you don’t always end up where you’d anticipated.
this is never more true than when you are building. plans invariably change. ideas co-allese. priorities shift. it is the nature of things to evolve.
when buildings evolve during the process of initial construction, it can mean re-doing things, unless you are fortunate enough to have pre-empted the change, or are building in such a way that allows for change.
the second phase of constructing this timber frame wasn’t easy. what is? there is a phrase in construction known as incremental error, build enough, and you will find it sooner or later. the key to it, is how you overcome it. sometimes it just means stopping it and other times it means undoing, and or, adjusting things.
when i started this frame i had certain reservations about how i was going to be able to put the frame up, those reservations didn’t subside until i had essentially finished. until then it was all hypothesis. only in the pudding was there ever going to be any proof.
much of the process of construction for the second phase was the same as the first.
several of the parts are essentially acting in the same way. however, at the first floor, whereas the most external lateral intersecting beams (girts) ran from post to post (in the same way as the 1st floor floor joists),
at the loft level there are two 8×10 roof plates scarfed and pegged together, that run the full length of the building, front and back. housed into these plates are the tie beams (the first part of the roof truss), which run from the front to the back plate.
because i didn’t have a crew of 56 people or a crane i had elected to half the posts, essentially building in a more modular way, a way i could manage on my own. this caused me a series of problems i needed to overcome. principally it meant the need for additional support for each of the posts, particularly at the point where they intersected, as i was creating what is known as a nodal point (a point of compressive stress that wants to act like a knee) and i needed to restrain it.
this meant the addition of knee braces throughout the structure, where they had not been allocated in sobon’s design, and the pegging of all the first floor joist ends to effectively create a raft of floor joists, girding beams and girts.
subsequently, i have thought that a slightly different system might have been preferable, to have treated the first floor post-tops as above and used beams in the manner of roof plates to house the posts more stiffly, which would’ve obviated the need for all the additional bracing, and reduced my build time. however, at that time i had issues with doing it that way, firstly and most importantly, i didn’t have sufficient material, and secondly i had joinery conflicts there, and strength issues if i’d changed the choices of joints.
as it was, each of the top section posts had a mortice housing underneath the post that had to correspond with a tenon on the receiving post from the floor below. beginning to sound complicated?
it was imperative that all the posts lined up, so that the plates could line up, so the roof line would be continuous. this is where the possibility of incremental error is likely to rear its ugly head. get one thing out, lower down the sequence, and it throws everything else out of true or square, or plumb. this was exactly the scenario i was facing as a shift in the line of the lower posts had occurred, and since it wasn’t possible to shift the frame back into perfect alignment (i tried, but it was too well secured and too damn heavy), it meant altering the point at which the top posts sat on their corresponding tenon below, to avoid continuing or worsening the error.
braces were cut, then posts and beams had brace mortices let into them.
the braces and posts needed to fit exactly into the roof plates with millimeter accuracy in order for everything to line up, which meant they had to be pre-assembled.
the posts were then set out on their new line and braced
the roof plates had to have long and complicated scarf joints to make the span (approx 9m),
the 8 post tops had a 4″ (100mm) x 1 1/2″ (37.5mm) tenon that ran the full width of the post, this was to locate into the underside of the roof plate, since both the front and back plates were in two parts that were joined by a long and complicated scarf joint it meant there really couldn’t be much play in the mortice housing, no more than 5mm between the 2 posts per length of plate, or put it another way, each post was only allowed 1.25mm play either side laterally within its unique housing, and less than 1mm from front to back (that means tight enough to need drifting in).
i pre-assembled each pair of posts, their respective braces, and the roof plate they all housed into, to make sure everything fitted well. knowing that when the posts were in and i was lowering the plates onto them would not be a good time to have to adjust the fitment.
the process of getting the plates in was nervy, to put it mildly. the plates weighed a lot, they had to be suspended in mid air, with a makeshift scaffold, and eased onto the post tops one at a time, lowered and tapped down until they fit, then lifted out enough to get the braces in, re lowered and sent home. which probably sounds easy, but when you have a 5m beam that weighs several hundred pounds swinging around, willing itself to slip out of its sling, and fire you off the scaffold and down to the ground 8m below, was pause for thought.
the first plate each side was relatively easy compared to the second. as that not only had to line up with all the post top tenons and braces, but the scarf joints had to line up aswell, these joints were over 2ft long, and had to fit as tightly as possible.
the really awkward part of this is that you can’t tell if there is any twist in the plate over its entire length that is going to affect how both halves of the scarf joint will fit together. it only has to have the tiniest set and they won’t go together, not when they are housed into numerous other components. even though i had pre-assembled the scarf joints together, it was always going to be a slightly different story when the plates were sat on-top of the posts in situ.
in the end it all went well.
preparation was the key.
following getting in all the roof plates, it was time to drag the nearly 21′ x 8×10″ (7m x 200mmx250mm) tie beams down the side of the mountain, one at a time, mount them on the scaffold, the set out the cuts.
the roof plates had one half of the connecting dovetail joint (the female half) and additionally a step housing on the internal face.
the tie beams had the male half of the dovetail joint which would tie them into the plates, and the male part of the step housing, set below the dovetail.
the complication here was again the need for everything to fit exactly, and without play. if there was play in the joint it would’ve been sloppy and the dovetail would not have acted as it needed to, to tie the building together at this point. too tight and the beam wont go in. you can’t crush the fibers of the joint trying to force it in as it prevents the beam from housing properly. you can’t even kerf cut the beam into the joint, as there isn’t room to get a saw in. it has to be set out and cut right.
some of the dovetail joints had to be off-set and resized, as the wane in the face of the plates, and in one case the proximity of a scarf joint prevented it from having a full dovetail.
the tie beam also had braces, that ran from each post to the underside of the tie beam. these all had to be cut and housed and tried prior to insertion.
the scaffold had to be built to raise each tie beam in turn. to take it out past the front of the frame in order to bring the tail end back over the top of the plate. this was an awkward manouver. following this, the tie beam had to be raised and then lowered into position into the corresponding dovetail joints and housings. this meant it had to contact four different joints with millimeter perfect dimensions in order to fit. once the tie was in there was no real way it wanted to come out. whilst this occurred it was necessary to insert the braces in their respective housings. since i was doing this on my own it meant tying in the braces as the beam was being lowered, so they didn’t fall out of place.
when everything was set down, it was time to drill and peg the joints.
the principal rafters.
up to this point i had been generally following jack sobon’s book. from here on in it wasn’t going to help as i had other issues to deal with that he didn’t. i plan to build a room at the back of the house that will become the bathroom. the roof of which i plan to segue into the main roof (at a different pitch to the main roof) and join-in part of the way up at the back.
this meant that i would be putting additional load on the roof. rafter dimensions and construction is generally based on estimated load. this is derived by two things live and dead loads. a dead load is a static load, such as the weight of the materials that go to make a roof. and a live load is that which acts upon the roof, such as snow, or more likely in my case, wind.
in order to increase the carrying capacity of my roof, without increasing the rafter number or size, i felt it would be appropriate to add in principal rafters. together with collar ties, and queen posts. the point of the queen posts is that they help to transfer part of the load through the frame. in addition i put in purlins and windbraces. all of these additions help to stiffen the roof, to prevent collapse of the rafters, or spread of the walls, and to help prevent the rafters from racking (lateral and downward movement in a domino effect).
where the collar ties went was a matter of judgement. i felt they wanted to go up somewhere between a third and half way down from the apex of the roof. so i set the rafters out on the temporary deck on the first floor, and layed out where i wanted the collar ties and queen posts to go. moved them around on the layout until i felt they looked right. it was a feeling thing, not a measuring thing. the collar ties i felt needed a thru mortice, even thou they were only there to prevent compression, the action of the timbers drying could twist them out of alignment if the collar tie tenons were not deep enough into the rafters.
the queen posts were of a much smaller section and i felt could be restrained the same as all the other braces with a 3″ (75mm) tenon the depth of the post.
each pair of principal rafters was pre-assembled with all joints finished.
this included a thru mortice and tenon connecting joint at the apex. when ready each pair was dis-assembled, then lifted in section and set into position.
finally, when all assembled they were pegged.
the last part of the main part of the roof structure was to cut and assemble purlins (the logitudinal beams that run under the rafters, connecting them together), they were clasped into the collar ties where they met, and additionally held thru the rafters, and by windbraces.
the windbraces keep the spacing between the pairs of principal rafters, and help to prevent racking from wind load acting on the gable ends of the building.
cutting roofs is my favorite thing in carpentry, i don’t think there is anything better.
it is more than the end of the frame, it is the culmination of a lot of hard work, often the most complicated part, and the most rewarding, for me the roofing carpentry makes sense of the building.
in a previous post i suggested that there was something to be found in hard manual work, that it gave you certain understandings. it does something else too, it gives you a sense of fulfillment, and satisfaction at having achieved something, it gives you value, it gives you worth in a way that is different from any other way of achieving it.
making a shelter, providing warmth, nourishment, safety, these are the five principal basic needs of survival. fundamentally, building achieves these needs.
if you asked me why build with timber? or why did i choose to build with timber in the way i have? there are many answers i could give. its the material i have a lot of, its the material i know best, but i think the most honest answer from my point of view is its because it felt right. and it still does.