Geyrfugl – building

Nick Schade’s book purports to be all a complete novice needs to read to be able to build a beautiful and functional cedar-strip sea kayak.

Geyrfugl from astern - click for bigger imageWell, I had never met or spoken to anyone who had built a cedar-strip kayak, nor had I ever seen such a boat until I had completed my own first project. Whilst I cannot deny having read rather more widely than Nick’s book, and, in my next boat, used some of the techniques noted in other sources, Geyrfugl was built using only methods and designs described in “The Strip-built Sea Kayak” or of my own invention based on that. Nick says that the book will enable “anyone who has ever picked up a hand tool” to build such a boat. My five-year old son was always picking up my hand tools (especially any hammers left unattended for a moment) and perhaps such a boat was beyond him for some years, but in essence, all you need is that book and a generous supply of concentration and patience.

Having cut the forms and threaded them onto my strongback, I found getting them accurately aligned to be a lot harder than I imagined from reading the book. After a day and a half, I started again from scratch, using a slightly different technique, and using thin wedges to hold the forms in place and move them gently by a fraction of a millimetre at a time. This was more successful, and after another day or so, I was satisfied. It later became apparent that one form in the middle was slightly twisted out of alignment (I think this must have happened after I started stripping) but one would have to know this and look for the consequences to actually spot the problem in the finished boat.

Since I was not using cove-and-bead strip, and my strips, being cut from 3m planks, were shorter than the boat, I had a potential problem in joining strips together and keeping them aligned. Initially I approached this by making through-the-thickness scarfs to make strips longer than the boat, but I only had enough space and clamps to do a very small number of strips per day this way, which was keeping progress very slow at first. On a future boat, I would probably do this sort of scarf for the sheer strip and maybe the next, but after that I would go to butt-joins or across-the-width scarfs glued in situ.

Geyrfugl bow showing wide spruce cheater stripTo avoid too much bending in strips at the stem and stern, Nick suggests using “cheater strips”, which also help later on to ensure that the transition at the keel occurs at the same level fore and aft. Here I deviated from Nick’s recommendations in two ways. The first strip below the sheer was laid much straighter, and I then filled in the gap with several such cheater strips, rather than alternating long strips with single cheaters. I think this works OK. My cheater strip nearest the sheer was of a contrasting material (spruce) in a wider strip. This was not as successful – not so much for the contrast of material (which is used as background to the name of the boat) but for the width of the strip, which made it difficult to plane and sand a fair curve in the bow.

I also found it very hard to make the transition from stripping down from the sheer to stripping out from the keel at all tidy. Indeed, I ended up with an unsightly gap right in the keel line. This was readily dealt with later by adding a hardwood stem piece, but fitting this involved planing down so far that I could not easily avoid getting epoxy onto the end form, which needed some very careful masking in order to get a hull which would actually be possible to remove from the forms. In future I would be tempted to make an inner stem piece and glue the strips directly to this, as Nick describes as an alternative technique. (In fact this was academic on my next project: a seaboat scaled down slightly from Squeedunk’s Cormorant 16, with a stitch and glue hull and stripped deck and the Borealis project will involve a very different stem and stern construction).

With the hull stripped, I immediately went on to strip the deck, with the hull still on the forms. The design I had adopted has a very abrupt transition from the hull to an almost flat deck near the stern and a peaked deck near the bow, but with a much more rounded transition near the cockpit. This is actually quite a hard thing to do, as you need either a lot of twist on one or two strips, or some other way to get all the angles to blend in. Perhaps one way would be to use some narrower, but thicker strips, and then get the shape fair by sanding later. However, what I actually did was add a strip the full length of the deck, adjacent to the sheer at bow and stern and some way above the sheer, where the deck angle was similar, near the cockpit. I then filled in the gap between this strip and sheer, which involved a great deal of planing difficult bevels onto long thin strips later. I suspect this was a lot harder than I could have made this part of the job, but by this stage I was gaining confidence in my ability to get difficult strips to fit, and the result certainly looked OK when the boat was finished.

Having got this first strip above the sheer, I then put a double strip along the deck peak, added some stringers to the forms, and stripped between them with a series of short diagonal strips to form a broad chevron pattern on the deck using three contrasting woods (red cedar, mahogany from an old piano and spruce left over from some building work – I didn’t actually buy any timber at all for the boat). This stripping pattern involved a lot of glued joints between edge grain at the ends of the short strips, and the long smooth sides of the longitudinal strips. This was to prove to be a problem later in construction, though not in the finished boat.

After stripping both hull and deck, a niggling worry now loomed large. I knew that I had got glue onto the end forms during stripping, beyond the area which I’d masked with tape, and I was anticipating having a lot of trouble getting the hull off the forms. There was also a risk that I’d got epoxy onto the end forms when adding the hardwood stem pieces. I reckoned the deck would come off a lot more easily. So I went ahead and faired the hull and deck with plane and then sanding and then decided to remove the deck from the forms first – without glassing it. In fact, all those short strips had involved a lot of glue, and the deck was very thoroughly stuck to the tape over the forms – getting the deck unstuck was really hard! Doing this did indeed make it a lot easier to get the hull off the forms later, but in the meantime two untoward things occurred. Firstly, one of the seams between edge grain of short strips and a longitudinal strip on the deck peak split completely, and a few other glued joints looked pretty fragile. Secondly, work on our house electrics resulted in the barn, where I was building the boat, being without any power (as it remained, for several months). This meant that I stopped work on the boat with the (damaged) deck off the forms, for several weeks in autumn when everything started to get cool and more humid. I do not recommend this !!

My resin and glass arrived, and it became clear that power to the barn was not going to return. Autumn was heading for winter. I discovered that such a short boat could just be fitted into a garage which I could heat, and so the hull was moved and left to get warm and dry for a while ready for glassing. I think this was the step I was most dreading, as it seemed to me that anything going wrong here would not be easily mended. In fact, Nick’s instructions are more than detailed enough, and the working time of West System 105/207 long enough that this was a much easier task than I had anticipated and I would urge anyone put off by the prospect of glassing and epoxying a boat to stop worrying – this is far from the hardest part of the job ! I laser-printed the name “Geyrfugl” onto tracing paper and applied this to the bow along with the first sealing coat of resin.

Once the outside was glassed, I was able to remove the hull from the forms with a lot less trouble than I had expected, and soon faired and glassed the inside. I had eventually decided on a slightly heavier lay-up than my initial ideas, with 4oz on the outside, doubled on the bottom of the hull, and a single layer of 6oz glass on the inside. However, perhaps during this stage the hull was still in transition from the cool humid conditions of the barn to the warm dry conditions I had found for doing the resin work. It transpired that the hull had curled inward very slightly and after glassing the inside, the beam was less than it should have been. However, I did not
realise this until rather late in the process…

Ebony veneer detail I now returned the deck to the forms and cleaned up the damaged glue joints. A few staples held the deck back on the forms whilst these joins were reunited – using epoxy this time. The big long joint which had split was repaired well enough not to show, but a small area towards the rear of the cockpit where more strips also joined at odd angles didn’t fare so well. The rejoined edges were far from level and looked a bit tatty. I sanded these down flat, and covered an area each side with a veneer (of Macassar Ebony) cut and laid parallel to the chevrons of the aft deck. The veneer stands slightly proud of the surrounding deck area, but this is partly offset by butting one layer of glass up to the edge of the veneer where the glass from fore and aft decks overlap. All the tattiness is hidden and the very dark veneer fits well with the design of the rest of the deck.

Great Auk logo on Geyrfugl's foredeckOne of the broad chevrons of light-coloured spruce was chosen to receive a Great Auk graphic from the Royal Ontario Museum website – one each side of the deck ridge, laser-printed onto tracing paper as for the boat’s name and a “maker’s plate” inside the hull.

Once I’d glassed the deck, I cut the cockpit recess (I know this is the wrong order, but it was clear that the deck wasn’t strong enough to do this until I had glassed at least the outside). Upon taking the deck off the forms again and trying the fit with the hull, it became clear that there was a big mismatch. Given my experience with the deck, I assumed that I had allowed it to flatten and become wider than the hull, and started taking steps to make the deck peak angle more acute. Big mistake ! If something is wrong, measure first before assuming where the problem is. Once I had realised that it was the hull that was wrong (mainly because I couldn’t get the deck to fit at all and was forced to measure everything) I had quite a bit of glass to remove from the inside of the deck to undo my attempted fix. The correct solution was to wedge the hull wider to fit, at the stage of joining hull and deck, but that came later.

Another step that sounded a lot harder than it proved to be was building the cockpit coaming from lots of short strips. This went in two or three sessions of only a couple of hours each. I chose wood to match the deck at each point, and made a few narrow strips to get the transitions at just the right points. I think another time I might not do this – the spruce strips later proved to be more difficult to sand fair than the harder mahogany which adjoined the ebony veneer. I put a 7 or 8 cm strip of glass cut on the bias around the outside of the coaming (in three overlapping sections so I could get each one fairly neat within the working time of the resin). This made the whole thing strong enough to mean that I could now handle the deck a bit less carefully, which was useful, as my earlier mistakes meant quite a bit of fairing on the inside of the deck. A bit of fairing putty filled a few divots which had been created when removing glass earlier, and after a further resin coat all over, the inside deck looked quite acceptable again by the time it was glassed.

Aft hatch with cover offCutting the hatches was less nerve-racking than expected. I used a power padsaw, but rather than lowering this onto the deck whilst running, as illustrated in the book, I simply started the cut by using a blade from the saw by hand. I used 4mm plywood for the hatch lips, cut in two pieces for each hatch. Detail of aft hatch reinforcement The hatches felt strong, but I worried about the risk if they were stood upon when off the boat, so reinforced them with two triangular pieces of hardwood inside, glassed over. For each hatch cover, one of these bits of ash had a hole cut to attach a retaining cord. Again, I hadn’t needed to buy this hardwood – all my ash was cut with the bandsaw from branches removed from trees on our property. This particular branch had seasoned by lying next to the compost heap for a couple of years !

Forward hatch with cover offThe boat almost looked ready to mate hull and deck, but in fact there was still quite a lot of work. I used some old 5/16 inch PMI pit rope (an originally white kernmantel nylon caving rope, some 20 years old and needing retiring from mission-critical load-bearing use 🙂 for the cockpit rim. Nick says that it is hard to get a watertight seal between a rope rim and a spraydeck, but I presume this applies to a laid rope. Once glassed over, the kernmantel rope gives a nice smooth surface, though it is quite difficult to get it at a constant height above the deck when tacking it in place with hot-melt glue. Perhaps this would be easier with new rope than rope old and somewhat kinked from much heavy caving use ! The details of the nineteen recessed deck fittings and steel wire security fitting are elaborated on their own page, but each fitting took quite a while. Drilling out inch diameter holes in the deck proved to be quite nerve-racking !

A series of pieces of wood cut to length and taper, and hammered into the hull got the width to match the deck quite closely – no way could fibre-reinforced packing tape exert this amount of force ! Access to the inside seams was made rather difficult by the recessed deck fittings, so that I elected to apply a series of short pieces of fibreglass tape on top of a fillet painted into the join, rather than Nick’s idea of soaking a large roll of tape in resin and laying it out all in one go. The size of the cockpit and hatches also meant I couldn’t get both head and arm into the inside of the kayak at once and reach into the ends, so the tape inside stops quite a way short of the ends of the kayak. An extra resin fillet went into these bits after the outside seam and end-pours were done.

By contrast, the outside deck-to-hull seam was easy, and done with a single piece of tape round the whole boat, which made the bows slightly neater. It was only at this stage, with the deck in place and recessed deck fittings finished, that I filled the weave of the glass outside the deck. Doing it this way made a squeegee pretty hard to use, but on the other hand, I don’t think I would have avoided scratches on the finish if I had filled the weave soon after applying the glass. Recessed deck fittings look great, but they certainly make the deck quite a bit more difficult to build.

The boat was now basically sea-worthy, but I managed to resist the obvious temptation long enough to do the end toggles and both forward and aft bulkheads, which were of unglassed 6mm ply. I had planned to put the forward bulkhead at a position which would make the boat a good fit for my wife, but had overlooked thinking about this when doing two of the deck fittings. They were in the way, and this bulkhead had to go a few centimetres further toward the bows. A consequence of this is that I can actually fit my legs into the boat, though I am some 40kg heavier than the largest intended paddler, so I did’t expect to be able to use it (but see the Geyrfugl in use page). The small amounts of resin left after fitting each bulkhead were used to add four pad-eyes inside the boat – two for the shock-cord to the back of the backstrap, and one for each hatch to tie the covers to the boat.

Newly finished Geyrfugl with my existing North Shore MistralWith bulkheads in place I sanded down the outside, creating a great deal of fine dust, so moved the boat into our conservatory for the first coat of varnish. It took only a few minutes to add the hatch seals, and then the various deck lines and retaining straps. I had built a back strap and this fitted neatly onto the cords already in place on the cheek plates. This was Monday 2002-02-11, which proved to be sunny, so no time was wasted in getting the “finished” boat outside onto the lawn and finishing the films in two cameras.

I had originally intended to fit footrests to make the boat usable by my growing daughter, and I needed a moulded foam seat as well as various bits and pieces for other kayaks, so a trip to Knoydart kayaking systems in Keswick on the next day provided an excuse for a short “sea” trial on Derwent water. The experience is included in the Paddling Geyrfugl page.

Well, snow will have to do until the water warms up! Does this count as white water, dad ?In the week since the trial launch, I had the deck lines off again and applied two more coats of varnish. The foam seat was in place, but needed to be trimmed down a little. Having finished in February – still the white water paddling season in the UK – there were still quite a few weeks for the finishing touches before the weather got warm enough for a real maiden voyage by Geyrfugl’s intended paddler.

By the end of August, all the family (and quite a few friends) had paddled Geyrfugl, and the conclusion is that she is fast, stable for small paddlers and fun to paddle. She surfs on pretty small waves and tracks really well for all but the smallest of paddlers who suffer somewhat from windage. Sarah and Michael could both climb into her from deep water, and the hatches don’t leak, even with the boat upside down, and two children stood on the hull punting her along with a paddle (yes, it was that kind of a summer:)