I am hoping someone can give some advice here, as I can't find any anywhere else!

I am fully aware that gaps around modern fire doors should be 3mm (+/- 1mm) and this is because the intumescent strip and smoke seal are rated for gaps that size.

However, I conduct a lot of fire risk assessments for residential blocks of flats built in the 1970s and 1980s where the CP3 standard applied, requiring 25mm rebates. The requirement for that standard was the door to be flush against the rebate, to prevent smoke and fire from getting past.

In accordance with the LGA guidelines, and also BS 9794 (Clauses 9 and 10.4) I am happy to accept these doors in principle, including the two hinges, on the basis that they are fully self-closing.

However, my questions is about the gaps around, and what are the tolerances permitted, but nobody seems to know about this.

It seems to me, that the 3mm requirement should only apply to fire doors where there is a reliance on an intumescent strip and smoke seal, so what tolerance should I accept for these "notional" fire doors? Would say a 5mm gap be acceptable? Or maybe larger one, provided the door frame fits tightly against the rebate (ie up to 25mm)?

Thanks in advance for any advice

Before the use of intumescent edge seals fire door tests were carried out to early versions of the BS476 test. The test preceeding the 1987 Part 22 test, we still use today, was BS 476 Part 8 of 1972 and door leaf edge to door frame gaps were 3mm without using intumescent edge seals. There has never been an absolute requirement for a 3mm (or indeed 4mm max gap) but test results have been successful using that gap on a first principles basis and guidance in BS 4787 'Internal & External Wood Doorsets' Part 1 Table 1 recommends a 1.5mm to 3mm gap but these are not necessarily fire doors and scope for adjustment is included to accommodate smoke seals. However, a very limited number of fire doors are permitted with larger gaps.

So two things to bear in mind:

1) Where smoke seals are necessary these are sometimes fitted to the edge of the door leaf and will usually be effective at 2mm to 4mm. But there would be a gap that could be exploited by cold smoke if the gap is larger. Where compression type smoke seals are used to seal between the rebate stop and the face of the door leaf, this larger gap should not cause a problem.

2) BS 8214: 2016 section 9.5.2, without making explicit reference to intumescent seals, states that a typlical gap to achieve good fire performance is between 2mm and 4mm. Therefore, it is possible that larger gaps may adversely affect fire performance. These gaps refer to test conditions though.

Edited July 10Jul 10 by Neil ashdown
To add more info

Hi Neil

I have reviewed BS 476 Part 8 of 1972 and I see that clause 7.1.2 reads: "the gap shall be at least 3 mm between door edges and the frame."

So why can the door not have a 4mm gap, 5mm or 6mm?

Moreover, I just don't see the logic of the door needing to have a particular gap, as long as the gap is not so big to 'miss' the rebate and that the door sits flush against the rebate.

Daniel

Gaps around the door are prescribed by the door manufacturer following a burn test. Anything outside the burn test, the door may not perform as expected giving less that the required protection.

As Mike said, the door leaf edge to door frame gaps are specified by the door leaf manufacturer and based on evidence from fire resistance peformance tests to BS 476 Part 22 and BS EN 1634-1. Commonly, the gap is 4mm maximum for timber-based fire doors.

BS 8214: 2016 section 9.5.2 notes that failure of fire resisting doors under fire performance test is often due to burn through at the gap between the door leaf edge and the door frame. A typical gap for good fire performance is between 2mm and 4mm.

There is no guidance, as far as I am aware, that suggests the gaps can be larger for older fire doors already installed in the built environment. Although many existing fire doors do have gaps outside the recommended 2mm to 4mm tolerance. UK Gov guidance may be useful for example at https://www.gov.uk/government/publications/fire-safety-england-regulations-2022/fact-sheet-fire-doors-regulation-10 and https://www.gov.uk/government/publications/making-your-small-block-of-flats-safe-from-fire The Responsible Person has a duty to make sure the fire doors are capable of providing adequate protection.

thanks Neil

I hate to be argumentative, but BS 8214: 2016 relates to modern fire doors; indeed the note that you refer to under 9.5.2 (which is not a normative standard) has text underneath it which relates to intumescent strips and smoke seals. It's clear to me that this standard is not appropriate for assessing period fire doors,

Darren Minton's article in the RICS journal is interesting for me; whilst it doesn't give specific guidance on door gaps, it does warn of the dangers of assessing old doors against modern standards.

The conclusion I am coming to is that there are no standards for door gaps for period doors, and it's just left up to the fire risk assessors judgement...which doesn't really help me much, as a fire risk assessor!

So given the lack of an independent standard, I am forced to think through the logic of the period door "technology", in that the critical element is that the door face needs to form a strong seal with the rebate. And on that basis, unless anyone can show me evidence otherwise, I plan to accept doors with bigger gaps.

Basically, as a Fire Risk Assessor, you can dictate what ever gap you feel that you can justify.  There are many ways to make a building fire safe, doors are only item in the equation, but one of the most important for protecting the escape routes.

However, you may find yourself in front of the man with the funny wig trying to explain why you went against the consensus of 2 – 4mm.

"I hate to be argumentative, but BS 8214: 2016 relates to modern fire doors"

No problem at all with that but, the point is that BS 8214: 2016 in the context of the 2mm to 4mm gaps, refers to doors being subject to the BS 476 Pt 22 or BS EN 1634-1 fire resistance performance tests. Its not that the doors are modern. Old doors and new doors are both made from timber-based materials and fire conditions in reality might not follow the same fire curve as the standard fire resistance performance tests.

I've dug out data from the 1951 fire door standard and even back then the normal fit for a fire door was a gap of up to 3mm with the door closing on a stop of 25mm (no intumescent back then).

To say the gaps don't matter because it's a pre BS8214:1990 (first edition date) door fails to appreciate there have been fire door standards for 75 years which were also very specific on measurements

Hi Ant

It's good to hear from you!

And I see now that I was wrong on the door gaps for period doors. But the research has been useful for me, and I feel I have more "backup" when I am writing my recommendations on my FRAs. Thank you.

I couldn't find BS 459 Part 3:1951, but I did find a really useful Fire Research Note no 855 from 1971 by W A Morris "An investigation into the fire resistance of timber doors" showing the critical importance of the 3mm (or less) gap in doors with 25mm rebates.

Neil, thank you, too, for your advices

Daniel

Hi Guys,

This thread has been an interesting read, and i am hoping you might be able to shed some additional light for me on fire doors gaps in both modern/period doors as i have some confusion around the actual physics and what exact gap we're supposed to measure.

The guidance re the gaps of 2-4mm - where is this actually measured from? For a door without a rebate, the smoke channel is between the door and frame itself so the gaps to measure are straightforward and that all makes sense to me.

But in a door with a rebated frame (1inch), when the door is closed, surely the relevant gap that would allow for smoke to leak through is the gap between the rebate and the door, not simply the door and the frame?

So, for example, a flat entrance door - if i were to measure the gap from the inside of the flat, the gap between the door leaf and frame is 5mm so outwith recommended tolerance levels, but, from the outside of the flat, where the door closes flush on to its rebate, the gap between the door and rebate is 1.5 mm. So - which gap should i be taking into account? If i look at it from a physics perspective, and the spirit of the guidance in that we are trying to prevent smoke leakage, the relevant gap that creates the smoke pathway is through is the 1.5mm gap between the door and rebate. Yes, there is a smoke channel of 5mm on the other side, but the smoke then hits that rebate and still has to funnel through the 1.5mm gap? Why can i accept a 3mm open-channel on a non-rebated door, but not accept a 5mm gap on a door with a 1inch rebate and an actual smoke channel of only 1.5mm when the door is closed?

I've tried to research this, i have spoken to various fire door surveyors, watched youtube videos - but i have not received a straightforward answer. I have seen fire door survey records where they've marked nearly all of the flat entrance doors within the block as non-compliant according to their checklists due to gaps between the inside of the door and the frame - but they could not explain why that gap applies when there is a rebate on the other side, essentially blocking the smoke pathway entirely and met with a kind of shrug and 'this is what our checklist says'.

From my understanding, prior to the introduction of intumescent strips/cold smoke seals - the rebates themselves were in fact considered the smoke seal?

In older UK fire doors, the purpose of a rebated frame (where the door stop is machined directly out of the solid timber frame rather than being a separate "planted" stop) was to create a deeper, interlocking, and more robust barrier against fire and smoke.

Key Purposes

Longer Flame Path: The stepped edge (rebate) forces flames and hot gases to travel through a tortuous, overlapping path rather than a straight line, slowing down fire penetration.

Structural Integrity: Machining the frame from a single solid piece of dense timber provided better stability under extreme heat, reducing the likelihood of the door stop warping or burning away as quickly as a small, nailed-on wooden strip.

Inherent Seal (Pre-intumescent era): Before modern chemical intumescent strips (which swell up in heat) became standard, older fire door designs relied on the physical depth of a heavy timber rebate—sometimes 25mm or more—to mechanically block smoke and fire around the perimeter edges.

(Information from AI, but with reference links that were genuine sources so not AI hallucination)

On 02/09/2026 at 14:02, Guest Guest82 said:

Hi Guys,

This thread has been an interesting read, and i am hoping you might be able to shed some additional light for me on fire doors gaps in both modern/period doors as i have some confusion around the actual physics and what exact gap we're supposed to measure.

The guidance re the gaps of 2-4mm - where is this actually measured from? For a door without a rebate, the smoke channel is between the door and frame itself so the gaps to measure are straightforward and that all makes sense to me.

But in a door with a rebated frame (1inch), when the door is closed, surely the relevant gap that would allow for smoke to leak through is the gap between the rebate and the door, not simply the door and the frame?

So, for example, a flat entrance door - if i were to measure the gap from the inside of the flat, the gap between the door leaf and frame is 5mm so outwith recommended tolerance levels, but, from the outside of the flat, where the door closes flush on to its rebate, the gap between the door and rebate is 1.5 mm. So - which gap should i be taking into account? If i look at it from a physics perspective, and the spirit of the guidance in that we are trying to prevent smoke leakage, the relevant gap that creates the smoke pathway is through is the 1.5mm gap between the door and rebate. Yes, there is a smoke channel of 5mm on the other side, but the smoke then hits that rebate and still has to funnel through the 1.5mm gap? Why can i accept a 3mm open-channel on a non-rebated door, but not accept a 5mm gap on a door with a 1inch rebate and an actual smoke channel of only 1.5mm when the door is closed?

I've tried to research this, i have spoken to various fire door surveyors, watched youtube videos - but i have not received a straightforward answer. I have seen fire door survey records where they've marked nearly all of the flat entrance doors within the block as non-compliant according to their checklists due to gaps between the inside of the door and the frame - but they could not explain why that gap applies when there is a rebate on the other side, essentially blocking the smoke pathway entirely and met with a kind of shrug and 'this is what our checklist says'.

From my understanding, prior to the introduction of intumescent strips/cold smoke seals - the rebates themselves were in fact considered the smoke seal?

A flat fire door has two purposes: to prevent a flat fire from compromising the fire escape and, if there is smoke/fire in the common areas, to allow the residents of the non-affected flats to remain safe in their flats (typically for 30 mins). Assessment from both sides is obviously important.

The primary fire safety feature of the old (pre-1991) fire door was the hardwood rebate which is designed to form a seal between the door and it's frame. You talk about a 1.5mm gap, that implies that the door or rebate has warped and the door is not closing snug against the rebate. If that's the case, then the action must be to remedy this. I'm not a chippy, so not an expert on remedial actions, but I would have thought that if the gaps were small then it would be possible to undertake some careful planing to ensure the door sits against the rebate all the way round, but most likely, the door leaf will need be replaced.

So to summarise and answer your question directly, the reason why you can accept a door with a 3mm gap whose leaf fits snuggly against the rebate, is that the functions of the door is met, in that there is a good door seal and the gap around is within tested tolerances. With the 1.5mm gap to the rebate and 5mm gaps around, then the door does not satisfy either of the two standards.

In the fire door sector, doors are currently tested to the Standards BS 476 part 22 and/or BS EN 1634-1. BS 476 is to be withdrawn from Approved Document B of the building regulations in September 2029 in respect of fire doors.

In the tests above, doors are tested for integrity without smoke seals fitted. Intumescent seals are fitted but failure at the gap between the door leaf edge and the door frame is a common reason for the door not to perform as intended in the test. Therefore, in most cases the maximum door to frame gap is 4mm.

In tests before the days of intumescent seals, the 3mm gap was more critical and failure at the gap between the door edge and door frame was more common. That's why doors were known as Fire Check rather than Fire Resisting. Anthony has earlier made reference to the1971 investigation into fire door performance, its an enlightening read so here's the link https://publications.iafss.org/publications/frn/855/-1/view/frn_855.pdf

In the case of restricted spread of smoke at ambient temperature, smoke under pressure is forced through the gap between the door leaf and the rebate stop and then through the gap between the edge of the door leaf and the door frame. This is why propriety smoke seals are availble to fit to the rebate stop of the door frame as well as combined with the intumescent seal for sealing of the gap between the edge of the door leaf and the door frame.

So, in a nutshell - the 4mm gap is primarily about satisfying the requirement for a successful result in a fire resistance performance test in a furnace.

In practice the smaller the gap, then the better performance in terms of smoke spread. For better performance, install smoke seals that seal the gap.