Episode 5

Thermal Envelope Interpretation: Where the Insulation Perimeter Actually Sits

1 April 2026 · 49 minutes · Hosted by Brian Haines, Matthew Graham and Andrew Hooper

Summary

Where exactly does the thermal envelope stop? The team open with a disagreement they had already been having off air, about whether a garage attached to a house needs to be insulated at all — and find that the answer turns on a single word in the code.

From there the discussion works outward through the elemental provisions that sit on top of a seven star rating, into thermal breaks and where they belong, and finally into what a continuous insulation barrier actually means when insulation is installed between studs, joists and trusses and therefore is not continuous at all. Along the way it exposes a persistent mismatch between how the National Construction Code and the NatHERS assessor handbook describe the same building elements.

The recurring theme is the gap between what compliance requires and what good building practice would do, and the awkward position assessors are left in when a private inspector on site takes a different view from either.

In this episode

• 0:57What Specification 42 requires on top of a seven star rating

• 2:52Clause 13.2.7 and the attached Class 10a garage

• 5:06The word "or", and why it changes everything

• 7:59Two valid envelopes: wrap the house, or include the garage

• 8:06Ceiling, not roof — a distinction the diagram makes for you

• 10:21Where the thermal break goes once you have chosen the envelope

• 11:44Internal walls to roof space: the same problem in a different place

• 13:39Cladding or lining? The NCC and the handbook use different words

• 18:19What a continuous insulation barrier means in practice

• 19:53External corners, lintels and the spots builders miss

• 21:27Insulation is never continuous — studs, joists and trusses

• 23:00AS 3999, and standards the NCC does not call up

• 26:57Where the private inspector's authority actually sits

• 28:08Compliance, best practice, and what you would do in your own house

• 33:00Why the NCC and the assessor handbook need to talk to each other

• 45:07Thermal bridging as introduced: steel only, ceiling only

Key points

Seven stars is not the whole requirement

Specification 42 clause 4 sets out the elemental Deemed-to-Satisfy provisions that apply on top of a seven star NatHERS rating. The team run through them: thermal insulation requirements at 13.2.2, thermal breaks at 13.2.37 and 13.2.55, compensation for loss of ceiling insulation due to penetrations at 13.2.35 — which NatHERS software handles for you — floor edge and under-slab insulation at 13.2.64 to 13.2.66, and building sealing at 13.4.

The garage clause hinges on one word

Clause 13.2.7 states that a Class 10a building attached to a Class 1 building must either have external fabric achieving the thermal performance required for a Class 1 building, or be separated from the Class 1 building by construction achieving that performance. Brian's initial reading was that the garage ceiling and its external wall must always be insulated. Andrew's position, which the discussion lands on, is that the "or" gives you two complete and alternative options: bring the garage inside the envelope and insulate it as though it were part of the house, or draw the envelope around the Class 1 building and isolate the garage from it entirely. Choose one. If you separate the garage out, there is no requirement to insulate its ceiling or its external walls at all.

The diagram only shows a double storey, and that causes trouble

Figure 13.2.7 illustrates the principle on a two storey building, where the element in question is an intermediate floor. There is no single storey version, so readers have to translate it themselves — and the thing being insulated in the single storey case is the ceiling, not the roof. Several minutes of the discussion are spent making sure that distinction is clear, because it is easy to say roof and mean ceiling.

The thermal break follows the envelope

Once you have decided where the envelope sits, the thermal break location follows. Include the garage and the break belongs on the outside of the garage. Separate the garage out — which the team consider the more sensible choice, since you are not going to seal a garage door — and the break belongs on the wall between the garage and the house. That is consistent with how internal walls to roof spaces are already treated, where the wall is the building envelope and carries both a bridge and a break.

A terminology mismatch that keeps causing arguments

The NCC refers to cladding, which most people read as external. The NatHERS assessor handbook refers to lining, which reads as internal. Table 11.3 in the handbook sets out default thermal breaks and air gaps and includes internal walls to unconditioned spaces, distinguishing a cavity — requiring a break but no bridge — from lightweight cladding direct fixed to the same insulated steel member as the wall lining, or with no wall lining at all. Internal walls to roof space are typically lined on one side only and have no cavity, which is precisely where the wording stops being clear. The team's view is that the handbook is more explicit than the NCC here, and that the NCC would benefit from an internal wall section of its own.

Two kinds of thermal break

A useful distinction drawn late in the episode: there is a physical break, an R0.2 material such as a proprietary strip, timber or polystyrene, and there is an air break, where a cavity performs the same function. That is why a ceiling requires a bridge but not a break — the attic space above it is doing the job, and Andrew puts the value of that space at roughly the R0.2 equivalent. It also explains why a flat roof, where the assembly terminates at the roof sheet, needs both.

"Continuous" cannot be read literally

Clause 13.2.2 requires insulation to form a continuous barrier with ceilings, walls, bulkheads and floors that inherently contribute to the thermal barrier. But insulation sits between studs, between floor joists and between trusses, so it is never actually continuous. Andrew's argument is that the qualifier "inherently contribute" does real work: an external corner where two studs meet has no direct interface with the internal room, and on his reading does not inherently contribute to the thermal shell. Brian's counter is that a private inspector with a thermal camera will see those corners regardless, and that AS 3999 does call for them to be insulated — though that standard is not called up in the NCC, which makes it voluntary to that extent.

Where that leaves an assessor

The team land on a position they can all live with: it is not strictly required, but a builder finishing the job properly would do it, and none of it can be modelled in a NatHERS assessment anyway. Brian's practical point is that the private inspector is the certifying authority on that job, and is perfectly within their remit to raise it. Andrew's is that best practice has to stop somewhere — he did not chase every gap when he built, choosing the big ticket items instead, and notes that expanding foam can restrict the movement a house needs and cause its own problems.

Thermal bridging arrived deliberately incomplete

Matthew describes the introduction of thermal bridging as "thermal bridging light" — applied to steel framing but not timber, and at ceiling level rather than roof level. He would have preferred to wait for a complete introduction. Andrew's counter is that this is how every energy requirement has arrived: four stars, five, six, seven. Introduce it whole and the outcry from a market that is ninety per cent timber framed would have stopped it. Both agree the underlying assumption still being carried — that insulation is continuous — is the elephant in the room.

Documents referenced

• National Construction CodeAustralian Building Codes Board — the code itself, and the Housing Provisions referenced throughout.

• NatHERS publicationsTechnical Note, assessor handbook and scheme guidance material.

Transcript

Lightly edited for readability. Hesitations and repetition have been removed and technical terms corrected; the substance is unchanged.

0:05Welcome to The Rating Room, the podcast where we unpack everything NatHERS. I am Brian Haines, and alongside me are my co-hosts Matthew Graham and Andrew Hooper. Together we dive into the technical, the practical and sometimes the controversial aspects of the NatHERS scheme and everything in between. Where are we starting today?

0:39We are starting with the insulation perimeter. Andrew and I have already had quite a heated discussion about this. Were the cameras rolling? They were not. I do not believe we can do it again.

0:57Let us get back to basics. If you are doing a lot of energy efficiency assessments you will understand Specification 42 — the requirements for performance requirements H6P1 and H6P2. Clause 2 details seven stars for the majority of the country, with the exceptions we know of in BASIX and in Queensland under the Development Code. Clause 3 is the Whole of Home benchmark. Then clause 4 covers the additional elemental Deemed-to-Satisfy provisions you need on top of the seven star rating.

1:51Those items are 13.2.2, thermal insulation requirements. 13.2.37 and 13.2.55, which are your thermal breaks. 13.2.35, the compensation for loss of ceiling insulation due to penetrations — the NatHERS software caters for that, so it is not something you need to think about too much. Then 13.2.64, 65 and 66, floor edge insulation and under-slab insulation, which is the one Andrew and I do not always agree on. And 13.4, building sealing, where you seal gaps and cracks.

2:52What I want to mention came up on a Facebook page — not the NatHERS Assessor Network page, a different one — where an architect was explaining 13.2.7 to me. That is a different elemental Deemed-to-Satisfy clause within Part 13. It talks about the insulation perimeter and where you need to put insulation. What I was trying to point out is that if you are doing a NatHERS assessment to seven stars, you do not have to comply with 13.2.7.

3:38Reading it verbatim: a Class 10a building attached to a Class 1 building must have external fabric that achieves the required level of thermal performance for a Class 1 building, or be separated from the Class 1 building with construction having the required level of thermal performance for the Class 1 building.

4:09There is a diagram, but it is a double storey diagram. It would be nice if there were a single storey one as well. My understanding is that if you were doing an elemental Deemed-to-Satisfy report and had to comply with 13.2.7, you would be insulating the entirety of the Class 10a garage and its external wall. Looking at Figure 13.2.7 now, the insulation perimeter either goes completely around the Class 1 and down the external side of the Class 10a, or it goes across the top of the 10a and down the internal wall between the house and the garage.

5:06Andrew, do you agree that every garage ceiling, whether single or double storey, and the external wall, needs to be insulated? The short answer is no. Look at the diagram — it shows the Class 1 component has a continuous insulation barrier around the home, and the garage is separated away from the Class 1. It does not matter whether it is single or double storey. Even with a single storey, if you insulate the ceiling, insulate the external walls, and insulate the wall between the garage and the house to the same level as the external wall, and do whatever needs to be done to the floor, you have created that continuous envelope around the Class 1 component and isolated it from the Class 10a.

6:18Then what would be the point of insulating the roof of the 10a? If the thermal envelope has completely encapsulated the Class 1, why would they say you need to insulate the roof but not the external wall to atmosphere?

6:51Where does it say you need to insulate the garage ceiling? Not separately — as per the house, because 13.2.7 says the external fabric of the Class 10 needs to be the same as the external fabric of the Class 1.

7:04It is a separate option, because you have the "or" at the end of that statement. You either include the garage and insulate it fully exactly as you would the whole home, or you create a thermal shell around the Class 1 and isolate the garage out of it. So you are not going to insulate the ceiling, because you do not need to — you have separated the roof space from the Class 1 component by insulating the ceiling.

7:40So my understanding then is the internal wall between the house and the garage is insulated, the ceiling of the Class 1 is insulated, and you do not worry about insulation across the top of the 10a or its external wall, because you have separated it from the Class 1. Yes, with a caveat.

8:06Brian, can I just ask — you are referring to roof. Do you actually mean roof, or do you mean ceiling? Yes, I mean ceiling. Not roof. Because this diagram is really showing ceiling. In this double storey case it is a floor. If we imagine a single storey, it would be ceiling, not roof.

8:43I feel like this lends weight to what Andrew was saying, by having it at the ceiling and the wall.

10:21It does lead on to the second question about thermal breaks. If you are insulating the external wall of the garage to encapsulate the entirety of the Class 1 and the Class 10, where is the thermal break? You cannot have a thermal break on the external wall of a garage. Does that mean you do not have to have one on the internal wall between the garage and the house either?

10:46Again, it comes down to where your thermal envelope is, because that is the space you are trying to isolate and keep at a controlled temperature. If you are including the garage, then yes, the break should be on the outside of the garage. If you are putting the division between the garage and the home — which to me makes much better sense, unless you are going to seal your garage doors and everything else that goes with it — then the break should be between the garage and the home.

11:26Which is how every NatHERS assessor would be doing their assessment. It is consistent with the way we treat internal walls that go to roof space: we are saying that internal wall to the roof space is the building envelope, and we put a thermal bridge and a break in it. The roof space is unconditioned space we are trying to separate from.

13:00Would every internal wall between a habitable space and a garage have to be thermally broken? Basically you are going to have plaster on both sides of a stud wall, and that is the thermal envelope of the house, so it is going to have an air gap.

13:39Internal walls are really interesting, because that clause refers to external walls and does not really use the word internal wall. It also specifically says cladding — and for me, cladding is something on the outside of a wall. It says lining. Are we talking NCC or NatHERS? Because there are two different terminologies. That is true. I was referring to the NCC, where I think it uses the word cladding, and you are saying it refers to lining.

14:32This would not be the first time where there is supposed to be alignment but the terminology differs slightly. For me, cladding is external and lining feels internal. Looking at the NCC: between the external cladding and the metal frame, if the wall does not have a wall lining, or has a wall lining fixed directly to a metal frame and is clad with weatherboards, fibre cement sheet or the like. So are we saying "or the like" is plasterboard, where we are referring to an internal wall?

15:07What I am referring to is the definition of the building envelope. Are we making it the wall between the garage and the habitable rooms, or the external wall between the Class 10 and atmosphere? That is at the discretion of the assessor, the designer and the client as to which way they want to treat it. You can have either option — going back to that "or". So it is a job by job decision.

16:51That is why the diagram has those two examples. It just uses the two storey, which throws out the question of single storey, but you have to imagine it without. And you would not have insulation in the ceiling of the garage if that was the case. So it literally either encapsulates the Class 1, if you are separating out the Class 10a, or it has to include the 10a and then it is irrelevant what happens to the wall between the two.

16:59So if you are doing elemental, you need to choose one or the other. Yes. And that is part of elemental being a provisional basis of the code rather than the performance basis in NatHERS — although you can do one or the other in NatHERS as well if you wanted to.

18:19Let us discuss what constitutes a continuous insulation barrier around a house. On a double storey with an intermediate floor, would good practice be to insulate the external wall where those joists are? And where the corners of external walls meet at ninety degrees, there is that little spot on the outside that does not normally get insulated. Is this best practice, or is this NatHERS?

19:58In all those instances there is probably no heat flow as far as Chenath is concerned. There is no heat flow on that external corner. But you are talking best practice. Is it best practice in construction, or AS 3999, which is not called up in the NCC? And if you are doing a NatHERS assessment you cannot specifically put insulation in those areas in the assessment itself.

20:32But when you have a private building surveyor or inspector going around the house, they will one hundred per cent pick those items, as well as insulation not sitting in front of lintels above windows.

21:27I think you are taking continuous barrier a little literally. In a timber stud wall there is no insulation in the studs — it is between them, so it is not continuous. In reality it is never continuous. The same between floors, where insulation sits between the joists, and in the roof between the trusses. It sits between all of the building structure.

22:03But it does say where practical. So describe one of those where it would actually be practical to get continuity. Under the roof it is continuous but compressed. On the external corners where the studs meet you could insulate that corner, but you would have to do it before the wrap goes on. You could also insulate where internal walls meet an external wall, but again before the wrap. It is also three-dimensional heat flow that you would need to model to understand whether it made a difference. I do not know if it makes a difference. Chenath certainly cannot tell you.

23:00I agree. But when a private inspector says these things need to be done because of AS 3999, or clause 13.2.2 which talks about a continuous barrier with ceilings, walls, bulkheads and floors that inherently contribute to the thermal barrier — I think they do contribute. I can talk my way out of it, and have many times, but morally I think it needs to be done.

23:33Morally, probably. Certainly AS 3999 says those spots are to be insulated, but as you said it is not a primary standard called up in the NCC, so the Australian standard becomes voluntary to that extent. If you go back to the NCC as the higher level regulatory document, there is nothing that specifically says you have to do it.

24:07But that line — "inherently contributes" — I find it hard to say that a corner inherently contributes to the thermal shell. Everything about the NCC and the building regulations talks about the internal lining being the start of the envelope. If you are calculating the volume of a house it is from the internal face inwards. So does wall thickness start to come into play, and to what extent does a corner with no direct interface with the internal room come into play? To me it does not inherently contribute.

25:08Then why do we seal gaps and cracks under building sealing? Air leakage is completely different. There would still be some form of air leakage other than the plaster, because the plaster is the only joint in that area.

26:57The reason I am bringing it up is that dozens of times this has come up with private inspectors going out on site and picking these items. They are the certifying authority. Their word is gospel on that particular job unless you want to take it to a higher authority. So they are perfectly within their remit to raise the question.

28:08So morally, ethically, would we insulate, or leave it out? We have come to the conclusion that you do not necessarily have to, but should a builder be doing it anyway? If you want to finish your job properly you are going to fill in all the gaps and holes. I am not convinced there is air leakage in a plastered corner — there is under the plaster, under skirtings, around architraves — but the corner should be finished cleanly.

29:00If you want your house properly insulated then yes, I would fill in those little gaps with expanding foam or whatever. In the same context, at the moment we can say we do not need to apply insulation where there is a stud, a nogging or a ceiling joist. But best practice would be to run a continuous layer across the face of it. So where do you draw the line between best practice for this and not for that?

29:47If you ran around the house with a thermal camera you would see there is no insulation in those corners. You would also see there is no insulation where the studs and the rest of the framing are.

30:37So we agree you do not necessarily have to do it, but you really should as best practice, in a utopian universe. I would want to see it done in my own house. Would I go to that level? Probably not. We built a few years ago and I did not — I found the big ticket items rather than the little fiddly things.

31:16Was it even an option? We did talk about it in terms of expanding foam, which they were not keen on, because that can limit the movement a house needs. Every house is going to move, and if you put expanding foam in you might find your windows do not move in the frame as they should and you end up with cracking. So you have to find the right balance. Would they have objected if I bought a heap of batts and shoved them into the gaps? Probably not — I did not ask, but they were reasonable with us. I just did not feel it necessary. If you are doing passive house you are going to want to fill every spot, because the airtight barrier plays more of a part.

32:51From what I see in the NCC it is not a requirement, but it is good practice. I just do not think these things talk to each other — the NatHERS assessor handbook and the NCC do not talk to each other in this case.

33:11Where I think there is not a direct conversation is with internal walls. The NatHERS assessor handbook breaks down the various elements — floors, external walls, ceilings and roofs — but it also includes internal walls as their own element. Internal walls adjoining a garage, or other spaces, or roof spaces. We are not interested in internal walls around a laundry. When you look at the NCC, I almost feel like there should be an internal wall section of its own. It covers the other building elements, but internal walls are sort of inferred towards the bottom of that section, whereas it is more explicit in the handbook.

34:34From the NCC point of view you have a Class 1 part of the building and a Class 10 part. From a NatHERS perspective you have multiple possible unconditioned spaces, so you can insulate them out — internal walls, external walls, that is fine. But from a thermal break point of view you are only doing the external walls when required, and the internal walls when required. You are also doing subfloors and roof spaces — bridge and break, timber subfloors and roof spaces.

37:20If we look at the NatHERS assessor handbook, chapter 11.4.1, and Table 11.3 for default thermal breaks and air gaps, under internal walls to unconditioned spaces we get two scenarios. A cavity, where a break is required but no bridge. And lightweight cladding, described as direct fixed to the same insulated steel member as the wall lining, or where there is no wall lining. So again we get this mixed language between lightweight cladding, typically referring to an external wall, and lining as an internal wall. And internal walls to roof space are typically lined on one side and not the other, and inherently do not have a cavity.

38:39The wording is what makes it grey. If it said an internal wall adjacent to these unconditioned spaces, such as the floor and the ceiling, I could live with that. We have two different types of break: a physical break, which is an R0.2 and might be a proprietary product, a strip of timber or polystyrene — or an air break. If we look at a ceiling, it has a thermal bridge but does not require a thermal break, and I believe that is because of the air gap between the ceiling and the roof sheet. Whereas with a flat roof it is limited at the roof sheet and that becomes both a bridge and a break. So I see the attic space as the equivalent of the R0.2 — the handbook and the technical notes say 0.16, so round that up.

39:50For that reason I look at a skylight shaft or a bulkhead and wonder whether that has the same impact. It faces into the attic space and does not require the thermal break. For the sake of argument I will say it has to be on the skylight shafts. I think it needs to be there — that is my personal view. I just do not see that as being how it is written. And how about the internal wall to the roof space? To me it should be done. But I do not read that as how it is written. If we are doing it to the internal wall to the roof space, should we not be doing it to the ceiling as well, for consistency? It is the same sort of adjacency.

41:29So what is our recommendation — that the NatHERS assessor handbook has more diagrams, more explanation? Absolutely. This has always been the case. If we go all the way back we had no handbook, we had a technical note, and the technical note did not have a lot of diagrams. Over time the diagrams have become more explanatory, and I think the handbook is essential. I am constantly telling my students, have you read the handbook. I am doing well to get them to read the technical note. It is not perfect, and like all documents including the technical note, they are supposed to be living documents that evolve. This would be a great area to see a little clarification — and I would like to see it flow back the other way into the NCC as well.

43:26We are not going to get it into NCC 2025, but potentially the next one. I do not think it is high on their priority list. They deal with safety, with the essentials. Thermal breaks and bridging are maybe a little less important at this point. Condensation requirements are significant in 2022 and were barely mentioned prior to that, and a lot of that was written around the Tasmanian code — and Tasmania did not even take it up.

44:26Should we have alignment? Yes, I think we should. It would make life a lot easier. We do not want elemental saying you only have to do this and this, but NatHERS saying you can do something a little different. We are dealing with the same shell of the building whichever pathway we take, so the condensation requirements should be the same and the thermal bridging should be the same in both. I would like consistency and equality in both approaches.

45:07Thermal bridging is relatively new to the industry, and to NatHERS, so assessors have not had the chance to live with it or explore it in depth. There probably has not been as much rigorous debate over thermal bridging and breaks as there should be. For me, thermal bridging was introduced as a little like thermal bridging light — not a full introduction. Personally, to do things properly, I would rather have waited for a more complete introduction. We do not do it for roofs, only ceilings. And we only did it for steel frame, not timber. So the elephant in the room for me is that we are still assuming the insulation is continuous.

47:26Is it wrong to bring it in light though? Look at how we started with energy ratings — four stars, five, six, seven. You progress. Maybe the intent is to progress thermal bridging over time, because if you gave them everything in one hit, imagine the outcry, especially with timber framing which is what, ninety per cent of the building market. The builders would scream. To be fair we can test that in some software — HERO is a good example, in non-regulatory mode you can compare timber framing like for like. So it depends a little on the NatHERS tool. But I agree, you have to start somewhere.