External walls: What would we build today?
Updated: Sep 10
CREU Cartref Technical Extra
If you walked onto one of our live timber-frame projects today and opened up the external wall, there is a good chance you would find quite a lot of a rigid, closed-cell thermoset plastic foam board 'polyisocyanurate' or PIR for short.
There is a reason for that.
It performs thermally.
It allows us to keep walls relatively compact.
And, importantly, there are established timber-frame systems with published evidence of fire resistance that give designers, contractors and manufacturers a clear technical route to follow.
So this isn't a story about something being wrong with the walls we are building today.
It is more a question we have been asking ourselves:
If we were starting again, would we use quite so much PIR?
As part of the technical development behind CREU Cartref, we have been looking at exactly that.
Not because we think we have found the answer.
But because we think the external wall is a good example of the sort of technical question that deserves a little more attention than simply:
“What U-value does it achieve?”
Where we are today — a PIR dominant wall - Lets call this option A for the purpose of this Study
On current live schemes we are using wall constructions based around a PIR-dominant timber-frame approach.
There is plenty to like about it.
Thermally it is extremely efficient.
In our comparative study, a nominal 140 mm timber frame fully insulated with PIR, together with continuous PIR across the frame, could achieve approximately:
25 mm continuous PIR → 0.149 W/m²K
40 mm continuous PIR → 0.133 W/m²K
50 mm continuous PIR → 0.124 W/m²K
So if the brief is:
high thermal performance + compact wall thickness
This type of construction makes a lot of sense.
It also has the advantage of sitting within an established Structural Timber Association (STA) Pattern Book evidence route, provided the final wall remains within the published system requirements and limitations.
The Structural Timber Association is the UK trade body representing the structural timber sector. Its technical guidance, research and testing are widely used to support confidence in timber-frame construction, and its Fire Safety in Use Pattern Book sets out tested wall and floor systems with defined fire-resistance parameters. The STA also operates the STA Assure quality assurance scheme for member manufacturers.
Wall options B, C and D within our study were all developed from existing STA Pattern Book systems, which gave us a much clearer evidence base from which to compare the different insulation strategies.
But there are trade-offs.
PIR has a high carbon footprint during manufacturing, relies on expensive raw chemicals like MDI, and cannot be easily recycled, meaning old boards usually end up in landfills.
Rigid insulation between timber studs needs careful fitting.
Small gaps matter.
Manufacturing tolerances matter.
Inspection matters.
And you end up using a significant quantity of rigid petrochemical insulation.
None of those points makes the system a bad one.
But they did make us wonder whether we could achieve a better balance.
Can we use less PIR without giving up the evidence?
That became the interesting part of the exercise.
Rather than immediately jumping to a bespoke wall, we deliberately looked first at other existing evidence-backed systems.
The question became:
How far can we reduce our reliance on PIR while still retaining a recognised technical evidence route?
Two options stood out.
Option B — Full-fill mineral wool
The simplest move is to remove the rigid insulation altogether and use mineral wool between the timber studs.
It is an appealingly simple wall.
Mineral wool is familiar, forgiving around timber-frame tolerances and straightforward to install. There is no rigid insulation to cut accurately around every stud, and importantly, there is already an STA Pattern Book route for this general wall architecture.
Our first calculations were deliberately conservative and gave no additional thermal credit to the reflective membranes or the airspaces alongside them. On that basis, the repeating timber bridge through the structural frame had a significant effect, and the frame needed to become increasingly deep to approach the performance we were looking for.
But when we reviewed this against an existing calculation, the picture became more interesting.
When you take into account the enhanced thermal resistance that can be achieved where properly specified low-emissivity membranes face controlled airspaces. Applying equivalent assumptions to our comparative Option B calculations gives indicative results of approximately:
140 mm frame → 0.19 W/m²K
184 mm frame → 0.16 W/m²K
195 mm frame → 0.15 W/m²K
220 mm frame → 0.14 W/m²K
These figures remain subject to confirmation through a project-specific calculation, particularly around membrane emissivity, cavity conditions and the actual timber framing fraction, but they materially change how we view the option.
The timber bridge has not disappeared. The studs still pass through the full depth of the insulated zone. But the wall is now working as more than simply timber plus mineral wool. The membranes, service zone and controlled airspaces are contributing to the overall thermal strategy.
That makes the middle ground particularly interesting.
A 195 mm frame potentially gives us a wall around the 0.15 W/m²K level without rigid insulation, while a 220 mm frame could move towards 0.14 W/m²K.
Going deeper again could theoretically take a mineral-wool-only wall towards 0.13 W/m²K, but at that point another question starts to matter: how deep do we really want the structural timber frame to become?
Removing PIR is attractive, but replacing it with significantly more timber and a substantially thicker wall is not automatically a better solution.
So rather than ruling this option out, we now see it as a credible alternative worth developing further — particularly if our ambition is to reduce the amount of rigid insulation within the platform while staying within an established STA-certified wall architecture.
The question is no longer simply whether a mineral-wool wall can perform.
It is where the best balance sits between thermal performance, wall depth, material use and simplicity.
Option C — Keep the mineral wool, but deal with the timber bridge
This is where we landed on what we think is the most interesting option today.
Instead of choosing either mineral wool or PIR, use each material where it does the most useful job.
Mineral wool goes between the studs.
A smaller quantity of continuous rigid insulation deals with the timber bridging across the frame.
This also retains an existing STA Pattern Book evidence route where the construction remains within the published system requirements.
And thermally, the difference is significant.
With high-performance mineral wool within a 140 mm frame, allowing for the contribution of the low-emissivity membrane facing the protected internal service void, our comparative calculations indicate approximately:
40 mm continuous PIR → 0.15 W/m²K
60 mm continuous PIR → 0.13 W/m²K
That is particularly interesting because it means we can reduce the quantity of PIR without simply replacing it with a much deeper timber frame.
We are using mineral wool where its tolerance and non-combustibility are useful, retaining rigid insulation where it has the greatest thermal benefit — continuously across the timber bridge — and making use of the enhanced thermal resistance created by the low-emissivity membrane and service zone.
The result is a relatively compact wall that still performs strongly, but with less dependence on rigid insulation.
And then something else became apparent.
The 140 mm frame starts helping elsewhere too
The attraction of Option C isn't just thermal.
A nominal 140 mm structural frame is a very useful construction dimension.
At ground-floor level, for example, it can coordinate neatly with a standard 140 mm block below the sole plate.
That sounds like a small thing.
But anyone who spends time developing construction details knows that small dimensional decisions have a habit of repeating themselves everywhere.
A rational 140 mm timber frame potentially gives us:
simpler sole-plate and masonry alignment;
a more conventional ground-floor bearing condition;
fewer bespoke masonry dimensions;
reduced overall wall depth;
easier slab-edge and foundation coordination;
standard foundation widths;
simpler opening details;
cleaner intermediate-floor interfaces;
easier eaves and verge coordination; and
more consistent setting-out through the model and drawing package.
None of those individually decides the wall.
Collectively, they start to make the system feel more coherent.
This was probably the point where we stopped looking at Option C simply as a U-value solution and started seeing it as a whole-wall strategy.
But what if we want to go further?
So far we have been talking about solutions where there is already a relatively clear published evidence route.
But we also wanted to ask a different question.
What might the wall look like if reducing rigid insulation became a stronger design objective?
That took us back towards two deeper-frame ideas.
Tai Ar Y Cyd Pattern Book Wall - Bio-based deep-frame construction - We will call this Option D for the purpose of this study.
Our original reference construction uses a substantially deeper timber frame with bio-based insulation and an insulated service zone.
There is a lot to like about that material philosophy.
More timber.
Bio-based insulation.
Less reliance on rigid petrochemical insulation.
And a construction approach that places embodied impact much more visibly alongside operational performance.
But there is an important technical distinction.
For the purposes of our study, this wall does not automatically benefit from the STA Pattern Book fire-resistance evidence route used by Options A, B, and C.
So although it provides an important reference and a potentially attractive direction, we shouldn't treat its evidence position as equivalent without further verification.
What if we develop that idea using mineral wool?
Our Option E is really an evolution of the same deep-frame philosophy.
Rather than bio-based insulation, the concept uses high-performance mineral wool within both the structural frame and the service zone.
Thermally, it starts becoming genuinely interesting.
Our comparative study suggests that an all-mineral-wool version of the wall could achieve approximately:
195 mm structural frame + 50 mm insulated service zone → around 0.15 W/m²K
and, with a deeper frame:
235–240 mm structural frame + 50 mm insulated service zone → around 0.13 W/m²K
So from a purely thermal perspective, an all-mineral-wool wall appears entirely achievable.
And there is a compelling material logic to it:
predominantly non-combustible insulation;
good tolerance around timber-frame construction;
no PIR between the studs;
no continuous PIR layer;
and a relatively simple, consistent material palette.
The trade-off is depth.
To reach the same thermal performance as the hybrid Option C wall, we need considerably more mineral wool and a substantially deeper structural timber frame. That brings other considerations with it — more timber, thicker walls, deeper reveals and potentially greater implications at floor, foundation and roof interfaces.
And then there is the question of evidence.
This particular arrangement does not currently align directly with the published Pattern Book systems supporting Options B, C and D. Before it could become a standard CREU Cartref solution, it would therefore need a clear evidence route through assessment, testing or manufacturer-specific certification.
That doesn't make the idea wrong.
It simply means it is not yet resolved.
And those are two very different things.
So where does that leave us?
There is a fairly clear progression through the options.
What we are commonly building today
A PIR-dominant wall.
Compact, thermally efficient and supported by an established evidence route, but reliant on significant quantities of rigid insulation and tight installation control.
The simplest lower-PIR move
A mineral-wool-filled frame with no continuous insulation.
Robust and evidence-backed, but difficult to push to high thermal performance without making the structural frame significantly deeper.
What currently looks like the best balance
A 140 mm mineral-wool-filled structural frame with continuous rigid insulation.
Still evidence-backed, significantly reducing the amount of PIR compared with the full-PIR approach, while giving us a rational structural timber frame and strong thermal performance.
Where we would like to keep exploring
Deeper timber-frame constructions using bio-based or high-performance mineral-wool insulation, potentially eliminating PIR altogether — but requiring further work to establish an equivalent evidence position.
What would we build today?
If we were starting a new timber-frame housing scheme today, based on the evidence we have looked at so far, we would probably begin with:
Wall Type Option C: 140 mm Timber Frame + High-Performance Mineral Wool + Continuous PIR
Not because we believe it is the perfect wall.
We don't.
It is because right now it seems to sit in a useful middle ground.
It reduces our reliance on PIR without abandoning it completely.
It retains a published evidence-backed system.
It performs thermally.
It accommodates normal timber-frame construction tolerances.
And the 140 mm structural frame starts creating useful efficiencies in the details around it.
That feels like a sensible place to be today.
But it isn't where we want the conversation to stop.
Where could this go next?
The option that probably interests us most long term is not necessarily the one we would specify tomorrow.
Could an all-mineral-wool system secure the evidence it needs?
Could a bio-based wall achieve the same performance within a more rational structural depth?
Could new insulation materials change the calculation completely?
Could manufacturers develop tested systems that allow us to reduce rigid insulation further?
Could better embodied-carbon information shift the balance away from the solution that currently looks strongest thermally?
We don't know yet.
And that's partly why we're sharing this.
CREU Cartref isn't intended to be a fixed catalogue of construction details that we decide once and never question again.
The useful bit is the process:
understand what we are doing now, challenge it, test the alternatives, retain what works and keep watching for something better.
We'd like to hear what others are doing
There is an enormous amount of knowledge already sitting within RSL technical teams, timber-frame manufacturers, contractors, product manufacturers, warranty providers and consultants.
And external walls are one of those subjects where everybody seems to have learned something — often the hard way.
So we're interested.
Are you still favouring full PIR?
Are you already moving towards hybrid mineral-wool systems?
Would you accept a deeper wall to remove PIR entirely?
Have you found an evidence-backed bio-based solution that we should be looking at?
Or is there another approach that makes all five of these options look dated already?
Our answer today is Option C.
We're quite comfortable with the possibility that somebody might change our mind tomorrow.
If you're working through the same questions, we'd genuinely like to compare notes.
And keep following the CREU Cartref Technical Extras.
Comments