Hi all, I'm a long-time BuildHub lurker and stealer of others' good tips. I started a loft conversion in April this year and I've been meaning to start a BuildHub diary for a while, partly because I've found the forum enormously useful both prior to and while doing it, and partly because I've ended up learning more about loft conversions than I ever intended to and it may help others in future.

This isn't necessarily going to be a 'look what I built' thread (albeit I'll happily pepper it with photos where appropriate). There are plenty of those and, while they're very useful, I think my personal experience on this project is potentially more valuable in documenting the decisions, problems and compromises that happened between the drawing and the finished building. One of the biggest things I've learned is that you can plan a project to death and still have absolutely no idea what you're going to find when you start cutting holes in it.

The project

We have a detached house, built in 2001 on a typical family housing estate, with a Fink truss roof and the project is a loft conversion with a full-width rear dormer. I wrote and submitted the full plans Building Regulations application myself, mostly because I have been the primary builder and so I wanted to design something that I was comfortable actually building. Just on me briefly - I'm not a professional builder, but it's something I've been doing for a long time and I come from a proud lineage of stubborn people who would rather mess something up themselves and learn from it than pay someone else to do it. It's a dangerous mindset to inherit.

The loft conversion structure itself is based on the EcoTrus light gauge steel system, rather than the conventional approach of using steel beams and timber joists. I chose this for a number of reasons, which I'll detail in a later post, but the primary one was because it was a system that I could build largely by myself with minimal heavy plant and additional hands.

There was also quite a lot going on around the LSG structure: external wood-fibre insulation, mineral wool, a warm-roof approach but with a ventilation layer, a full-width dormer, new rooflights, new stairs, fire protection, upgraded acoustic performance, ventilation and a new bathroom downstairs.

The application therefore became a fairly substantial document covering everything from structure and fire safety through to thermal performance, ventilation, drainage and electrical work. I took several months to put it together and was quite pleased with it when it was done, and it passed inspection without any clarifications. I genuinely thought I'd worked through most of the major construction issues before I'd even got started...

The illusion that you can plan everything

Glibness aside, I'm not trying to say the planning wasn't worthwhile - it absolutely was. I think it has genuinely saved me enormous amounts of time and money so far. The Building Regulations drawings gave me something to work from. The structural calculations gave me a defined structural solution. The thermal details established how the insulation and airtightness strategy was supposed to work. The roof details dealt with things like ventilation, membranes, flashings and the dormer and the interfaces between them.

But the fundamental problem with trying to design an alteration to an existing building to the finest degree is that you don't actually know what you're dealing with until you expose it. I drew multiple scaled CAD illustrations, but ultimately there are parts of the building that haven't been visible since the day it was constructed. A good example of this was the cavity wall and the height offset between the wall plate and the brick outer skin - that relatively small piece of unknown geometry suddenly has consequences for several other things. Where does the fascia go? Where does the soffit finish? How does the gutter sit? How does that affect the external insulation plane? How do the various layers meet? How do you deal with the junction without creating an ugly or unnecessarily complicated detail? Those aren't really Building Regulations questions. They're 'you're standing on a scaffold looking at an actual building, what are you going to do now?' questions.

The value of my time

Over the last six months, I've often found myself weighing up several possible solutions to an unexpected problem and often spending far too long chasing the neatest or most technically elegant answer, particularly because it is my own house. I'm gradually learning that the perfect solution isn't always worth the extra time and complexity.

My time has value too. Sometimes a straightforward solution that works well is better than spending hours trying to make everything perfect.

I'd like this diary to cover both sides of the project: what I planned, using my Building Regulations application as the starting point, and what actually happened during construction. I'll cover the EcoTrus structure, insulation, roofing, windows and all the inevitable surprises along the way. If I waffle too much, please tell me!

Hopefully, it'll be useful to others considering a loft conversion, and an honest record of what happens when detailed plans meet the reality of an existing house. I've added a few illustrations of the project to date to set the scene.

Looking forward to this - thanks for sharing

5 hours ago, Bouncimon said:

The loft conversion structure itself is based on the EcoTrus light gauge steel system

Great to see folk embracing light gauge steel (LGS).

I have a background in this. When doing my training I wrote many structural calculations that underpins the likes of the software that one of the biggest CF building suppliers in the UK and Australia may use and / or their Engineers. There is a place and time for using LGS. If you hit the sweet spot it can be a very elegant and cost effective solution.

I've had a glance at your photos. Some elements such as the connections look technically ropey, could be improved! It looks a bit over engineered but lacking on and understanding of connection design behaviour. The thing that strikes me most is the cost of fabrication, lots of cutting back flanges and so on.

In the round though I think what you have done is great!

Hi @Gus Potter, thanks for reading - I've taken a lot from your previous posts on here.

Interested to know which bits concern you about the technical connections? I'll post some proper detail on how the whole system goes together next; worth saying that some of those photos were taken before all the required fixings were in. But in short, it's 16mm bolts at all connections between the main members and then a lot of 5.5mm Tek screws at 600mm centres. And I mean a lot!

10 minutes ago, Bouncimon said:

I've taken a lot from your previous posts on here.

Thank you for reading some of my posts, much appreciated, some are lengthy not least and a bit ranty!

12 minutes ago, Bouncimon said:

Interested to know which bits concern you about the technical connections?

14 minutes ago, Bouncimon said:

Interested to know which bits concern you about the technical connections?

Ok to be a bit pedantic, you say this was an earlier photo. The use of one bolt at the rafter ceiling joist connection is not common practice, I would look for two.

Looking at your roof, it's of modest size, the loads are relatively small, cuts you a fair bit of slack.

The thing that made me wonder is this. Looking at your photo it all looks over engineered for CF. Now while EcoTrus and their SE's may not have the depth of knowledge and industry experience that I might have, My Masters thesis was testing top hat CF purlins for example, I kind of know a bit about this. I wonder why does this look over Engineered, is there another part of the roof that we can't see from the photos that would explain why I'm seeing what I'm seeing?

But in the round, I really like what you are doing here!

Edited 14 hours ago14 hr by Gus Potter

10 hours ago, Gus Potter said:

I wonder why does this look over Engineered, is there another part of the roof that we can't see from the photos that would explain why I'm seeing what I'm seeing?

No I don't think so - the roof plan is a straightforward box with the exception of a small mini gable on the front and that wasn't included in the steel system, so this structure is purely a pitched roof plane joined to a big flat roof dormer at the rear.

I'll get the full details of the EcoTrus system up later and would welcome your views. Ultimately it's all backed up by the structural calcs, but I've no doubt it's probably over-engineered for what it is - there are essentially only three or four versions of it designed to fit all scenarios, so I guess what they save in bulk manufacturing makes up for some designs being over the top for the problem they're actually trying to solve.

The Ecotrus system - why I chose it

As promised, here's a few more details on the light gauge steel system that I used, and in particular why I chose it. Unlike Gus, this isn't something I'm familiar with - the vast majority of any previous building work that I've done has been with timber, and if I could have done this conversion using wholly timber, I would have happily done so.

My key issue was the span, combined with my desire to do as much of the work myself as possible. Our house is essentially a square box, and it's approx 8600mm from wall plate to wall plate. There are no supporting walls at first floor level and so the existing roof was supported entirely by a typical Fink truss structure.

I therefore kept going round the loop of how to bridge the span without needing the conventional approach of two RSJs at loft floor level and probably another at the ridge. I realise I'm not a structural engineer and there are probably a lot of possibilities which simply didn't occur to me here, but I did do a fair amount of research and cycled through considerations such as:

• Posi-joists: theoretically possible with the span, albeit would have been very deep. I also couldn't find any clear details, despite speaking to a few manufacturers, as how they could sit on the wall plates and be tapered on their top chord to fit within the roof slope.

• Timber I-joists: again, theoretically possible with the span, albeit right at the upper limit. Simpson Strong-Tie used to have a system to allow you to cut the necessary taper on the I-joists to fit within the roof slope, but it appears to have been discontinued now. Regardless, for the span I needed, this wasn't a particularly cost effective option and still didn't resolve what to do at the ridge.

• Glulams: as above, and much heavier to move as well.

• RSJs: probably the most conventional and likely the cheapest option overall. However, not something I felt confident managing or directing on my own, and would have required a crane to get them into position. If I'm honest, the main barrier here was probably my fear factor of operating well outside my comfort zone, plus the difficulty of finding a reliable contractor who was willing to take on the structural steel elements only.

• Telebeam: similar to EcoTrus, but relatively expensive and still didn't resolve what to do at the ridge.

• Various other combinations of building supporting walls at first floor level (too disruptive to the rest of my family, who I've been desperately trying to keep onside!) and trying to span between them.

In summary, the EcoTrus system was the only one I came across which seemed to offer the required span, the freedom for me to install it myself and was backed up with a design service, structural calculations and LABC/NHBC Stage 1 pre-approval and LABC Partner Authority assessment.

To be clear, I'm not saying it was the only option or even the right one, but it solved the problem that was front and centre for me. Being able to say 'I did it' has been a big part of this for me.

How does it work?

I found the system pretty logical. One day, a lot of steel turns up at your front door - in our case, 117 individual pieces of light gauge steel. The main components are the floor joists, rafters, gusset plates, ceiling joists and the dormer faces/cheeks. The latter are the only bits that turn up semi-assembled: we had three panels which made up the front of the dormer and two further panels for the cheeks.

In advance of receiving the steel, you're provided with the assembly drawings and some standard guidance as to how put it together, starting with the floor joists. We removed the bottom three rows of tiles on both sides, fixed ropes to both ends of the joists and then lifted both simultaneously (with the back rope being rigged over a higher part of the scaffold) so that the joist ended up suspended mid-air at roughly loft floor height and in line with the relevant truss bay. Each one was then pulled in and slid through the roof until it was resting on the wall plates on both sides (with 45mm packers already in place). This was by far the most physical bit; each joist only weighed approximately 35kg, but the length made them awkward to handle. We managed it with three people and got them all in within four hours. There are two 250mm C-section joists per truss, fixed back to back through the truss at 600mm centres top and bottom with 5.5mm Tek screws.

Then came the rafters; again slid into the right place using the same technique and with two 165mm C-sections either side of each truss. They were connected to the floor joists initially with a 16mm bolt, later reinforced with the same Tek screw pattern through the truss. Starting with the rafters at either end, the idea is that you then pivot each rafter on the 16mm bolt to lift them up to the underside of the existing felt until they are a specified distance above the floor joists. Once fixed, you string a line between them and use that as your reference point for all the remaining rafters. This sounded straightforward, but we ran into issues with our trusses not being true - once I had the rafters at either end fixed into place, it was impossible to get some of the intermediary ones pushed up to the same height because their corresponding trusses and/or the roof battens in that area were sitting lower. Ultimately, I had to strip that side of the roof to get them all in plane, which was fine for me as we had a tin roof scaffold and because I was planning to re-roof anyway, but would have caused a lot more issues if I had wanted to keep the existing tiles.

Then you add the gusset plates - effectively reinforcing ribs which replace the need for a ridge beam - to the tops of the rafters and in line with your strings. Next comes the dormer face and cheek panels; I had messed up here and not allowed a sufficient gap in the scaffold to pull them through. The largest one was 2200mm by 3000mm and weighed about 120kg. The scaffolder was really helpful and sent a couple of his team round to adjust the scaffold and then help to haul up the panels, which we did with a scaffold hoist and some ropes to stabilise, effectively sliding them up the back of the house. They sit directly on top of the floor joists, more or less directly over the inner leaf of the back wall of the house. Once in position and squared up, you then install the dormer ceiling joists between the gusset plates and an L-bracket which is pre-fixed to the back of the dormer panels.

Notwithstanding the issues with the rafters, it was a straightforward process - quite repetitive (driving Tek screws gets boring very quickly), but logical and sequential in its nature. The below photo shows where we had go to after ten days with two people working full-time (and obviously using the system for the first time).

I'm conscious of length, so I'll end this post here and I'll cover the roof windows framing and some of the system idiosyncrasies next time.

Edited just now1 min by Bouncimon