Category: Home retrofit

Super-insulating a suspended floor

As I write this my feet and legs feel chilly, despite the cosy slippers and the radiator a metre behind me being so hot that I cannot hold a bare hand to it for more than a second or so. It's cold and windy outside today and the floorboards and carpet that are the sum total of the insulation and draughtproofing in my floor are clearly inadequate in terms of thermal comfort.

There really isn't any insulation under the floor is there, Daddy!?!

Insulating a suspended floor can be a DIY job, but it's often done in a way that performs poorly. There are two common pitfalls. Firstly when fibrous insulation is used (for example mineral wool) it isn't protected from air movement in the solum (the space under the floor, which is ventilated to remove moisture). This reduces the performance of the insulation in a similar way to the wind whipping through a wooly jumper on a windy day - the insulation is excellent but it only works to its full potential when you add a windproof layer on top. The second common pitfall is to use rigid insulation, which is difficult to fit between the floor joists without gaps, and these gaps substantially reduce the real-world performance since air movement can simply short-circuit the insulation (this is called 'thermal bypass'). This problem is present whether or not there is a windproof layer outside the insulation, but is especially severe when there is not. There is some excellent guidance on insulating suspended floors that addresses these pitfalls, and also moisture risk, from Ecological Building Systems, here.

I wanted to do something similar to what EBS suggest in their blog, but I wanted to add more insulation below the floor joists. To achieve Passivhaus levels of thermal comfort I need to get really low U values wherever I can. I'm already compromising a fair bit on the ground floor walls because of insulating them internally, and only adding insulation between the joists would give a similar, relatively poor U value of 0.23 W/m2K. That seemed a shame when there was still a lot of space underneath the joists in which I could add insulation while still maintaining a decent (150mm) depth of well-ventilated solum. Adding insulation beneath the joists also means that they are not exposed to the cold solum, reducing moisture risk to them.

In the past I've seen special joist extenders that look like half an I joist fixed to the existing joists and used to support a board that is used to hold the insulation in. This seems like a good, simple idea, but on further investigation it seemed impossible to get hold of these and the alternative of chopping I joists in half seemed expensive. While playing with some insulation samples I came up with an alternative idea; extend the joists using wood fibre insulation board, and use this to hold the wind-tight breather membrane in place. This would allow me to add an additional 100mm of insulation below the joists, bringing the U value to 0.14 W/m2K.

I'm going to be using loosefill wood fibre insulation since it will fill the spaces well and because it is much cheaper than wood fibre or Jute fibre batts. All of these options are very good from a moisture point of view since they are both vapour open (they allow water vapour to pass through them) and hygroscopic (they will absorb liquid water and allow it to dissipate). This reduces the risk of damage should any moisture find its way into the build up, but care is needed to reduce the risk of this happening in the first place.

No insulation or draughtproofing under the floorboards makes for chilly feet. Thankfully no sign of damp so far in the places I've pulled up boards.

From the three places I've looked under the ground floor, the solum seems very dry, which is a great start, but I'll be taking the following measures to reduce the risk of moisture getting into the floor build up.

Airtight vapour barrier installed on the top (warm side) of the insulation; this stops both vapour diffusion and bulk air movement which if not controlled can bring significant quantities of water into the build up
Wind-tight, breathable membrane installed underneath (cold side) the insulation to ensure any moisture that gets into the build up can escape into the solum, while protecting the insulation from wind-wash
Check there are sufficient ventilation bricks into the solum, and that ventilation paths are not blocked by insulation
Extend the damp-proof course where necessary to make sure that insulation is not touching foundation walls below it
Ensure there is a 150mm clear space underneath the insulation for ventilation to take moisture to escape from the solum

The last few months have been pretty busy getting the building warrant completed and ironing out some complicated design choices about what to do with the first floor walls and roof (for another blog!). Planning permission has been granted and the building warrant is now in. I'm looking forward to getting started on the floor and starting to see some dramatic improvements in the comfort of our house.

What to do with the walls?

Five minute read.

Our house is like many houses in Scotland; from the outside the walls look like white-rendered masonry walls. But in fact the rendered concrete is primarily a weather screen - the roof and floors are held up by a timber frame. There is a ventilated cavity between the timber frame and the rendered concrete block wall.

A plan view of the existing wall (what you would see if you chopped the top off the wall and looked down on it). The insulation (red squiggles) between the studs is very thin, about 15mm at most.

Investigations show that in the case of our house the insulation in the timber frame is extremely minimal (about 15mm thick). What should we do to improve it?

Some of the most obvious solutions don't work very well at all; If we add insulation to the cavity between the timber frame and the blockwork wall then we lose the important function that cavity plays in taking moisture away from the timber frame (this is why it is ventilated) and stopping moisture from the outside from reaching the timber frame. This has already been done, inadvertently, in at least one part of the house where an old doorway has been closed up and insulation simply fitted against the blockwork wall that fills the old doorway. This has resulted in a damp problem, pictured below, likely due to moisture moving through the wall (probably via both bulk air movement and vapour diffusion through the structure) from the inside and condensing on the cold blockwork, and maybe also some rain getting through the blockwork.

This is an old external doorway that has been filled in. Insulation has been fitted up against the external blockwork wall, resulting in a damp patch seen here as darkened insulation at the base of the opening. This is probably due to moisture condensing on the cold blockwork and trickling down to the bottom of the cavity.

What about adding external wall insulation to the blockwork wall? This is an attractive option since it would involve no loss of space and minimal disruption. But that insulation would be outside the ventilated cavity. If ventilation to the cavity is maintained then the insulation will be almost completely useless since cold air will be able to bypass the insulation straight to the cavity. If the ventilation is blocked then we risk a moisture problem in the wall since that would remove the current route that moisture takes to escape the wall. Damp in structural walls is never a great situation, but it's especially worrying in a timber frame house, where the structural integrity of the house depends on the timber staying reasonably dry.

If neither of the two obvious and simple solutions to improving the thermal performance of the walls will work, what are our options?

Initially, to avoid losing space internally, because it would give better U values, and because it would be less mess internally, I was keen to find a solution that we could do from the outside. We developed a really good detail that involved removing the blockwork wall, removing the sheathing board on the outside of the timber frame, adding insulation between the studs, adding an airtight sheathing board, adding a lot more insulation outside of this and then cladding with a timber rainscreen. The only one small problem with this strategy was the cost. The quantity surveyor estimated that this option was going to be £33k more than insulating from the inside!

So that left us to come up with a detail for insulating the wall from the inside. In the end we came up with something similar to that suggested for timber frame walls by Chris Morgan (who's working for John Gilbert Architects on this project) in his excellent guide to domestic retrofit (detail from page 80 onwards). I'm actually really pleased with this detail now. The U value will be quite a lot higher (about 0.22 W/m2K instead of about 0.14 W/m2K, so about 60% more heat loss through the walls) than the previous strategy, and some of the thermal bridging details will be trickier, but I'll be able to do a lot of this work myself (which I wouldn't have been able to do for the 'from the outside' plan), it won't involve removing the perfectly functional external blockwork and it will be a good opportunity to redecorate the rooms. Here's the plan:

Remove the existing plasterboard and insulation
Insulate between the studs using woodfibre or jute insulation batts. Using insulation batts, rather than rigid insulation boards, allows the insulation to be fitted tight on all sides, eliminating gaps that can dramatically worsen real-world performance. Wood fibre and jute insulation are both vapour permeable (meaning water vapour can move through them) and hygrosopic (meaning they can wick liquid water so it can disperse instead of causing damage). These properties are important since, by insulating so well, we are reducing the amount of heat flowing through the wall that can dry out damp areas (although we're reducing the risk in other ways, see below). They're also ecological options, with low embodied carbon emissions (the carbon emissions associated with manufacturing, transporting, installing, disposing of, etc.) and low toxicity.
Add another 40mm of woodfibre insulation board, fixed to the studs in the walls
Add airtightness and vapour-barrier membrane on the inside of the insulation board. This will be an 'intelligent' membrane that will allow drying to the inside when conditions permit. Getting the airtightness and vapour control right is crucial to reducing the moisture risk to the wall, see this excellent blog by my former MSc lecturer for a good explanation of why.
Add battened and insulated service cavity
Add new plasterboard to finish the walls

We may have to add another sheathing board internally (dependent on an assessment from the structural engineer following finalisation of what we're doing in the rest of the house), and we might have to do some remediation to the existing sheathing board to prevent wind passing through the new insulation (wind washing), but this is the gist of what we'll do.

Here's how the proposed wall will look, in plan view:

We'll be insulating the existing stud wall properly, then adding more insulation, and an airtightness line, inside of that.

In total this solution will 'cost' us about 100mm of space internally on each external wall on the ground floor. Not a huge amount, but the house already feels small so I don't want to lose more than this. This plan will radically improve the comfort of the downstairs rooms - improving the U values from over 1.3 W/m2K (current wall, which likely performs considerably worse than this in reality due to poor installation) to 0.2 W/m2K will increase the internal surface temperature during cold weather (0°C outside, 20°C inside) from 16.5°C to 19.5°C. For an explanation of why this is important see my blog post on thermal comfort Why do I feel chilly? We'll lose a little space downstairs, but we'll be adding space upstairs. That will require a different solution for the first floor walls, a topic for a future blog post!

What are we starting with?

Five minute read.

The plan for our house is to retrofit it to meet the rigorous EnerPHit standard, the Passivhaus standard for existing properties. But before starting that it's important to have a good idea of how the home works (or doesn't work!) at the moment, so as to understand what will be needed to reach the levels of performance required by the EnerPHit standard.

Our house, pre-retrofit, from the front.

Our house is a small (about 84m² treated floor area, a Passivhaus measure of useful floor area), 1.5 storey, three bedroom timber frame house built in 1975. Without doing any investigation of the fabric I know that we use about 1,100 litres of heating oil, costing about £600 per year and emitting over three tonnes of CO₂each year. Almost all of this oil is for heating the house - the oil boiler does do hot water but it is not very effective meaning we usually wash our hands with cold water, fill the kitchen sink from the kettle and the bath from the electric shower. That's three tonnes of CO₂ for a house that we deliberately run colder than we would like (thermostat set at 18°C), that feels uncomfortable even when the heated to 20/21°C (Why do I feel Chilly?) and that suffers from damp and mould problems.

So why is the house so inefficient? The EPC says the walls and roof are insulated, and there's pretty new double glazing on all but one window. Let's do some investigations...

The EPC for our house (produced before we bought the house) assumes that the loft and walls are pretty well insulated. 4 out of 5 stars sounds pretty good, right?

First of all let's look in the loft. The EPC for that assumes that there is 200mm of insulation throughout. When you pop your head into the loft it indeed looks like there is relatively new, thick insulation on the flat (because the house is 1.5 storeys some of the roof insulation is sloping), go a metre or so in either direction from the loft hatch and the insulation changes and thins to about 15mm. I suspect someone has been paid to insulate the loft and has deliberately bodged it to deceive either an EPC assessor, the home owner, or both. The construction industry is crazy.

Thick insulation next to the loft hatch, not so thick beyond there...

So we've got 15mm of insulation in the loft, what about the walls, which the EPC also says are insulated? Looking at the gables in the loft it looks like the same insulation as is in the loft continues down into the timber stud walls, and removing a few bits of plasterboard in key places confirms this - 100mm deep timber frame but with only ~15mm of insulation. Couple this derisory thickness with the fact that the insulation is fibrous, in many places has no wind protection and is in a very leaky house (more on this below) and you end up with a house that is performing a long way below the assumptions in the EPC.

Current loft at the west gable. 15mm of insulation above the ceiling and the same in the studwork walls (to the right here)

Thin insulation in the timber stud wall, also stopping short of the timber stud.

This is looking into the roofspace at the eaves. This space is well ventilated (as it should be), but the insulation visible is fibrous and very thin. The wind washing of this insulation will render it almost completely useless during windy weather (think fleece or wooly jumper on a windy day with no windproof layer on top).

What about under the floorboards? Well the good news is that the EPC was, for once, extremely accurate here; there is no insulation whatsoever, nor any draughtproofing, under the floorboards. Remove a section of floorboard and you can feel a strong breeze, this ventilation is critical to maintaining the timbers used in the floor, and we'll need to be careful with the design of retrofit measures here to not increase the moisture risk to these timbers. No insulation, and air able to enter or leave the house directly through gaps in the floorboards explains why, even with thick carpets, the floor feels cold most of the time.

We also had an old friend Ben Wear (who you can find at Ben@skyedesign.co.uk) do a pressure test to establish how airtight the existing building is. Not suprisingly it's super leaky. Depressurise the house and all the carpets lift (as air comes in from the underfloor space), and in some places the air just pours in. What's perhaps more surprising is that despite being leaky our house is not what I would call 'well ventilated'. Bedrooms are stuffy in the morning, some cupboards suffer from damp and mould and it is very difficult to dry clothes indoors, even in summer, unless the heating is on. All this despite our house being much more leaky than is sufficient for trickle vents+extractor fans to be deemed an acceptable method of ventilation by building regulations. Our pressure test result was somewhere around 15 m3/m2/hr at 50 pascals, three times the 5 m3/m2/hr above which natural ventilation is deemed acceptable. The video below shows considerable leakage from under the kitchen sink (where services come into and leave the house), from above a sliding door and through the window/door seals.