Installing Rooflights

[Paul Singleton]

I've been trying to find some advice with regards to the best technique to install rooflights to minimize thermal bridging. As I couldn't find any I thought I'd ask the question.

Velux brochures show what I assume is a foam surround that fits into the opening created in the rafters before fitting the roooflight. It says in the brochure this is now necessary to satisfy building regs. As I regard building regs as a minimum requirement should I be doing something more efficient (I'm aiming for CSH Level 4)?

[Mark Siddall]

Roof lights are a problem for thermal bridging as due to the need for an upstand. I've not give this much thought at this time. I'd be interested to see what you come up with and discuss the detail.

If you thinking of Code 4, presumably integrating the PassivHaus standard as a first step, then based upon Belgian experience Velux roof lights should not considered airtight. Fakro are reported to be better. I have no direct experience with these products.

If you do not have site restrictions perhaps you could consider a roof monitor or dormar detail rather than rooflights. This approach would help to reduce over heating risks in the summer, ensure better winter solar gains in the heating season and would allow the use of suitably airtight windows systems.

Mark

[Paul Singleton]

Thanks for your comments Mark. I'm afraid we have too many restrictions to allow roof monitors or dormers. Looks like I'm going to have to give this more thought. If anyone is going to the AECB Conference maybe you could ask the question there on my behalf?!

[Nick Grant]

Paul

A radical option is to install roof light in the insulation plane, insulation over frame, the works, but protect it all with a single glazed roof light or fixed glass in plane of the roof covering.

Saw a version of this in a Passivhaus School near Frankfurt and I believe David Olivier has done something like this for his house.

I think Alan Clarke has some ideas about improving the installation of standard rooflights including choice of flashing kit.

When we built our house only double glazing was available but can now get triple and Chris Herrings says (if I understood him right) that this makes a big difference with roof lights.

Nick

[Chris Herring]

Its certainly true that when glazing is inclined the reduction in U value for triple is very much less significant than with double glazing.  Installation is the big issue, as discussed.  The only soluition I have seen yet is the one Nick proposes  Can't see any inherent design reason why rooflights should be less airtight than vertical windows.

[Paul Singleton]

Chris,

Do you know how much less a reduction in U-value for inclined triple glazing compared to double glazing? Is there something I can reference which covers this as I'm going to have to try and convince my planning officer to allow me to use triple glazed roof windows instead of double glazed conservation ones.

Looking at the information for one of the conservation rooflights it quotes a U-value of 1.0 W/m2K using a 24mm argon filled double glazed unit.

When I look at the information for a triple glazed roof window (which has an optional central glazing bar to give it the conservation look) it also quotes a U-value of 1.0 W/m2K.

Are both of these U-values correct? Seems odd they are both the same. If they are the same then I'll struggle to convince my planning officer that triple is better.

[Alan Clarke]

I have both Velux double and Fakro triple glazed rooflights, and none showed up as a leakage problem on an airtest. The Fakro ones do not have a seperate vent, which could be a weakness in the long term (triple glazed ones only), and have an extra draughtstrip - I don't expect any airtightness problems with these.

The centre pane U-value for the triple is around 0.5, but overall roughly 1. Double glazed would only approach this as a centre pane figure, not overall, which would be 1.5+.

To minimise the edge thermal bridge, as Nick mentioned, I used the most recessed flashing kit available, which isn't the standard one, but will probably please your planning officer (the recessing means that more of the frame is within the line of the insulation).

[David OLIVIER]

European roofllights can quote U values which are for product in a vertical position and centre of glass only. Hence you get absurd claims like a U-value of 1.0 which can't be used for design heat loss calcs., they're as useless as SAP calcs of 1995.

North American data on the installed U-value for multiply-glazed low-e Ar- or Kr-filled products shows it can easily be 3.0 allowing for kerb, higher frame and edge U-value, extra losses from sloping glass, etc. The Conservation Roofligfht will be worse, because it has so much metal not wood.

I wouldn't advise anyone to fit a standard-type rooflight into more than a Silver Standard building because the edge losses from a raised kerb are so high. The SUs aren't very good either!

I was only copying in my house what the 1st Canadian Advanced House did (in 1989)! You have to be extremely exacting in the construction if you split the rooflight into a single outer pane (or dome) and an inner sealed unit but it's workable on small projects.

D.