Cold bridge at new patio doors – sill resting on brickwork? Installer says “nothing more can be done”

[fandyman]

Hi all,

I’d appreciate some advice from people who know more about window/door installation than I do.

We had new REHAU Rio patio doors with sidelights fitted earlier this year by a small FENSA-registered company. We are getting very noticeable cold and draughts around the bottom of the frame/sill junction, especially now that the weather has turned colder.

After taking some thermal images, there is a very clear cold line exactly where the sill meets the frame. It’s a consistent pattern across the whole width. I’ve attached the thermal photos below.

I also looked underneath the exterior cill and it seems like the sill is resting directly on the brickwork with no insulation or thermal break. The installer told me:
 

“There’s only a 3–5 mm gap, we can’t insulate that, proper insulation isn’t required for small gaps.”

and also:
 

“It’s your floor that’s cold, not the door.”

The floor is finished up to the frame, and the thermal camera shows the cold coming specifically through the frame/sill junction, not from below.

 

The coldest areas appear most pronounced at frame junctions and corners rather than uniformly across the opening.

My questions for the forum:

1. Is a sill resting directly on cold brickwork with only tiny packers (1-4mm) and without insulation acceptable practice?
2. Should a proper installation normally allow some allowance for insulation or a thermal break beneath the sill, rather than bearing directly on masonry?
3. Would resealing or adding more foam internally achieve anything, or is this fundamentally a threshold installation problem?
4. Is this something that FENSA would normally get involved with?
5. Has anyone here had to escalate a similar issue under the Consumer Rights Act?

I’m trying to resolve this with the installer first, but they say the installation is correct and that nothing more can be done.

Any professional opinions or similar experiences would be really helpful before I decide how to proceed.

Thanks in advance.

Edited 21 hours ago by fandyman

[Redbeard]

55 minutes ago, fandyman said:

My questions for the forum:

1. Is a sill resting directly on cold brickwork with only tiny packers (1-4mm) and without insulation acceptable practice?
2. Should a proper installation normally allow some allowance for insulation or a thermal break beneath the sill, rather than bearing directly on masonry?
3. Would resealing or adding more foam internally achieve anything, or is this fundamentally a threshold installation problem?
4. Is this something that FENSA would normally get involved with?
5. Has anyone here had to escalate a similar issue under the Consumer Rights Act?

I’m trying to resolve this with the installer first, but they say the installation is correct and that nothing more can be done.

 

Answers:

 

1. I doubt if any of us here would say it was acceptable but I suspect it is common.

2. Yes for me, and I guess for all of us here.

3. Not much, and yes.

4. I don't know. I fear that there's no obligation to insulate below a cill. I'd do it anyway, but whether Bldg Regs would *make* me do it, I think perhaps not, but not certain. The last patio doors I was involved with were sat on Compacfoam.

5. No, sorry.

 

I think if I had been presented with what you have there I would be cutting out a brick at a time and replacing with something like Compacfoam - expensive but effective load-bearing polystyrene.

 

I am sure there are lots of really good uPVC window and door installations. Unfortunately I have seen a large number of average-to-awful ones, so this detail upsets me but does not surprise me. 

 

Was the uPVC door a replacement of an existing door, or is that a 'dropped' window opening? The latter, I think, given the odd cuts on the bottom half, RHS.

 

Edit: Thinking about it, the patio doors we did were in a wall with EWI, so the compacfoam was 'buried'. In your case you'd have a 'raw edge'. It's late, so I'll think in the morning how one might deal with that. How far above ground level is the external cill? If it is close you'll get a lot of splash-up. OK, it's not a timber door, but neither is it ideal. (And whatever the Building Regs say, rain splashed up more than 150mm in many areas!

Edited 20 hours ago by Redbeard

[Andehh]

I know how it feels OP, I laboured over every bit of insulating detail for our place until our 2.4 x 7m sliding doors were installed (I was abroad with work). They removed insulation from cavity and back filled with solid concrete and sat the flush doors on them (from what I can tell). 

 

The installers refused any other installation method or approach due to our big, heavy and expensive the doors were, but it has left me a massive cold bridge which can be felt 1 foot into the room. Not sure id have accepted the risk of forcing them to do it differently, but I'm pretty sure mitigations could have been found for some if it (between them and builder). 

 

As the doors are nigh on fully flush inside and outside, nothing I can do about it. 

 

You can't win them all.  

 

 

(edit: also be wary of knock on implications, fix one problem may create others, and adding flex or tweaking installed doors may create other issues) 

 

 

Edited 12 hours ago by Andehh

[Oz07]

You're worrying too much if it were a passive new build fair enough. To me it looks like the installers have done an ok job with the circumstances. What did you expect the door to be made 25mm short then insulated below and sit on pvc packers which are cold bridges anyway? To me it looks like your screed has no thermal break to the outside bricks. Completely beyond the remit of some pvc door installers. Atleast they've put you some dpc down, thats above and beyond IMO. If you wanted to sort any potential defects with the way your building was put together you should of ripped the old frame out, done what you needed to do, then had the frame made to suit the alterations. I think your out of order trying to seek redress for this. 

 

If your getting a draft a good healthy bead of silicone around the internal and check the locks and seals are adjusted right. That's fair enough. 

Edited 11 hours ago by Oz07

[Andehh]

Your very last picture, I would masking tape it off and then apply a very thick black bead along the bottom. 

[fandyman]

11 hours ago, Redbeard said:

 

Answers:

 

1. I doubt if any of us here would say it was acceptable but I suspect it is common.

2. Yes for me, and I guess for all of us here.

3. Not much, and yes.

4. I don't know. I fear that there's no obligation to insulate below a cill. I'd do it anyway, but whether Bldg Regs would *make* me do it, I think perhaps not, but not certain. The last patio doors I was involved with were sat on Compacfoam.

5. No, sorry.

 

I think if I had been presented with what you have there I would be cutting out a brick at a time and replacing with something like Compacfoam - expensive but effective load-bearing polystyrene.

 

I am sure there are lots of really good uPVC window and door installations. Unfortunately I have seen a large number of average-to-awful ones, so this detail upsets me but does not surprise me. 

 

Was the uPVC door a replacement of an existing door, or is that a 'dropped' window opening? The latter, I think, given the odd cuts on the bottom half, RHS.

 

Edit: Thinking about it, the patio doors we did were in a wall with EWI, so the compacfoam was 'buried'. In your case you'd have a 'raw edge'. It's late, so I'll think in the morning how one might deal with that. How far above ground level is the external cill? If it is close you'll get a lot of splash-up. OK, it's not a timber door, but neither is it ideal. (And whatever the Building Regs say, rain splashed up more than 150mm in many areas!

Thank you — that’s really helpful and aligns closely with what I’ve been suspecting.

Just to clarify one important point about the opening:

• When I purchased the house over three years ago, the opening already existed at full height exactly as it is now. There were no windows, doors, or dwarf walls present, and no structural changes have been made since (there was a conservatory that has been demolished since then and new patio paving laid in its place). The installer therefore worked with a pre-existing opening and was responsible for detailing the door installation within that condition.

The sill appears to be bearing directly on brickwork, with no packers or thermal break beneath. Internally, the cold line corresponds very precisely with the sill/frame junction and the side-panel-to-door frame junction, which is where condensation and mould are occurring.

Thermal images show localised internal surface temperatures at those junctions significantly lower than adjacent areas, suggesting a concentrated cold bridge rather than general floor temperature influence. The floor itself does not show the same thermal signature away from the frame.

Your comment about Compacfoam is interesting — that’s broadly what I would have expected in terms of maintaining thermal continuity at a load-bearing threshold, even if not explicitly mandated by Building Regulations. As you say, “I’d do it anyway.”

What I’m still trying to understand is whether, in a retrofit scenario with a pre-existing opening, best practice would normally allow a sill detail that results in the internal frame/sill junction dropping below dew point under normal winter conditions — even if the regulations themselves are not explicit.

For clarity, the patio paving was installed after the door, and the final external level is approximately one brick (~75 mm) below the sill.

Thanks again for taking the time to respond — this has been extremely useful.

Edited 10 hours ago by fandyman

[fandyman]

1 hour ago, Andehh said:

I know how it feels OP, I laboured over every bit of insulating detail for our place until our 2.4 x 7m sliding doors were installed (I was abroad with work). They removed insulation from cavity and back filled with solid concrete and sat the flush doors on them (from what I can tell). 

 

The installers refused any other installation method or approach due to our big, heavy and expensive the doors were, but it has left me a massive cold bridge which can be felt 1 foot into the room. Not sure id have accepted the risk of forcing them to do it differently, but I'm pretty sure mitigations could have been found for some if it (between them and builder). 

 

As the doors are nigh on fully flush inside and outside, nothing I can do about it. 

 

You can't win them all.  

 

 

(edit: also be wary of knock on implications, fix one problem may create others, and adding flex or tweaking installed doors may create other issues) 

 

 

Thanks for sharing that — I can definitely relate to the frustration.

I agree that large, heavy sliding doors and fully flush thresholds are particularly difficult details, and I can see why installers default to solid bearing in those cases. In your example it sounds like the constraints of door weight and flush internal/external levels severely limited the options.

My situation is slightly different in that the installation is a uPVC patio door with sidelights rather than a very large sliding system, and there doesn’t appear to have been any discussion or design decision around managing thermal continuity at the threshold — the sill simply bears directly on masonry.

What I’m trying to understand is whether, in a relatively conventional retrofit like this, best practice would normally expect some mitigation to prevent the internal sill/frame junction becoming the coldest surface in the room and repeatedly dropping below dew point, rather than accepting that outcome as unavoidable.

I completely take the point about knock-on effects, and I’m not looking to “tweak” the installed doors myself — this is more about whether the original detailing should reasonably have anticipated and avoided this condensation risk in the first place.

Appreciate you taking the time to reply — it’s useful to hear real-world experiences, even when the outcomes aren’t ideal.

[fandyman]

1 hour ago, Oz07 said:

You're worrying too much if it were a passive new build fair enough. To me it looks like the installers have done an ok job with the circumstances. What did you expect the door to be made 25mm short then insulated below and sit on pvc packers which are cold bridges anyway? To me it looks like your screed has no thermal break to the outside bricks. Completely beyond the remit of some pvc door installers. Atleast they've put you some dpc down, thats above and beyond IMO. If you wanted to sort any potential defects with the way your building was put together you should of ripped the old frame out, done what you needed to do, then had the frame made to suit the alterations. I think your out of order trying to seek redress for this. 

 

If your getting a draft a good healthy bead of silicone around the internal and check the locks and seals are adjusted right. That's fair enough. 

Thanks for taking the time to respond.

I agree this wouldn’t be an issue worth discussing if it were about chasing passive-house levels of performance — that’s not what I’m aiming for. The concern here is that the internal sill/frame junction has become the coldest surface in the room and is repeatedly dropping below dew point under normal winter conditions, leading to condensation and mould.

I’m not suggesting the door should have been made short or that it should sit on compressible packers alone. My question is whether, in a conventional retrofit like this, best practice would normally allow the sill to bear directly on masonry with no thermal break or mitigation, given the predictable internal surface temperatures that result.

I accept that the solid floor and external brickwork form part of the thermal boundary, but the thermal imaging shows the cold concentrated at the sill and frame junctions rather than uniformly from the floor. That suggests a localised bridge associated with the installation detail rather than the slab alone.

I’m also not proposing post-installation tweaks or DIY fixes — sealing or adjusting locks doesn’t address a conductive cold path, and I’m conscious of knock-on effects. This is about understanding whether the original detailing reasonably anticipated and avoided this condensation risk.

Appreciate the alternative perspective — I’m mainly trying to establish where the line sits between “unavoidable compromise” and “avoidable detailing issue” in this type of installation.

[fandyman]

1 hour ago, Andehh said:

Your very last picture, I would masking tape it off and then apply a very thick black bead along the bottom. 

Thanks — I understand why that might help with draughts or water ingress.

My concern here isn’t primarily air leakage, but internal surface temperature. The condensation and mould are forming because the sill/frame junction is becoming cold enough to drop below dew point under normal conditions.

Adding a thick internal bead may improve airtightness, but it wouldn’t change the underlying thermal path through the sill into the masonry, and therefore wouldn’t raise the surface temperature that’s causing the condensation. There’s also a risk of sealing moisture into the junction rather than resolving it.

That’s why I’m trying to understand whether the installation detail itself reasonably anticipated and mitigated this thermal bridge, rather than looking for post-hoc cosmetic fixes.

I appreciate the suggestion though.

[Oz07]

I think these responses are being written by AI 

 

Is there any way bots can join the forum and post these type of queries or is op just utilising ai to draft their responses. You suspect the original query is copy paste due to the formatting. There have been a couple of posts on this forum recently that make me suspicious. The user the other week quoting replies and inserting spam links. 

[Oz07]

FYI this is exactly how I'd expect a frame to be installed in an existing opening. End of story. 

21 minutes ago, fandyman said:

My question is whether, in a conventional retrofit like this, best practice would normally allow the sill to bear directly on masonry with no thermal break or mitigation, given the predictable internal surface temperatures that result.

What would you propose as an alternative chat gpt?🤣

[fandyman]

I’m here to understand the technical aspects of the installation and the causes of the observed condensation.

I’ve tried to be clear and structured in explaining the issue, but I’m happy for people to disagree on the substance.

If you think the installation detail is acceptable and that the symptoms are expected, I’d be interested in the technical reasoning behind that.

[JohnMo]

Think the trouble with a lot of these questions and issues, they are raised after the event, when it's basically too late. People expect wonders to occurs without pre planning and deep discussion on expected outcomes. Thermal bridging isn't great, but we have a work force of poorly trained labourers, not skilled crafts people.

 

Only things open to you without taking everything out and starting again, fill all gaps with expanding foam, seal the outside with mastic to form a waterproof seal.

 

But looking at the thermal images, I would comment

 

Door cill area doesn't look that bad, worst bridge appears to be at the side lights, most likely caused by air movement.

 

Your doors most likely need adjustments as you appear to have drafts at the door seals.

 

So as mentioned fill any remaining gaps with expanding foam, seal with mastic and move on with your life.

 

 

 

 

[RedRhino]

I imagine that the vendor would have confirmed your order with the size of the patio door they were installing. 
That defines the contract - their glazing to match your hole. The patio doors fit the hole and they have fulfilled the contract. 
IMHO it would take higher level knowledge to think beyond this (e.g. it would be reasonable to expect an architect to anticipate this problem). 
But I wouldn't be too upset. What's done is done. We all learn through experience. 

[Mr Punter]

13 hours ago, fandyman said:

cold and draughts

 

The draughts are to do with airtightness, not insulation.  Externally you need to run sealant between the frame and the brickwork and under the cill.  Get a nice tooled finish on the sealant.  Add extra sealant between the cill and the base of the frame.

 

If the floor adjacent to the doors is cold it may be because of air leakage, insufficient insulation or no thermal break, such as a perimeter upstand of PIR.

[fandyman]

24 minutes ago, RedRhino said:

I imagine that the vendor would have confirmed your order with the size of the patio door they were installing. 
That defines the contract - their glazing to match your hole. The patio doors fit the hole and they have fulfilled the contract. 
IMHO it would take higher level knowledge to think beyond this (e.g. it would be reasonable to expect an architect to anticipate this problem). 
But I wouldn't be too upset. What's done is done. We all learn through experience. 

I do not think "it fits the hole" is correct. In UK law (Consumer Rights Act 2015) the supplier’s obligations include that the service is carried out:

• with reasonable care and skill

• fit for purpose

• as described

I do not believe that fitting the opening dimensionally exhausts the duty of care.

 

If that were true leaking windows would be acceptable if they “fit”. It would also mean draughty doors would be acceptable if they “fit” and cold bridging would never matter right?

 

I am confident that is not how CRA works.

[Redbeard]

Deleted: 

Edit: I had not read @RedRhino's post above, so I had just effectively repeated their theme. Sorry!

 

 

Edited 7 hours ago by Redbeard

[Redbeard]

1 hour ago, fandyman said:

I do not think "it fits the hole" is correct. In UK law (Consumer Rights Act 2015) the supplier’s obligations include that the service is carried out:

• with reasonable care and skill

• fit for purpose

• as described

I do not believe that fitting the opening dimensionally exhausts the duty of care.

 

If that were true leaking windows would be acceptable if they “fit”. It would also mean draughty doors would be acceptable if they “fit” and cold bridging would never matter right?

 

I am confident that is not how CRA works.

 

I think the points being made by those who have responded to your OP suggest that, however regrettable it is, most window/door re-fitters do 'fit the hole'. In the absence of any other instruction (such as re thermal bridging mitigation, which might be included on an architect's or other professional's spec.) that's 'what they do'. If they have not 'fitted the hole' very well (the sealant is poorly-detailed or the opening lights do not 'kiss' the draught-stripping well) this can and should be amended. Then they would have done the job 'with reasonable care and skill', 'fit for purpose' (it's a door without draughts) and 'as described' insofar as their 'offer' probably said something along the lines of  'supply and fit XYZ door'. (What did it say?) If peripheral insulation measures had been promised and not installed then the CRA (of which I know nothing, TBH) may 'kick in'.

 

Knowing that you are unhappy about the cold threshold I suggested a careful and painstaking 'stitching' job with Compacfoam or similar. That could be done (with caution - don't accidentally introduce 'droops' as others have said) but it would be an extra over and above (as far as we know) what was quoted for.

[fandyman]

3 hours ago, Redbeard said:

 

I think the points being made by those who have responded to your OP suggest that, however regrettable it is, most window/door re-fitters do 'fit the hole'. In the absence of any other instruction (such as re thermal bridging mitigation, which might be included on an architect's or other professional's spec.) that's 'what they do'. If they have not 'fitted the hole' very well (the sealant is poorly-detailed or the opening lights do not 'kiss' the draught-stripping well) this can and should be amended. Then they would have done the job 'with reasonable care and skill', 'fit for purpose' (it's a door without draughts) and 'as described' insofar as their 'offer' probably said something along the lines of  'supply and fit XYZ door'. (What did it say?) If peripheral insulation measures had been promised and not installed then the CRA (of which I know nothing, TBH) may 'kick in'.

 

Knowing that you are unhappy about the cold threshold I suggested a careful and painstaking 'stitching' job with Compacfoam or similar. That could be done (with caution - don't accidentally introduce 'droops' as others have said) but it would be an extra over and above (as far as we know) what was quoted for.

 

We agree to disagree then. I am pretty sure that under UK consumer law a consumer does not need to instruct a trader on how to avoid creating defects and the trader must still perform the service with reasonable care and skill.
 

You don’t need an architect’s spec to expect no persistent condensation, no mould and no unreasonably cold internal surfaces under normal use.

 

If the outcome is defective, the absence of instruction does not automatically excuse it. That would be nonsense.

[fandyman]

I get what you’re saying about how most re-fitters operate and the reality of “fit the hole” jobs - I’m not disputing that this is common practice.

Where I struggle is that we’re not talking about chasing passive-house performance or some theoretical improvement. We’re talking about persistent condensation and visible mould forming at a very specific junction, under normal use, that wasn’t there before the installation. That feels like an outcome problem rather than an expectations problem.

I didn’t ask the installer to redesign the opening or go beyond what was quoted, but equally I don’t think it’s unreasonable to expect a newly installed door not to create a cold internal surface that repeatedly condenses and grows mould. Even if the trade norm is to accept that detail, it doesn’t automatically follow that the outcome is acceptable in every case.

I’m not saying the installer acted in bad faith - just that something about this particular junction isn’t working as intended. That’s why I’m pursuing independent/regulatory review rather than arguing it out on forums.

[craig]

1: Is it a bad install? 
A: No, not the best, I’d expect them to clean up the expanding foam and trim/silicone internally (if within remit).

2: Threshold, gaps and no insulation.

A: I would have expected this to be foamed and mastic applied, they’ve forgot.

3: Have they breached consumer rights under care & skill?

A: No.

4: Draughts/leaking etc.

A: Around the frame and on occasion via the door / window is the installers responsibility. They’ve not foamed under the door and I’d ask them to come back and do it plus mastic.

 

Other than that, they cannot do anything about the existing cold bridges present. It’s not up to them. 

 

What I see and read a minor snags on the finishing internally and whether that was within the remit and the underside of threshold (foam/mastic). If that was reported to me as a supplier. I’d apologise and have the team attend to sort out at least one of the two issues that I see.

 

Edited 2 hours ago by craig

[Redbeard]

@fandyman

19 minutes ago, fandyman said:

That’s why I’m pursuing independent/regulatory review rather than arguing it out on forums.

sounds like a 'bowing-out' post. I was about to trawl back for/ask for further details of the condensation/cold, and exactly how  it was manifesting. I thought I had read more of cold spots than of condensation of mould, but I may have skipped past a significant bit. If you're bowing out I won't trawl back/ask my Qs (which would have included what the issues were with condensation and mould on the old door. Edit: I think maybe you did not have a door there at all, as you had a conservatory.). Is the condensation forming on the frame and running down onto the screed, or forming on the screed, *not* on the uPVC? The more detail you can give the more help we can be. I wish you luck with it, whether you pursue it here or by other means.

Edited 2 hours ago by Redbeard

[Mr Punter]

Those plastic sills are never brilliant as thermal breaks but I suspect the issue may be that the floor adjacent to the doors is not sufficiently thermally isolated from the wall.  Do you have a thermal image of this area?  Also, do you know what insulation is under the floor?

 

Door thresholds are often a real thermal weak point.

[saveasteading]

29 minutes ago, craig said:

It’s not up to them

Quite so. I'd advise you listen to the advice here. it may be free and we aren't always right,  but it is by people who have done it/ do it for a living/ know the theory or the reality or both.

 

Draughts are worse than cold bridges and that thermal pic may make it look worse than it is. draughts can be sorted with standard brush strips. so what do you do if you get awarded £30 towards fitting these?

 

33 minutes ago, fandyman said:

I’m pursuing independent/regulatory review

Up to you. You will spend many hundreds and they might agree that it isn't a justifiable claim. 

I've sat with an industry expert (selected by my solicitor) because of a cost claim by a client. I think it was about structural design.  After a very expensive hour he said how interesting it had been, and I clearly knew more than him on the matter in question. That might be the case here, with the free advice above being equally expert. Not mine: the others who know this subject backwards.

[Nestor]

15 minutes ago, Mr Punter said:

Those plastic sills are never brilliant as thermal breaks but I suspect the issue may be that the floor adjacent to the doors is not sufficiently thermally isolated from the wall.  Do you have a thermal image of this area?  Also, do you know what insulation is under the floor?

 

Door thresholds are often a real thermal weak point.

Not an expert but have similar thoughts about where the internal floor build up meets the external walls, no sufficient  thermal break visible to me.

I fitted 20mm Pavatherm reveal boards in the window apertures, timber frame build. Not suitable under large doors due to compression so used 25mm plywood. Walls are 390mm deep.

Edited 1 hour ago by Nestor