The Dome Story
Part 1: A Catalogue of Errors
Many people just want to tell you about their successes, but much of the most important information is gained by taking wrong turns and discovering what is down there. Mistakes are valuable lessons, and since my aim is to save you the trouble of making them again I must tell them to you or - if you are at all like me and look down every street that seems to go the right way thinking "why aren't people going that way?" - you will take every wrong turn I did along the way.
My journey with domes started in 2002 with (as with many people) a visit to the Eden Project in Cornwall, England.
These domes are massive - the largest is 100m across, made of a hex-pent structure filled in with 12m diameter 4-layer plastic inflated bubbles.
But size is not the most impressive thing - it's walking in from a cold early spring day to a steamy jungle, the completely different environment that it contains.
So I started to think about dome greenhouses and looking at dome structures that other people had made, mostly with round poles and blue-pipe ends bolted together.
My journey with domes started in 2002 with (as with many people) a visit to the Eden Project in Cornwall, England.
These domes are massive - the largest is 100m across, made of a hex-pent structure filled in with 12m diameter 4-layer plastic inflated bubbles.
But size is not the most impressive thing - it's walking in from a cold early spring day to a steamy jungle, the completely different environment that it contains.
So I started to think about dome greenhouses and looking at dome structures that other people had made, mostly with round poles and blue-pipe ends bolted together.
It was a couple of years till I started to actually make anything though, but I started big. My first dome structure was a 50% scale trial for the second, a 12.5 meter wide flattened dome that was not a true geodesic at all (though most people didn't understand the difference), but based on horizontal circles, and quite complex. I didn't like the bolted blue-pipe design because of the offset it produced and inability to put sticks on without loosening all at that vertex. I designed a plastic hub with the ability to hinge out but not in so that the top-down (no-ladder) construction method was still possible. I used plastic because I was into fairies at the time and the fay folk don't like Iron. The smaller trial had used copper but I couldn't think of a way to make the only copper I knew where to get (copper pipe) into a strong connection so I switched to plastic drainpipe for my material. It looked very impressive in a friends field in South Molton but I got my first lesson quickly, USE GOOD MATERIALS. I had wanted to make as much dome as possible for as little money as possible and had made a 12m structure for about £150, by cutting some low-grade plywood into 5cm strips. Plywood doesn't stay strong when wet, and some rain and sheep brought it down a week later.
But the lessons were learned well, and so I decided to try a local hardwood sawmill, bought some slices of a beech tree and tried again, on a more modest scale. This time I was planning a structure for just me and Annie (the previous structure was intended as an instant dwelling for an entire community) and also went for the more common 2-frequency hemispherical geodesic geometry. But unlike others I had seen, I decided that because it would have to carry snow-load I would stick with rectangular section wood. The hubs had worked quite well but had buckled under strain as the structure collapsed, showing me some weaknesses, so I overhauled that design too, and switched from poly-ethylene pipe to 4mm flat poly-propylene as my starting-point. This design has worked well, carried snow and people climbing on it for many years and the materials have withstood years of abuse out in the weather or otherwise in damp storage. The slot-together plastic in the hub is a weak-point though, and the flaps can snap off when dismantling it in the cold. I would not advise anyone build in exactly this way because it is excessively fiddly to do, requiring the cutting of >65 plastic strips with 4 accurately cut and angled slots and 2 accuratly drilled holes in each one. They are also not strong during construction, not strong in directions other than those they were designed to take once the dome is complete for the same reason - the slots create a weakness. This structure cost about £120 at the time (2005).
With a strong frame, we set about making a cover. We found some canvas in a local army-surplus-type store. It wasn't treated (although we didn't know this). This made it easier to stitch, but being heavyweight (about 23oz/yard²) it still wasn't easy. We used the biggest needle and the biggest thread we could find in the local sewing shop in our little singer sewing machine. I made an accurate pattern for the triangles and the fill-in hexagons (the roll wasn't quite as wide as the triangles were) and started to cut the roll into triangles and sew them together again into a dome. It took nearly 3 months to sew. There are 26 hubs - each had a hexagon, and 40 triangles - with all their corners cut off to match to the hub-hexagons. we decided to do French seams to avoid putting the entire thing through the sewing machine's arch. We would cut the triangles from the roll with the pattern, then sew them together, then go back to the roll and cut some more... anyone who works with canvas off the roll will know what's coming... The day finally came to try the cover on the frame. I'd measured everything perfectly, making the patterns to match the frame and including seam allowances, the patterns were made of chipboard and so couldn't change... but the cover didn't FIT! It didn't even get to the bottom of the frame, even with the door open, by over a foot all the way round! What had happened?
Canvas lesson 1: I didn't know that canvass shrinks by 10% in the 2 weeks after coming off the roll... most people unroll the entire roll and let it 'relax' (the rolls are rolled under tension) before cutting, so it doesn't shrink after cutting.
The simplest solution was to make the frame smaller, so I cut about 5cm off the end of every stick (we still have some of the wooden blocks left around the place, Annie kept them for kids to play with) and re-drilled the fixing holes. 5cm wasn't enough (I should have cut 15cm, but I didn't believe I could have got it THAT wrong) so we eventually (a year later) added a strip round the bottom. For that summer, having a 6 inch gap round the bottom was fine...
We didn't know whether it was waterproof. We left it up for a couple of weeks while we got ready to come to Italy for the second time. Oddly for South Molton it didn't rain the whole time it was up so we took it down and packed it for travel still not knowing. We got to Italy and put it up on the only level bit of ground here on this mountainside, still not knowing. We installed our futon. We slept a few nights, glad of the canvas between our ears and the racket from the crickets outside. It rained. It leaked. The canvas soaked the water through and the french seams provided drip-points. We'd had a plan, though - Annie had been a multi-level marketer for GNLD (an eco-freindly swedish chemicals and suppliments company) in a previous life and still had some product left, including something that was intended to seal floors for easy cleaning but claimed to waterproof canvas. She'd used it herself on a teepee in Australia and it had worked, so we put it on our canvas. It actually made it worse, now it was dripping all over the top where we'd put this stuff on as well as from the french seams. So: plan B. Someone had said to use beeswax and linseed oil like on a waxed jacket. We didn't know where to get that sort of thing here but we found a beeswax and linseed oil furniture polish in a local hardware store and put that on... worse again! We ended up spending the rest of the summer sleeping with a sheet of black plastic over our bed...
We went back to UK after the summer. Over the winter I built a floor for the dome - 20 radial 2×4 joists connected in the centre by 2 4-foot plywood decagons and 31 coach-bolts. The ends connected by more 2×4's and more coachbolts and PTG floorboards laid from the plywood centre out. All this is supported flat by 3 2×6's that each reach from one of the ten corners of the floor to another corner 3 sides round. This sub-frame allows the floor to be levelled on a slight slope by propping up the two corners where the 2×6's are connected, and on a steeper slope with a frame. This floor has survived well, and Annie and the dome are still on it in spite of it being left dismantled out in the weather for a year with only the plywood hub covered by a bit of plastic sheet.
Canvas lesson 1: I didn't know that canvass shrinks by 10% in the 2 weeks after coming off the roll... most people unroll the entire roll and let it 'relax' (the rolls are rolled under tension) before cutting, so it doesn't shrink after cutting.
The simplest solution was to make the frame smaller, so I cut about 5cm off the end of every stick (we still have some of the wooden blocks left around the place, Annie kept them for kids to play with) and re-drilled the fixing holes. 5cm wasn't enough (I should have cut 15cm, but I didn't believe I could have got it THAT wrong) so we eventually (a year later) added a strip round the bottom. For that summer, having a 6 inch gap round the bottom was fine...
We didn't know whether it was waterproof. We left it up for a couple of weeks while we got ready to come to Italy for the second time. Oddly for South Molton it didn't rain the whole time it was up so we took it down and packed it for travel still not knowing. We got to Italy and put it up on the only level bit of ground here on this mountainside, still not knowing. We installed our futon. We slept a few nights, glad of the canvas between our ears and the racket from the crickets outside. It rained. It leaked. The canvas soaked the water through and the french seams provided drip-points. We'd had a plan, though - Annie had been a multi-level marketer for GNLD (an eco-freindly swedish chemicals and suppliments company) in a previous life and still had some product left, including something that was intended to seal floors for easy cleaning but claimed to waterproof canvas. She'd used it herself on a teepee in Australia and it had worked, so we put it on our canvas. It actually made it worse, now it was dripping all over the top where we'd put this stuff on as well as from the french seams. So: plan B. Someone had said to use beeswax and linseed oil like on a waxed jacket. We didn't know where to get that sort of thing here but we found a beeswax and linseed oil furniture polish in a local hardware store and put that on... worse again! We ended up spending the rest of the summer sleeping with a sheet of black plastic over our bed...
We went back to UK after the summer. Over the winter I built a floor for the dome - 20 radial 2×4 joists connected in the centre by 2 4-foot plywood decagons and 31 coach-bolts. The ends connected by more 2×4's and more coachbolts and PTG floorboards laid from the plywood centre out. All this is supported flat by 3 2×6's that each reach from one of the ten corners of the floor to another corner 3 sides round. This sub-frame allows the floor to be levelled on a slight slope by propping up the two corners where the 2×6's are connected, and on a steeper slope with a frame. This floor has survived well, and Annie and the dome are still on it in spite of it being left dismantled out in the weather for a year with only the plywood hub covered by a bit of plastic sheet.
view in from the door of the dome as is was in 2006
Back to the canvas story...
On its new floor, we added a strip of canvas all round the bottom to come down to the ground and added a porch for standing entry to the frame. Then, before putting the canvas back on the frame we had another go at waterproofing it, this time with pure, raw linseed oil. It took about 8 liters to cover the entire 40m² of canvas. We put it on, it rained, no drips - we thought we'd won! We celebrated, put music on - goaded the clouds to test us - they spat ice from the thunder, but still no drips - so we started to hang the wool insulation we'd made from the inside. We'd sewn some raw wool we'd got into big flat quilts about 5-10cm thick. It smelt of sheep dags, but in a good way. we fitted a small but ferocious woodburner with a huge metal heat shield between it and the canvas wall. We stayed till February, waiting for the snow, which never came, and were very cosy, reading stories aloud and spinning wool in the nude in the heat from the burner and the insulation. We decided we had to go back to the UK eventually to MOT the van, we weren't going to be gone long so we left the dome up for when we returned.
Things did not go to plan.
The van was written off as uneconomical to repair by the pre-test mechanic due to rust in the chassis. We had a dome to build for a friend of a friend so we went and earned our money while the van was still legal (we did not use untreated canvas, we did not cut it into triangles, we did not use French seams, we made it in 2 pieces which laced together like a marquee, it had a removable hat like a yurt...) and then had a choice.
Option 1: downsize to a small estate car that someone had left on my brothers driveway to go join a cult (or so I was told, its amazing what people will call a cult who don't know that they're in one as well)
Option 2: learn to weld
I took option 2. I thought it would take me 6 weeks. Ha. Ha. Ha.
The Welding Story is for another page, maybe, but suffice to say it took me nearly two years to complete the work and get back.
On its new floor, we added a strip of canvas all round the bottom to come down to the ground and added a porch for standing entry to the frame. Then, before putting the canvas back on the frame we had another go at waterproofing it, this time with pure, raw linseed oil. It took about 8 liters to cover the entire 40m² of canvas. We put it on, it rained, no drips - we thought we'd won! We celebrated, put music on - goaded the clouds to test us - they spat ice from the thunder, but still no drips - so we started to hang the wool insulation we'd made from the inside. We'd sewn some raw wool we'd got into big flat quilts about 5-10cm thick. It smelt of sheep dags, but in a good way. we fitted a small but ferocious woodburner with a huge metal heat shield between it and the canvas wall. We stayed till February, waiting for the snow, which never came, and were very cosy, reading stories aloud and spinning wool in the nude in the heat from the burner and the insulation. We decided we had to go back to the UK eventually to MOT the van, we weren't going to be gone long so we left the dome up for when we returned.
Things did not go to plan.
The van was written off as uneconomical to repair by the pre-test mechanic due to rust in the chassis. We had a dome to build for a friend of a friend so we went and earned our money while the van was still legal (we did not use untreated canvas, we did not cut it into triangles, we did not use French seams, we made it in 2 pieces which laced together like a marquee, it had a removable hat like a yurt...) and then had a choice.
Option 1: downsize to a small estate car that someone had left on my brothers driveway to go join a cult (or so I was told, its amazing what people will call a cult who don't know that they're in one as well)
Option 2: learn to weld
I took option 2. I thought it would take me 6 weeks. Ha. Ha. Ha.
The Welding Story is for another page, maybe, but suffice to say it took me nearly two years to complete the work and get back.
Other people went to our place in Italy where the dome was still set up and tried to use it. the first was my mum and a friend of hers from Uni in the first spring. The dome seemed OK but the stitching had come apart round the extra bit at the bottom so the wind was blowing through. In the summer, some friends we were intending to set up in community with decided to go out ahead of us (through all the welding I never thought I was more than 6 weeks away from finishing). That summer, the canvas started leaking, the insulation started stinking so our friends burned it and then a hailstorm literally shredded the canvas [picture middle left]. Something had gone terribly wrong in just 6 months from when we'd left, because other ordinary tents that had been up at the time had NOT been damaged. When we finally got back the canvas was so weak I could poke holes in it with my finger, but not everywhere...
We don't know what actually happened to the canvas but it had turned black and weak in many places and had gone hard and leathery in others. It could have been the sun (someone had told me that in the Welsh teepee valley they rot from the bottom up but in the Spanish teepee valley they rot from the top down) and that would fit the pattern, the shredding was all on top, the sides were mostly fine, especially the skirt which hadn't been treated at all. It could have been the linseed oil itself, the beeswax polish or the PE based floor sealer. It could be that the canvas was slightly damp when we sealed it with linseed and like baby oil it sealed in moisture to this non-rot-proofed canvas. I lost heart for a long time. The dome stood there covered with tarps and carpets and more tarps and fragments of black plastic, ugly and draughty. |
Part Two: From the Ashes
It stayed another 2 years like that till I took it apart, dismantled the floor and started trying to move it all to a new place up the hill. In the end it was dismantled like that for about a year, some of the spars from the frame went missing even, ants made nests in the piles of floor-wood waiting for re-assembly. We got as far as putting the floor star on props in a new position and then I could go no further, broken by a symbol (for me) of hopes dashed and impossible dreams. It waited a while longer. A fire took our caravan. Annie set up the dome next to the top gardens, without its floor. A mushroom ended our relationship. It took the wisdom of a woodland fairy and the drive of a London girl to get it moving again - Annie needed a floor to set up for winter. I had given it to her, and this had cleared my head. Her needs moved me on. Winter here can be severe, she needed insulation. The sheep round here are not woolly, so I started to think expedience with winter approaching fast. Funds were limited, so it had to be cheap too. Polystyrene is nasty and also hard to fit to a dome, so I started to think about glass fibre when in the building supplies depot I was presented with a display of the G3-touch fibreglass with a box inviting you to touch it. Annie had already sewn together a whole lot of blankets into an inner cover and we had a new roll of sheet black plastic so I came up with a plan. We went home with 31m² (2 rolls) of the 5cm thick G3-touch fibreglass insulation for ~€70.
|
We took the covers off and dismantled Annies dome, moved the futon onto the floor about 30m away and reconstructed the dome over it with the inner cover and a big bit of black plastic over it in one day.
The next day we got that plastic in order by trimming and stapling it to the floor all round, except at the doorway where we stapled it to the doorframe. Then we started to fit the fiberglass by stapling it to the dome frame, through the black plastic. One roll went right round the wall, the other in sections over the top. The wall roll didn't quite reach the floor. Then we covered it with another layer of plastic, quickly - because it was dark and beginning to rain.
|
On the third day we re-did the outer plastic properly stapling it to the floor and door frame and put the 7m yurt cover back on. Because the yurt cover was too short by about 45cm we used the remaining canvas to extend it, folding the yurt cover into pleats as we went. The door is covered by two blankets sewn together with hollow-fibre pillow-stuffing inside, fixed to the doorframe down one side and able to flap open on the other. Gravity holds it closed. The yurt cover is held out into a porch by a pole tied up to a nearby tree. We fitted a window so she would wake up in the morning, and a woodburner for heating by essentially the same method - cut a hole in everything and then attach the edges to the thing you cut the hole for - making sure that the water would flow to the outside.
The primary purpose of the sheet plastic was to hold in any fibreglass dust from the insulation, but it also serves as a vapour barrier preventing condensation from the breath of the occupants. It has so far (4 years, including winter snow and summer heat) been excellent in both regards, the air inside is clean and clear, and there is no condensation to be found anywhere, even next to the floor where there is no insulation and very little airflow. That said, there has been not much cooking in there as Annie prefers her firepit outside, but she does habitually put a pot of goup (our name for linseed tea) on the indoor burner overnight in the winter, not boiling but warming. The insulation works fabulously, with just 3 2-inch sticks able to heat the space from cold and keep her warm all night, in deep snow and -4°C outside. Three more sticks in the morning and she could do her yoga naked in the morning. She stacked up a few sticks in the autumn, a loose pile about 3 feet by two feet by one and it lasted most of the winter because most of the time lighting the burner made it too hot to sleep. The canvas hasn't lasted though, it disintegrated a year ago, and I put an additional layer of (white) plastic to protect the most important waterproofing layer below. At some point I will grow moss on it.
Outside it's winter, but inside is snug and warm, with a dog curled up by the fire at night ...
|
I have since made another dome for myself with the same layering techniques (except that I fixed my insulation to the inner layer of plastic with tape, meaning that I have no staple holes in my inner plastic and can poke things in behind the spars) but a different kind of hub, which I have been living in for 2 years now. It is also easy to make too hot with the woodburner. These hubs are a cone of metal, set in a slot on the ends of the rectangular spars which are cut to angles so they fit together. The angles are quite precise, and I cut them with my bandsaw, but they could also be made by hand with a box guide. They meet at a 56° and 68° angles tilted away from flat by 15° at the 6-way connections, and 72° angles tilted by 18° at the 5-way connections. The spars themselves I cut 1" slices from 2" thick used scaffolding planks (chosen for strait grain from a batch of 200 we bought for flooring purposes) and in the process of making them it took longer to sand them (with an angle grinder sander and finishing paper) than it did to saw them and make the ends right. construction time was a total (for me) of 2 weeks but anyone else who doesn't suffer from a restricted amount of physical energy could probably do it in one. Building the floor took one week with one helper, and putting it up and covering and insulating and fitting the burner all together took one week more. This is the method of spar-connection I would recommend, if using rectangular (sawn) wood. there is also a clear plastic skirt which stops wind blowing up between the floorboards and gives extra warmth inside from the winter sun when the leaves have dropped. I have to admit, though, the black plastic outside is ugly - I should have put the white side out. I plan to grow moss on this one too.
|
A third dome has also been constructed (not by me) for Liz, but this one with Hemp fibre insulation. Hemp fibre requires airflow in order not to rot, so it has been contained between two layers of semi-permeable (plastic) membrane. The cover was intended to be canvas, but even with an air-gap the top sections still rotted (because even though it was "waterproofed" it still leaked, and so got covered "temporarily" with a piece of plastic for 6 months) and has been replaced with EPDM pond-liner (a kind of tough, weatherproof synthetic rubber). The main frame in this structure was done with 1½" roundwood harvested from this land and has "proper" doors, since Liz wanted to be able to control whether dogs and cats can get in (the padded flaps that Annie and I have on our domes have the opposite intention). It's poles meet at small sections of larger round poles and have dowels in the ends and small metal screw-in brackets holding them together. It is not a portable construction like the other two, the plastic membranes are taped together (they came as 1.2m wide rolls) and to the hubs, the liner fabric is stapled to the frame... dismantling and reconstruction in a new place would take a long time and lots of new materials. But it is warm with a woodburner, and it is (now) dry, and it is a beautiful space to be in, which liz uses sometimes to sleep, sometimes for doing her therapy and sometimes for teaching her therapy to small groups.
|
We also have a small yurt, also insulated with Hemp fibre but as an experiment we decided to do it without the plastic membranes. The walls have been insulated with a wood-fibre based solid insulation board about 4cm thick. It's been there for a year now with no problems in the hemp or the canvas that we can detect, we'll see how it is after a second winter when we take the cover off...
|
Well, that's the end of the story so far, but there's still a lot more to say because I've left out so much of the information and experience I've gained during this process, details of construction and methods and much more.
Part 3: The Details
The frame
There are 65 spars - 30 short and 35 long, in a ratio of 0.886:1 . The floor area is PI·r², the circumference 2·PI·r. The long spar is 1/10th of the circumference. calculate distances from fixing point to fixing point. include the size of the hubs, from fixing point to fixing point. if the 6-way hubs are different from the 5-way hubs include half the size of each in the short spar length.
I cut the spars of Annies dome from raw wavy edge beechwood slices from the sawmill with a small (46mm depth, 700W) circular saw. I marked which piece of the wood had the grain strait enough to be strong and cut that piece out, following the marked lines by eye wearing a full-face mask so my head could be right in the firing line of the sawdust and I could see where the saw blade was actually cutting. I marked the lines with a biro so I could see them. A bandsaw would also cut this well, but planks are big and heavy and awkward to manoever through a bandsaw.
I recommend adding a pillar drill to your toolkit. It allows fast accurate drilling of the fixing holes because a pattern can be clamped to the table that you just shove the spar end up to and pull the lever and the hole's in the right place. A pillar drill also makes drilling with a circle saw in inch-thick hardwood a possibility.
Although the dome can theoretically be assembled from the top down, the change in hub angles as the last layer flaps down by 75° is too great and the required number of people (5) has rarely been available. I normally assemble from the bottom up with a step-ladder or work-table to reach the upper fixings.
I cut the spars of Annies dome from raw wavy edge beechwood slices from the sawmill with a small (46mm depth, 700W) circular saw. I marked which piece of the wood had the grain strait enough to be strong and cut that piece out, following the marked lines by eye wearing a full-face mask so my head could be right in the firing line of the sawdust and I could see where the saw blade was actually cutting. I marked the lines with a biro so I could see them. A bandsaw would also cut this well, but planks are big and heavy and awkward to manoever through a bandsaw.
I recommend adding a pillar drill to your toolkit. It allows fast accurate drilling of the fixing holes because a pattern can be clamped to the table that you just shove the spar end up to and pull the lever and the hole's in the right place. A pillar drill also makes drilling with a circle saw in inch-thick hardwood a possibility.
Although the dome can theoretically be assembled from the top down, the change in hub angles as the last layer flaps down by 75° is too great and the required number of people (5) has rarely been available. I normally assemble from the bottom up with a step-ladder or work-table to reach the upper fixings.
Although the hubs described above have lasted for 10 years now, and been taken down and put back up several times, I do not recommend building a dome this way. I recommend the method I have used for my own dome, to cut the ends of the spars to points and connect them with a metal (could easily be copper if you wished) cone. The angles of the points are quite precise and are detailed in the (hand-drawn) diagram (full-resolution version here). The longer spars come to points of 68° when viewed at 15° from perpendicular and the shorter spars have a point of 44° at the 6-connect end and 72° at the 5-connect end, both viewed at 18° from perpendicular. The lengths should be measured from point to point. I have used M6 coachbolts to pin the spars to the cones and provide the needed support for the weight the spar must carry. Without the nut on the end, pressure from the outside splits the small amount of wood that is outside of the cone off. I did it this way so that it could be dismantled and rebuilt easily, but it requires that the holes for the bolts be pre-drilled with some precision to get them all the same distance from the points, and then also the holes in the metal drilled equally precisely. as you can see the distance to the centre of the cone in the metal was slightly large and this is holding the wood from actually touching as I intended it to, but it doesn't seem to matter. It would also be possible if you don't expect to need to move the frame ever to drill smaller holes during construction through both wood and metal and then use a screw to hold the pressure. Then you would not see nuts on the inside and risk them coming off. It would be possible to move a frame done this way, but you would have to put every stick back into every hub exactly as it was because no two stick-hub connections would be the same.
|
Canvas
A canvas dome cover which fits is a rare beast. I have yet to make one, though I have tried twice. You might succeed if you know what I have learned by failing.
An old singer sewing machine will sew a large canvas - I have got 3.5m width and 7 meter length through the 12 inch arm of a singer 99k that we got for £35 on ebay. New singer sewing machines have plastic parts and cannot take the strain. Ours was made in the 50's. we have used spun nylon '25' thread with a size 22 needle with this machine (the largest available for short needle machines), using a hand crank rather than a motor because it gives better control from the low speed. The eye of a needle is large enough for the thread if it can pass the thread 3 times. We have used a heavy-duty walking-foot industrial machine (for the 7m dome mentioned above) but owing to the speed it works at it is difficult to operate. A dressmaking industrial machine is not suitable because it is designed for speed in light fabrics - the motor is fast, the gearing is fast, the mechanism is made extra-light to cope with the speed and it will probably bend if faced with heavy canvas and a big thread.
|
Canvas shrinks by 10% (linear) or more in the 2 weeks after unrolling, known as 'relaxation'. Expect this, buy 15% extra, unroll it if you bought it still on the roll and wait the 2 weeks before measuring the width and finalising your cutting plan.
It takes 2 people to sew a large canvas with a static sewing machine, whether a modern walking-foot industrial or a 50's singer - one to operate the machine and the other to move the canvas past it. No sewing machine can move 50kg of canvas in a straight line on its own. Having a long thin workspace with a long thin table with the machine set into it in the middle would be ideal, but a series of different height tables putting the flat surface of the sewing machine in line with the table that it is feeding from is a good second. The 'hauler' must be as aware of where the seam is going as the 'sewer', if the hauler is pulling the wrong way, the sewer has little chance of putting it right. The hauler must get it just right - not too hard, not too gentle - otherwise the needle can get pulled sideways and miss its hole in the base of the machine or miss the thread-catch underneath, causing a clash and possibly bending the needle tip over. I have found it possible to re-sharpen a bent needle-tip with a fine sharpening stone. Polishing would then be helpful too. A strait 7m seam can be done in 5 mins, but preparing the canvas can take much longer - laying, rolling, folding on the feed table (concertina style).
There is no need to hem cut edges, use simple lap seams so you don't have to sew through more than 3 layers where two seams meet. A 3cm overlap is about right. Make sure all overlaps will overlap downhill to shed water (sewn seams are watertight to a depth of 20cm, but don't rely on this). All fabrics have the 'self-edge' at the edges where the weft threads fold back into the fabric. If you are overlapping self-edge to self-edge, the seam may be sewn just once, but two rows will make it stronger and neater. with cut-edge to self-edge, two rows of stitching are needed - one to fix and the other to stop the parallel threads unravelling till the crossing threads can pull out of the stitches. With cut edge to cut edge the seam needs three rows of stitches: the centre row fixes, the outer two rows preventing each of the two pieces from unravelling in opposite directions.
It takes 2 people to sew a large canvas with a static sewing machine, whether a modern walking-foot industrial or a 50's singer - one to operate the machine and the other to move the canvas past it. No sewing machine can move 50kg of canvas in a straight line on its own. Having a long thin workspace with a long thin table with the machine set into it in the middle would be ideal, but a series of different height tables putting the flat surface of the sewing machine in line with the table that it is feeding from is a good second. The 'hauler' must be as aware of where the seam is going as the 'sewer', if the hauler is pulling the wrong way, the sewer has little chance of putting it right. The hauler must get it just right - not too hard, not too gentle - otherwise the needle can get pulled sideways and miss its hole in the base of the machine or miss the thread-catch underneath, causing a clash and possibly bending the needle tip over. I have found it possible to re-sharpen a bent needle-tip with a fine sharpening stone. Polishing would then be helpful too. A strait 7m seam can be done in 5 mins, but preparing the canvas can take much longer - laying, rolling, folding on the feed table (concertina style).
There is no need to hem cut edges, use simple lap seams so you don't have to sew through more than 3 layers where two seams meet. A 3cm overlap is about right. Make sure all overlaps will overlap downhill to shed water (sewn seams are watertight to a depth of 20cm, but don't rely on this). All fabrics have the 'self-edge' at the edges where the weft threads fold back into the fabric. If you are overlapping self-edge to self-edge, the seam may be sewn just once, but two rows will make it stronger and neater. with cut-edge to self-edge, two rows of stitching are needed - one to fix and the other to stop the parallel threads unravelling till the crossing threads can pull out of the stitches. With cut edge to cut edge the seam needs three rows of stitches: the centre row fixes, the outer two rows preventing each of the two pieces from unravelling in opposite directions.
Even rot-proofed canvas will rot if kept in damp conditions for more than a few weeks. Our second canvas suffered this in the 'cool-room' on the ground floor of our house. If this has happened, it can be repaired without a sewing machine using offcuts and outdoor grade silicone sealer. For small holes we just glued the patches on with silicone, for larger ones we hand-stitched the patch on with one of us on each side using a back-stitch before glueing down the edges with silicone. Don't underdo the repairs. Look for weak places around the holes by trying to poke your finger through or tear the edges without much force. Neaten up the edges with scissors before repairing. These repairs have lasted a year so far in the weather including the process of moving the dome.
Fitting canvas to a dome frame
Getting flat fabric to fit a spherical object takes some getting your head around, so if you have not done some sewing before (or even if you have, a yurt or teepee is nowhere near as complex as a dome to get fabric to fit since a cone is still just a flat circle with a bit cut out) some techniques to try could be handy. I've collected here various methods I've used and thought of.
- Methods I've Tried
1. Cutting into triangles then sewing together.
This method is very laborious but produces a beautiful result. Any size of canvas can be fitted to any size of triangle by having hub-pieces which need to be made conical with a pleat or seam if they are much larger than the hub. The main areas are covered by triangles but owing to the fabric being narrower than the hight of a triangle a hexagon or pentagon hub-piece fills in. Getting the overlaps in the right direction can be complex. There is some wastage from the corners that get cut off, the excessive number of seams, and the odd-shaped hub-pieces. |
2. wrapping the map.
This involves designing a 5th of the cover as a piece, then stitching or lacing them together. I started with a single vertical strip then added pieces, mostly more strips, but cut and pleated to give conical points where the hubs of the frame were supposed to end up. This results in a fitted cover, but with much less cutting and stitching and less wasteage, but the sizes of the triangles of the dome must fit with the width of the canvas for the method to work well. The measuring and cutting are much more complex (to say nothing of the designing) but it is easier to get the overlaps in the right direction and one can make five identical pieces and stitch them together at the end. (The Map is the flattened out version of a 3D surface. Take all the triangles and connect each one by one side to the growing map, leaving gaps where the angles don't add up)
3. approximate
Fabric is flexible and stretchy. The fit only needs to be approximate, and it will stretch over the points. Sew a cone big enough to cover the top two layers at the right angle for the top part, or perhaps a little more (8° cut out or a bit more). Then add bits at the bottom in circles to manage the more vertical stuff, easing the seams together (easing means stretching one side whilst squeezing the other of a seam whilst sewing to get two edges that are not the same length to nevertheless fit), using the frame like a tailors doll to fit and pin bits on. The bits you add can have angled ends and the angles can match each other - there's no "inside side" on canvas so you can turn them over to fit. This method is probably the easiest as it involves little maths or forethought, and is just practical common sense. It could be the most efficient too, if your cone isn't too big (as ours was) and needing darts and pleats. If your dome has thick fluffy insulation, the points almost disappear and the approximate method is much the best fit. The red and green cover you see on Annies dome above happened this way, although by accident. That's how approximate it can be!
4. Cross and fill (another approximate method)
Make a cross with arms long enough to reach to the ground on opposite sides, about 2/5ths of the total diameter wide (2×90cm strips wide each arm for our 4.8m dia. dome). Drape it over the frame and then cut pieces and pin them in place to fill the gaps. Take it off the frame and sew all the pinned pieces on. It is probably necessary due to the weight on weak pins to do one row of pinning at a time, sew them on, then re-drape over the frame and pin some more. I believe this is the method that was used on Liz's cover (I wasn't directly involved) and it worked well (except for shrinking before it was stapled down... Oh no! not again!). This is the method I would use if I was to make any more canvas covers (see the note at the end about why I probably won't).
This involves designing a 5th of the cover as a piece, then stitching or lacing them together. I started with a single vertical strip then added pieces, mostly more strips, but cut and pleated to give conical points where the hubs of the frame were supposed to end up. This results in a fitted cover, but with much less cutting and stitching and less wasteage, but the sizes of the triangles of the dome must fit with the width of the canvas for the method to work well. The measuring and cutting are much more complex (to say nothing of the designing) but it is easier to get the overlaps in the right direction and one can make five identical pieces and stitch them together at the end. (The Map is the flattened out version of a 3D surface. Take all the triangles and connect each one by one side to the growing map, leaving gaps where the angles don't add up)
3. approximate
Fabric is flexible and stretchy. The fit only needs to be approximate, and it will stretch over the points. Sew a cone big enough to cover the top two layers at the right angle for the top part, or perhaps a little more (8° cut out or a bit more). Then add bits at the bottom in circles to manage the more vertical stuff, easing the seams together (easing means stretching one side whilst squeezing the other of a seam whilst sewing to get two edges that are not the same length to nevertheless fit), using the frame like a tailors doll to fit and pin bits on. The bits you add can have angled ends and the angles can match each other - there's no "inside side" on canvas so you can turn them over to fit. This method is probably the easiest as it involves little maths or forethought, and is just practical common sense. It could be the most efficient too, if your cone isn't too big (as ours was) and needing darts and pleats. If your dome has thick fluffy insulation, the points almost disappear and the approximate method is much the best fit. The red and green cover you see on Annies dome above happened this way, although by accident. That's how approximate it can be!
4. Cross and fill (another approximate method)
Make a cross with arms long enough to reach to the ground on opposite sides, about 2/5ths of the total diameter wide (2×90cm strips wide each arm for our 4.8m dia. dome). Drape it over the frame and then cut pieces and pin them in place to fill the gaps. Take it off the frame and sew all the pinned pieces on. It is probably necessary due to the weight on weak pins to do one row of pinning at a time, sew them on, then re-drape over the frame and pin some more. I believe this is the method that was used on Liz's cover (I wasn't directly involved) and it worked well (except for shrinking before it was stapled down... Oh no! not again!). This is the method I would use if I was to make any more canvas covers (see the note at the end about why I probably won't).
- Other Methods I've thought of but haven't tried
1. Three cones
Extending from the approximate method, make 3 conical sections to fit the 3 levels of the frame (15°, 45°, 75°).
2. Five sections
This is similar to the wrapping the map, but approximate. start with a vertical strip, then add on more, but with them curved progressively narrower toward the bottom, to get the right amount of width at the bottom edge. The angle at the top is less than 72°, forming the correct cone at the top vertex. for a two frequency dome the angle is 70.5°.
Extending from the approximate method, make 3 conical sections to fit the 3 levels of the frame (15°, 45°, 75°).
2. Five sections
This is similar to the wrapping the map, but approximate. start with a vertical strip, then add on more, but with them curved progressively narrower toward the bottom, to get the right amount of width at the bottom edge. The angle at the top is less than 72°, forming the correct cone at the top vertex. for a two frequency dome the angle is 70.5°.
Fitting the Woodburner
There are a few ways to put a flue-pipe through a canvas roof but what I did is to combine the traditional flashing plate and storm collar into one big cone. The issue with a continuous sheet is that you can't get the flashing plate to overlap in the right direction at both the top and the bottom. The flashing plate, though, usually has a raised lip near the flue-pipe to prevent the water from running in below the storm collar, which is something that canvas (and sheet plastic, too) can be made to do. I got some galvanised steel sheet (aluminium is easier to work but much more expensive and conducts heat 5 times faster) and made a cone with an angled base. From more steel sheet I made an internal heat shield large enough to fix in 2 places to the dome frame, with a hole about 3cm smaller all round than the bottom of the cone. The inside of the hole in the heat-shield I then cut into tabs about 5cm×3cm which could be folded up exactly into the cone. Then I cut a hole through all five layers of the dome to match the one in the heat shield (without the cuts for the tabs), and cut away another 2" all round from the insulation layer so that the four remaining layers could be pushed together. Next I put everything together but without the flue-pipe in place - cone outside, heat-shield inside, and with a drill made holes through both bits of metal, one for each tab, starting from the top and working down the sides, for M3 set screws to go through. The screws go through the heat shield, then the four layers of cover, then the cone, before having someone outside put a nut on and the whole thing is tightened up. The fabric and plastic layers are pinned between the cut tab of the heat shield and the cone, stretched up so that water running down the outside must go round the hole. The stretching is strongest at the top and almost none at the bottom, so I decided that not every tab needed screwing up at the bottom.
The metal shape for the cone can be drawn using a half-circle for the flue hole slightly more than twice the diameter of the flue you will use, and then half an ellipse for the bottom edge with the distance to first circle at it's closes approach of about 30cm. this will be the length of the upper (short) side of the cone. If the diameter is double this minimum at the edges, then the cone will have a 45° angle at its base. Draw the ellipse using a piece of string or thread the length of the long diameter (right from edge to edge). Get someone to hold the ends firmly so that the centre of the free part of the string (between where it is held) can be just pulled up to the centre of the short diameter. Then run the pen pulling against the inside of the string along the arc this generates, all the way to the end-points of the ellipse. If it is 32cm from the straight edge of the metal to the top of the curve, then each side will be 64cm from the centre and the string will be held about 9cm inside on each side. Mathematically, the holding points in this example should be exactly 32×√3 = 55.4cm each side of the centre. Leave an extra 2cm of metal below the straight edge for a fixing flap. I have folded both flaps up at 90° and pop-rivited them together, which leaves little chance for water to get between. When you have bent this piece onto a cone shape and fixed it, draw the oval it makes on the triangular bit that will fix into the frame. I have made this triangle a 1m equilateral to fit into one of the long-sided triangles, with flaps on all three sides, two to fix to the frame and one to fold down to make it more rigid at the top. draw another oval 3cm inside the first and cut out the inner oval. then cut slits at 3 to 4 cm intervals out to the original oval. Bend them up so they match the inside of the conical section, match them together, then start drilling holes through both for the fixing bolts. I have always used 30mm long M3 set-screws for this, anything more is overkill. The length is important though, they have to reach all the way to the cone with the flap still quite flat during fitting, and then they pull the flap and the layers of the covers up during fitting. |
Fitting the Window
I used a similar method of stretching for the window, but because it rests on the floor the cover didn't surround it completely. The window was originally just one layer of glass but we had some flexible PVC window material so we stapled that on iether side of the frame to make it triple-glazed, folding it down round the sides on the side that would be outside. Having cut a window sized hole, we stretched the plastic up and over the frame and the outer window plastic so it just tucked over the outer top corner without actually breaking. the canvas we didn't stretch, just brought it directly to the outside top of the window frame and then stapled it on all round. So far (3 years) no leaks. The stapled triple glazing gets condensation inside, but it evaporates again.
Some Thoughts on Canvas
Canvas is not an ecological alternative. Most canvas is cotton, grown on land that should be feeding its inhabitants at enormous cost to the environment consuming vast resources of fertilizer and water, sprayed with pesticides. It must be treated with more toxic substances to prevent it decomposing quickly, and yet more chemicals to make it waterproof and fireproof.
Much is made of canvas' ability to "breathe", being porous it allows exchange of gasses (oxygen, CO2, water vapour) from inside to outside. Proofed canvas has only a tiny fraction of this ability left, and the longer lasting canvasses that get used for marquees and curtain sided trucks (and some yurt roofs) have been laminated in plastic which prevents breathing completely. Our most recent dome has 2 unbroken layers of plastic inside the canvas - I'm not relying on the canvas staying waterproof for long - the canvas is just decoration and protection for the plastic from falling chestnut prickles. The door is the only breathable part and that is not waterproof at all, its under the porch. In winter the woodburner keeps the airflow moving, I'm not sure if its pulling air through the door or from cracks round the bottom of the plastic between staples. There doesn't seem to be a problem without the woodburner either - even in summer heat with the door shut against insects there's no sense of airlessness.
Needless to say, I am looking for affordable alternatives to treated canvas for decorating and protecting future domes.
One alternative which comes immediately to mind is to grow a living cover - vines, ivy etc. (but how to keep the rodents from chewing in? How about 3 outer layers of plastic and ½" chick-wire as the trellis?). I am going to try a moss cover, seeded the same way as for moss-art on walls (moss mixed with yoghurt painted on), using garden fleece material for it to stick to. Another alternative is tree-bark (a bit labour intensive, perhaps. Still needing wire rodent-proofing even without ground contact? would give something to attach the bark to). Lap-boarding? Plaster? Plywood? Plastered Plywood Lap-boarding? Hempcrete with a lime render?
Much is made of canvas' ability to "breathe", being porous it allows exchange of gasses (oxygen, CO2, water vapour) from inside to outside. Proofed canvas has only a tiny fraction of this ability left, and the longer lasting canvasses that get used for marquees and curtain sided trucks (and some yurt roofs) have been laminated in plastic which prevents breathing completely. Our most recent dome has 2 unbroken layers of plastic inside the canvas - I'm not relying on the canvas staying waterproof for long - the canvas is just decoration and protection for the plastic from falling chestnut prickles. The door is the only breathable part and that is not waterproof at all, its under the porch. In winter the woodburner keeps the airflow moving, I'm not sure if its pulling air through the door or from cracks round the bottom of the plastic between staples. There doesn't seem to be a problem without the woodburner either - even in summer heat with the door shut against insects there's no sense of airlessness.
Needless to say, I am looking for affordable alternatives to treated canvas for decorating and protecting future domes.
One alternative which comes immediately to mind is to grow a living cover - vines, ivy etc. (but how to keep the rodents from chewing in? How about 3 outer layers of plastic and ½" chick-wire as the trellis?). I am going to try a moss cover, seeded the same way as for moss-art on walls (moss mixed with yoghurt painted on), using garden fleece material for it to stick to. Another alternative is tree-bark (a bit labour intensive, perhaps. Still needing wire rodent-proofing even without ground contact? would give something to attach the bark to). Lap-boarding? Plaster? Plywood? Plastered Plywood Lap-boarding? Hempcrete with a lime render?