I was up at 6.00 am and spent two hours at the kitchen table drafting up a way to change across to a 100 mm beam across the stairwell instead of 150 mm.

I then spent much of today reconfiguring the steel (interestingly enough - my original structural engineer, not Dave, had made the same mistake).

Basically - a piece of steel that terminated at another had to be made to continue through, and vice versa for another piece.

In the afternoon I cut the 100 PFC in half and welded it back to back.

I then cut a 100 x 100 mm hole through one of the beams plus a whole lot of additional work (brackets, reinforcing bar) to bring the strength back in to the cut piece.

This work continued over in to Sunday for an hour.

Sunday

I needed to weld the fish plates on today and I realised that this was risky unless I could lay my steel out on a level surface.

I had set out some 90 mm timbers previously to help me visualise the placement of the steel beams.

I spent a couple of hours with a spirit level and various packers to achieve a set of timbers emulating the top plates of the relevant stud frames out on site.

Laying out the steel beams accurately and squarely took me several hours.

Once I was satisfied with this I was able to tack weld the fishplates on to their beams.

This took me up to nightfall.

I still need to mark up the positions of the cleats, fabricate them and weld them on.

After that I still need to rust convert and paint.

With luck, I might be ready to install next weekend.

Wednesday

I took the 3m length of 6 mm x 65 mm and cut it in to 150 mm lengths.

I then realised that I should have bought 6m of flat bar as this only yielded 19 pieces and I needed 21 cleats.

Fortunately, in the workshop I discovered another short length that gave me another 5 cleats.

Thursday

During any break from work I ground a bevel on each cleat for a welding surface.

I also drilled 2 x 12mm holes in to each piece.

Friday

I started welding the cleats on to the beams.

This had to be done carefully with the welding spread out along the length of each beam and sufficient time to cool off so that the beams did not bow.