North face of the house. For the first few days of this plot I had the room thermostat set to 21 deg C, which turned out to be a bit warm, so I turned it down to 20.5 deg C, a change that can be seen around 11th November:
From the above, it’s clear that the outside temperature varies very wildly, whilst the room temperature is relatively stable, not changing by more than about 1 deg C. Most home thermostats can’t regulate to 1 deg C, the best they normally manage is around 2 deg C or so, and, as mentioned before, our old house changes temperature inside by several degrees per day, with the thermostat set to 21 deg C and the gas boiler running for a fair part of each day to try and keep the house around 21 deg C.
The most important single aspect of this whole under floor system, and the one that has the most dramatic impact on the evenness of the house temperature, is the fact that the circulating pump is running all day long, every day (but turned off at night) whether heating is needed or not. This circulates water around the whole floor slab and evens out the temperature, taking heat from the warmer areas and using it to heat to cooler areas. I can’t underestimate how important this is – it really does make a massive difference to the comfort level, all at the cost of less than 20 watts to run the low energy Grundfos pump on the under floor manifold all day.
The diagram above is mainly concerned with the DHW system, which uses pre-heated water from the ground floor buffer tank. I’ve included it here for completeness, and as it shows the expansion vessels and filling loops for the under floor piping and ASHP (filled with antifreeze mixture) and the buffer tank filling loop (filled with normal inhibitor). The buffer tank is run at a very low pressure, 0.3 bar cold and 0.5 bar warm, with a 1.5 bar pressure release valve. This is because the buffer tank isn’t a pressure vessel, and is designed for a working pressure of not more than about 1 bar. This arrangement was used to avoid having a header tank and overflow – it means the system can be sealed, a significant advantage.