I have a situation a bit like Peter's recent post although in this case we are upgrading a flat roof (currently just single ply membrane over 9mm ply on 200mm joists with 25mm PUR board between).

Initial calculations on the planned new roof suggest a condensation risk.
I am wondering about the best way to avoid this and how "safe" I need to be.

We will have, from top to bottom

a torch-on 2-layer membrane
100mm of Foamglas Readyblock (λ 0.036)
9mm ply
200mm roof joists with additional insulation between
12mm plasterboard

Taking as an example 22 C inside at 60% gives a dew point of about 6 degrees.

My original spec put as much sheep wool (λ 0.035) in as I could get... ie. 180mm ish of sheep-wool insulation between joists. This gets me an overall average U for the roof of 0.13 (0.12 between and 0.22 at joists) BUT this insulates well enough to put the lower surface of the ply between joists at 6.2 C. At -5 outside this falls to 5.1 C below the ply.

Adding a VCL on the warm side would be tricky and feels wrong. It would create a sealed airspace. And anyway we have quite a few downlights (thought they are pretty airtight) to fit into the plasterboard.

So presumably the sensible solution is just to use less insulation between the rafters? Just decreasing to 100mm fleece seems to give U=0.17 (average for 0.16 and 0.22) and raises the temperature below the ply to 7.4 at -5 outside (or 9.6 at -1 outside).

My questions:-

Is reducing the insulation (never thought I would say that in here) between joists the way to go?

Does the simple numerical approach really sensibly represent what happens with sheep-wool insulation given that it probably allows a little air movement and it has humidity stabilising properties? Does sheep-wool give me some sort of extra leeway?

What sort of temperatures in and out should we really be considering for these dew-point calculations? Do we need to consider rare extremes or are averages enough given that the building will smooth things out a bit? Is my 7.4 degrees C below the ply a safe enough margin when the dew point is 6?

Thanks

It needs a VCL.

We have a warm roof build up on parts of our ground floor. The ground floor rooms extend beyond the upper floor, forming balconies to the rooms above. The balcony floor is a rubber epdm finish and below this is a sandwich of timber firrings, then the warm roof build up. The firrings create the ventilation gap between the rubber roof and the warm roof below.

Sprocket, out of interest, was there a reason you didn't go for a higher performance insulation board above the ply deck? That would improve your overall u value and also keep the ply warmer. You lay the VCL on top of the deck so the ply stays in the 'inside the building' envelope.

Many brands available, just one example here:
https://www.kingspan.com/gb/en-gb/products/insulation/resources/product-literature/thermaroof-tr27-lpc-fm

Thanks, interesting material!

Is it possible to lay say 50mm foamglass on top of 100mm PIR?

Oh! Interesting idea.

Yes, compressive strength of PIR is pretty decent if you can spread the load. I can imagine though that stepping on a corner the PIR might give enough to cause the foamglas above it to crack (foamglas does not give much and might snap) but some of the fibre-faced boards might do OK in that sort of setup.

Please could you share more details on this, I am very interested as it used to happen all the time to chipboard and black chipboard decks but I haven’t seen problems for lots of years though I hate flat roofs so probably have missed a lot of problems.

I have a suspected mechanism that would be made worse by ventilation!

My examples were from 40 years ago England ( my cousin's house) new build roof collapsed walking on it after approx 5 years Reroofed ventilation added (as per my suggestion in earlier post). Okay after that at least 15 years when property sold. 2"nd example Wales (sister) new build after 3 years roof leaked boards rotted was not built to spec. Reroofed ventilated. Interesting point on both houses had building guarantees but because flat roof are classed as cladding they are not covered. So both had to cough up.
How much rework is author of post wants to do. How much original is being kept. If roof is fine not leaking an inside job would be easier cheaper and not dependent on weather. With 200 mm to play with I would consider taking ceiling down insulate to leave 50 mm gap holes in joists in that space for cross ventilation, and eaves ventilation for the run.
150 mm PIR will give you a very good u value about 0.12. Overlay with 25 mm insulation to avoid cold bridging through joist and plasterboard out with duplex board to give a vapour barrier. I would not cut holes in ceiling.
surface mount light fittings.

If going to do a warm roof then I would strip back re-board with something stronger than 9 mm (min 18 mm) and follow a prescribed method as detailed by insulation manufacturer. Kingspan have very good data sheets for warm roof there is even a tapered board designed to give you a built in fall.(Thermataper). there are guidance notes on overlaying an existing roof also.

End of day how much do you want to spend.

It is all OK then 🙂

Posted By: SprocketI really was not looking forward to pulling up all that ply and removing all the nails from joists whilst trying to keep the furring undamaged.

You wouldn't need to take up the existing ply if it was in good condition. Just glue/screw another layer of 9 mm over the top.

Posted By: WillInAberdeenIf your 9mm roof has been ok this far, it should be fine when protected with thick insulation.

I don't think the question is about protecting the ply with insulation; it is rather about whether the ply is stiff enough to protect the insulation. Foamglas is brittle and may break if loaded between point supports.

If the foamglas is glued down to the ply, then the two will form a stressed-skin structure that will be more resistant to bending and thus possible fracture of the foamglas core.