In order to have more insulation in the pitched roof sections, I had planned to use 50mm counter battens underneath the rafters. I need more than 50m now having looked at the various insulation options (can't use PIR due to design ethos). 75mm counter battens start to feel a bit crazy. I could go with 25mm battens and 50mm counter battens or v-v. A mistake not to have gone with full rafter fill, but having to roll with it now.

 

A thought occurred to me which was to use GL1 and GL6 to achieve a 75mm gap. This would be better thermally as would avoid the counter batten thermal bridge. I have sketched the make up below. I'd be able to get 2no. screws into the GL6 about 80mm apart. I've emailed BG Technical, but don't have a good record of timely responses.

 

 

I guess another mounting option if the GL6 has to be vertical is to add a timber batten cut to a triangle:

Have you considered mounting insulation on the external side and then membrane and batten?

2 minutes ago, MikeGrahamT21 said:

Have you considered mounting insulation on the external side and then membrane and batten?

 

Too late unfortunately as the roof is tiled.

I'd stick with 25x50 battens, but just cut 50mm stand off blocks to get the 75mm depth. i tend not to worry about small timber thermal bridges. 

 

Must admit, in a similar position I went for PIR. 

23 minutes ago, George said:

I'd stick with 25x50 battens, but just cut 50mm stand off blocks to get the 75mm depth. i tend not to worry about small timber thermal bridges. 

 

Must admit, in a similar position I went for PIR. 

 

Stand off blocks are a nice idea.

 

PIR certainly has it's benefits.

40 minutes ago, George said:

Must admit, in a similar position I went for PIR. 

 

I'm considering Pavatherm as an alternative. What type of scre do you use when you fix plasterboard through PIR?

This system is interesting and doing the sort of job I am after.

 

https://www.bigmat.fr/video/ki-comble-reno-2013/

https://www.youtube.com/watch?v=kXEAGvPTJuM

This is what I did

 

I needed an additional 100mm, so I cut short pieces of 47x50,  fixed these at 600mm centers, then further 47x50 were attached to the short pieces.  In my case the whole lot was filled with spray foam.  Then a reflective Air tightness vapour control layer and then a further set of counter battons.

On 22/10/2021 at 22:32, JohnMo said:

This is what I did

 

I needed an additional 100mm, so I cut short pieces of 47x50,  fixed these at 600mm centers, then further 47x50 were attached to the short pieces.  In my case the whole lot was filled with spray foam.  Then a reflective Air tightness vapour control layer and then a further set of counter battons.

 

Wow that must have given a very good U-value as you pretty much have a continuous ~300mm thick layer of foam insulation. I can see Posi-joists work really well with insulation like that.

It should be pretty good, the only thing that lets it down is the bridging caused by the metal ties of the posi joists.

 

The spray foam is expensive, but they were in and out in 2 days, we have approx 240m2 of roof.  It would have taken me weeks by hand and not achieved any better.  The good thing about the spray foam is its gap free.  The additional PHS Hi-Thermia Reflective Membrane (air tight and vapour tight), when installed with 25mm air gap adds an additional 0.78 to your R value.

Double post

Edited October 25, 20214 yr by JohnMo
Double post

I saw some videos of a larger renovation in France in which they used the GL6 connectors. Well I've given myself a bit of a shock working out the thermal bridge of one of these GL6 connectors. Steel has a lambda value of around 60 W/mK whereas pine has a value of round 0.18W/mk. That's a factor of 333. If these are at 400mm c/c x 600mm c/c that works out as 4.167/m2. If the average equivalent width is 25mm (reduced from 27mm by the holes) and the thickness is 2mm (guessing, will check) then the loss per m2 would be 4.167 * (0.025m *0.002m) * 60W/mk / 0.171m = 0.0125W/m2K which is a pretty large addition to the U-value. that assumes the metal extends all the way through the insulation, which it wouldn't.

 

 

The SAP calculation looks great as it ignores this thermal bridging, but the reality is more complicated.

 

 

On 30/03/2022 at 23:03, MortarThePoint said:

4.167 * (0.025m *0.002m) * 60W/mk / 0.171m = 0.0125W/m2K

Forgot to do the division: dU = 4.167 * (0.025m *0.002m) * 60W/mk / 0.171m = 0.073W/m2K

 

I've checked with Arthur Hough and there AH120 (aka GL6) is 0.8mm thick. BG suggest 600mm x 600mm centres (see extract below), reducing the number per square meter to 2.78/m2. That reduces the thermal bridging to dU = 2.78 * (0.025m *0.0008m) * 60W/mk / 0.171m = 0.020W/m2K.

 

That first number seems so large (e.g. changing 0.15 to 0.22) that it can't be right. I guess it is treating it as a bridge all the way through to cold which isn't really fair. To do it properly needs FEA.

 

Alternative model would be to consider GL6 in series with a block of wood to cold. the length of the block of wood would be the thickness of insulation remaining past the end of the connector, so 80mm in my case. What area to make the block of wood?

    27x27  :  r_wood = 0.080 / 0.12*0.027*0.027 = 914, r_gl6 = 143, r_total = 1067  --> dU = 2.78 / 1067 = 0.003

    47x50  :  r_wood = 0.080 / 0.12*0.047*0.050 = 284, r_gl6 = 143, r_total = 427 --> dU = 2.78 / 427 = 0.007

    47x81  :  r_wood = 0.080 / 0.12*0.047*0.081 = 175, r_gl6 = 143, r_total = 318 --> dU = 2.78 / 318 = 0.009

 

Feels reasonable to assume that the affect of using GL6 connectors vs some magic non-conductive ones is dU = 0.005 - 0.010, I'm happy to work with 0.005.

 

Removing 50mm wide by insulation_thickness_below_rafters timber battens improves the U-value by dU = 0.004 so the GL6s cancel that out and it comes down purely to cost and ease of installation.

 

Also putting it into context, it's all equivalent to a difference of about 7mm of Mineral Wool! I was uneasy that my pitched section's U-value was rounding up to 0.15 rather than down to 0.14. A single 1m2 window cancels out 120m2 of that difference!