Compound plywood bend in stitch&glue construction?

I’m only going to add to Rumars excellent post above by saying I’d not likely want to use 8mm foam. I see too much opportunity/need to prefair before glass and make it too thin, so 10mm all the way to 12mm for me, depends a bit on expected forming after glued up.

Another thing to add is the idea of a female mould or jig (he mentioned above) is not awful here. Wood stations of osb or pressboard would allow the hull to be planked, and glassed inside, then floors and tabbing (as needed) and even sole (if desired) and then the boat can be flipped and outside glassed without losing her shape; if you are careful not to set on the wings.

I would only use 300g biaxial over foam and 150g woven over wood strips with local reinforcement of the bottom inside and out with the 150 and perhaps more local reinforcement of the wing area if it is floppy. The wood boat is stiffer; I’m not doing the math for you on weights. This boat in foam really needs vac bagging. Rumars has used 12 sqm for the hull. This means 24sqm for glass at 300g is 7.2kg glass and 7.2kg resin. I think about 10kg foam at 10mm thick. Using vac and breather can pull about 3kg out and improve the bonding, but it is hard work to infuse or wet bag this boat. But hull alone in foam comes in at 25 kg no frames or glues. Moving down to 8mm and glass up to 450 is 7.7 core and resin and glass of 21.6kg or 29.3kg plus some glue. Going to 12mm core with 150g glass plus double bottom both sides is about 225gsm net glass or 16.2kg resin and glass and core of 11.5 or 27.7kg hand laid no vac needed. This is the same glass I’d use for 6mm wood core with weight est of 30kg for a total of 41.5kg for the wood hull no frames or sole or trunk. You need to load the math in a spreadsheet. I’m entitled to error here late at night!

never use green wood for encapsulation; it would also be unstable on the jig and the strips will twist out of shape before you glue them ~

ps ~ the foam I’d use would have properties close to Gurit Corecell M80 for the light woven skins. I would make sure to turn the second layer of glass so the tows ran 45 degrees to the first layer..with overlaps of 50mm min

 

For a comparison that might be useful,10 sq.m canoes are a little longer and have a minimum weight of 50Kg so that ought to give a figure to aim for.There are a few items that need to be considered with any construction method that I haven't seen mentioned so far;Internal framing will have to be present to support the double bottom and resist forces from the mast and the weight of the sailor.Additionally there will have to be some method of attaching a deck,so what material would the deck be made of and how would it be secured to the hull?The rudder will need more support than a very low transom can provide so how is it intended to provide a point for the upper gudgeon or pintle to attach to?A gantry is certainly possible but would require local reinforcement,but then the bottom of the hull might need that too.

There are practical matters too,several days have passed with all sorts of topics being discussed and the sailing season is getting closer very quickly.If the boat is to see the water this year things need to start happening and we have a perfect example of the way in which any boat is a combination of compromises combining availability of materials,practicality of construction,time and desired performance.Does it become easier if you can locate better materials for the boat but have to spend a long day driving to collect them?

 

Thank you guys for the wealth of info you keep providing. This topic has evolved way past anything I expected, and I recognize now that I should have started with a clearer explanation of why I'm building this boat. Better late then never I guess.

Basically, my first boat was a 3.1 m. S&G single-bottom lugger (link to an article about the build with photos), built with birch ply. That one came out at 35kg hull weight, 45kg fully rigged. It as a joy to use because of it's low weight, extremely fast in low wind conditions, very easy to car-top single handed. However, the waterline was too short, the hull shape not very hydrodynamic, 4mm bottom wasn't reinforced enough and kept cracking, and capsizing was a constant issue due to lack of open transom.

My most recent boat was a 4.2 m. flat bottom, built with okoume ply, tab-and-slot construction (link to photo album). Hull came out at ~55kg, 70kg when fully rigged. I totally overbuilt her, and she is a tank despite 4mm ply everywhere, but WAY too heavy to be car-topped single handed, slow in light winds, and the hard chines don't help on the water either. However, the double bottom was a bliss, never going back to single bottom again. She did have some compound curvature in the panels, but the whole build process was way overcomplicated and took me 2 summers to complete.

So, learning from these experiences, I wanted to design a racing boat combines the best from my previous experiences with the following requirements:
-4.5m hull length;
- <35kg hull weight, 40kg absolute maximum;
-Double bottom;
-Very narrow, canoe-like waterline (L/B ratio of 6 or more) and flares/wings to allow effective hiking;
-Smooth chines or otherwise hydrodynamic hull;
-Resistance to punctures, collisions, chafing and otherwise rough handling of the boat (not the way I want her to be handled, but the reality of our sailing venues and regatta culture)

Budget ~1500€, same as for my previous boat (actual total after the build was completed).

My initial idea was to use 3D printed core in a carbon fiber composite sandwich, borrowing the idea from Jan Herich, and on paper the boat I designed would fit all these requirements fairly easily. The boat would be 4.5 m LOA, 1.6m BOA, 4.4m x 0.75m waterline, with a free-standing mast and 10 m^2 lug rig. I did some testing in the materials lab in local university to validate 3D printed core as a suitable material for this use case, but due to insufficient/flawed testing methodologies and my general lack of understand of how composites work, I overestimated the compressive and shear strength of 3D printed core, and only after I built some larger parts of that boat I realized how ridiculously weak it is, and after numerous attempts to fix the issues, had to scrap it. It wasn't a total loss - at least I can build spars and some other small parts this way, since they don't rely on the core, it could even be removed afterwards. However, the hull needs a complete redesign.

That's how I arrived at this point - trying to find a method to build this boat that still fits the requirements I listed above.

For reference, here are a few pics from CAD of the original design to convey the idea of the hull shape that I was going for.






Unfortunately that idea failed, which is why I'm here, trying to build something similar within the requirements for this summer. Yeah, the time pressure is real, and I know my requirements could be impossible to fulfill without some compromises.


Regarding the availability of materials, I do have Okoume (Joubert) available. If I were to build the boat with it, I'd likely use it over birch. I got the impression that is not as strong, but weight savings are definitely more important. Available in 4mm, 6mm, 9mm, etc. Price is affordable. Also, just this morning I found one supplier who advertise they can get Canadian Cedar. I said that it's not available in Lithuania because just a year ago I did a thorough search for it, and came up with nothing, so this must be new. I will call them on Monday and try to find out if it's anywhere near affordable. Rumars, I will call Diab on Monday too and try and figure out if I can get some of their foam straight from the factory, maybe some written-off batch or something. Otherwise, I don't see many alternatives for foam, since the only real options here are XPS (big no-no) or PVC (which is not the same as your PVC, the one I have available is the same weight as lumber, so clearly a different formulation). I thought we had SAN, but after investigating it more thoroughly I realized these are SAN plates, not foam plates. So I guess that's out too.

TLDR, at this moment it seems there's birch ply, okoume ply, Canadian cedar or aspen.

Rumars, again thank you for that big post, that is truly excellent stuff, you did a lot of work for me there and I'm humbled. Regarding the weight, foam and glass option would be really nice, but unless Diab will let me in, I'm afraid there won't be any way for me to get any structural foam. So again, it seems my options are okoume stitch&glue or aspen/cedar strip planking. As per your calculation, if I understand you right, it seems that only okoume has the chance of hitting that 35-40kg weight limit, but not with any kind of wood strip planking.

Of course, again, this means stitch&glue, resulting in hard chines.

However, seeing that I'm driving myself into a corner here and must make compromises, early in this thread you guys suggested that it would be possible to achieve compound curves with plywood (maybe even my design from #1 post) IF I do it over a jig/mould, rather than stitch&glue. I dismissed that idea due to extra materials and labor, but perhaps I shouldn't have. In theory I could build a temporary mould from some cheap OSB boards with lots of stations and a few stringers, try to force plywood into that compound shape, glass the entire outside of the boat, and then release from the mould to do the interior work (frames and the deck/double bottom). But I'm really not sure how much of a mould I'd need (how dense the stations and stringers), and also whether the plywood would actually retain the compound curvature shape after being released from the mould, or if it would spring back to single-curvature shape. I would probably need to leave in some stations as permanent bulkheads, but again, not sure how many would be needed to retain the compound shape in the ply panels.

Also, like you said, maybe it could be done with some darts. As I understand a mould would still be needed. Gary from that trimaran post you linked said he did it with moulds, but unfortunately I couldn't find any photos of that stage of the build. Maybe I'm missed something, it's a big topic... I mean, it can't be just this...



...There ain't no way plywood could be forced to compound shape to these few flimsy stringers.. Yeah, I'm really perplexed on how he pulled it off.

I'm not sure if Gary is still around. He probably could answer on what is exactly needed to get plywood into such complex curvatures and make it stay there. Like moulds or no moulds; how many permanent bulkheads required to keep ply in shape, where to place darts, etc...

The ridiculous part is that the hull that Gary made only weighs 56kg. That's what he said at least. For a freaking 6.5 m. hull. That is insanely lightweight when you consider that length/weight relation is basically cubed. If I can pull off something similar, then it does seem to me that this would yield me a boat that is much lighter than strip planking...

EDIT: just found me mentioned another big thread where allegedly he explained the build process: Alternative to marvelous Buccaneer 24 https://www.boatdesign.net/threads/alternative-to-marvelous-buccaneer-24.32382/. I am going through it, maybe I'll find the answers there.

Thanks for this further info! Vac bagging is no problem, I have materials and equipment for that, and probably enough experience to do a decent bag for a boat of this size. I did consider infusing early in the original design (infuse over the 3D printed core), but had to abandon that idea when I realized that the core can't be properly sealed in any realistic way. Wet bag would be challenging but probably doable with some 10-12 hour pot life epoxy. In any case, I'll hold off this idea until I can find out if I can actually get any kind of decent foam here in Lithuania. I'll know on Monday.

Please check my #1 post in this thread where I posted some pics from CAD with the preliminary design of the internal framing to support the double floor. That was done with the assumption of compound bend ply panels though, without that I'd need more frames or some stringers. I make these assumptions based on the PaperJet design (link). As for deck material, it would have been plywood as well, basically everything made out of plywood except for stringers if any. As for the rudder attachment, I did plan for a doghouse (similar to my previous boat that I linked to in this post).

As for time pressure concerns, you are absolutely right. I am spending day and night on researching all the options you guys have provided me with so I can get designing and building ASAP. Right now, the Gary's method of compound curvature (tortured) plywood, using 4mm Okoume, framing similar to my #1 post and glassing inside and outside, seems the most realistic option to hit my requirements, especially the weight. But I'll continue exploring all the options.

 

In addition to my previous post; I just discovered this article on developable surfaces on this website (feeling kind of stupid I only found it now), which is very educating, I tried it out on CAD and i̶t̶ ̶s̶e̶e̶m̶s̶ ̶t̶o̶ ̶w̶o̶r̶k̶,̶ ̶m̶o̶r̶e̶ ̶o̶r̶ ̶l̶e̶s̶s̶, (EDIT: it doesn't, there seem to be ambiguities and mistakes in that article, and recreating it in CAD leaves curves underdefined), but it still doesn't explain it to me how Gary was able to design such complex shapes while still somehow keeping these surfaces developed. Contrary to what I thought before, according to the article, for a surface to be developable it must have zero compound (Gaussian) curvature. If I understand that right, such a surface would be relatively easy to make from a single flat sheet of plywood with very few bulkheads and stringers, as opposed to a heavy-built mould that forces the ply into an unnatural shape. However, looking at Gary's designs, it's hard to believe that there is no compound curvature in these shapes.

There is also a possibility that I am making a jump in conclusion that naturally achieved shape from ply equals zero compound curve shape. Perhaps that is not a prerequisite, and some compound curvature and the stress it induces into the ply can actually help it stay in some specific desired shape if deformed through strategically selected push points, like the bulkheads/stringers from Gary's design. Or maybe it does have to be zero compound curvature for that to work. Not sure.

I am still perplexed as to how he designed them. His trimaran main hull is indeed very similar to what I'm trying to do here, so if I could only find out his secret, perhaps I could pull it off too. I PM'd him, but his last post was several years ago, I'm not sure if he's still around in this forum...

 

You cannot rest a man’s body on a 4mm okume wing. The flex will be too high and you’ll end up heavy needing more glass. 6mm wood strip is needed.

Before driving yourself into corners; you need to recognize realities. I don’t have the deflection data a composites engineer keeps, but have a good sense that 4mm okume with 6oz glass will bend about 10” under my 100kg weight. The wings will fail. They need to be built like a box. Or at least the seat section needs special consideration.

 

Sorry if I wasn't clear enough, I don't intend the wings to be just a single 4mm surface. Of course that would fail instantly. Like I mentioned before, I do intend to have a box structure there, like in that PaperJet boat. Basically two 4mm ply panels separated by at least 20-30mm to get enough rigidity. Designing that "inner" (topside/cockpit) part is not an issue, right now I'm just trying to figure out how to do the outer (bottom) part with that complex curvature like in Gary's design while keeping surfaces developable.

 

If I remember right FreeShip had valuations for tortured ply on hull designs. Long time when I used it so can't recal precisely how it worked.

 

You are using the correct principle,but i have doubts about the outcome for two reasons;I had a Moth with wooden wings and the depth of the webs was at least 60mm where they left the main hull's topsides and they tapered.Which leads to the second point,the ergonomics of sitting out on a very narrow surface and I believe you would require around a 40mm wide panel at the outboard face.It can be sloped to match the angle one's thigh would assume.

Having reviewed the original proposal for the design,I'm wondering if the cambered surface of the mast support achieves as much as a simpler structure would.Perhaps a simpler,faceted structure would suffice and it might be worth increasing the fore and aft dimension at the sheer.I would be a bit curious about the amount of reserve buoyancy in the bow forward of the mast step and suspect that it would benefit from being increased quite a lot.I would also use a bulkhead similar to the image that I hope to attach to link the highly stressed parts of the rig supporting structure.In essence a 4mm okoume/gaboon bulkhead with 12mmX18mm spruce battens where indicated and I would cover the side sections with another skin of 4mm plywood to make them box beams.I left the area open to illustrate the principle.

The box that supports the rudder fittings would be a simpler job for the builder if it too was created from flat surfaces,at least flat sides and if curvature is desired for aesthetics,the top could be curved in profile and I expect you would incorporate a 100mm inspection hatch in the aft face to allow access to some fastenings for the fittings,its quite normal to do so.

 

Knowing the SOR a little better now, we can try to attack the problem of meeting the weight and toughness goals with available materials.
We abandon the one method fits all principle, because in this case it's counterproductive. For this we (that's you working on the CAD model) slice the boat into sections in order to calculate the surface area. Knowing that we adjust construction method and scantlings.
First slice is longitudinal, at the height of the inner floor. This leaves us with a canoe body, where most of the compound curvature will be. This section we strip plank. The result is a nice underwater shape, as designed. We use 6mm wood strips and adjust the glass later if necessary, we start with 200g each side as the upper limit. If necessary we adjust the wood species down to balsa.
The floor is a separate piece, 3mm okoume ply with stiffeners and a central truss frame, wood and/or XPS.
Second slice is also longitudinal at stem height, below the gunwhale. In the forward part of the boat this gives us the vertical part of the topsides, followed by the transition zone to the concave part of the wing flare. The vertical we plan in 4mm ply with some glass outside and inside to take the unavoidable impacts. The transition zone isn't subject to impacts because it's under the wing, so it gets 3mm ply.
The wings themselves are a box structure made from two 3mm ply sheets with wood and XPS spacers.
The gunwhale we make out of some stronger wood, because that's where the impacts are going to be on the wings.

The transition from strip to ply will be a rolling bevel. We make it a hard chine (several if necessary), with the fillets on the outside of the boat. The joint will be hidden by the other side of the wing structure. To bend this inner part to shape we use the ply with the long grain going transversly to the boat.

There's going to be only three plywood frames, two at the daggerboard ends (skeletonized) and the transom.

Of course I can't guarantee that this is going to work, you need to calculate the individual surface areas, do the math and play around with the numbers. The strip plank part is necessary just a little above the loaded waterline, after that it's a game of making it lighter.

 

Yeah, I used that one before, but as far as I remember it didn't provide any special developability tools except for flattening the panels and informing of the resultant Gaussian curvature stresses, which SolidWorks does too. I watched some tutorials, like this one, to try and get a better understanding of how to design developable surfaces in general, but the issue is that all these articles and tutorials I find only teach very primitive conical shapes, whereas Gary's design proved that it's possible to achieve much more complex shapes that are still (probably) developable, or otherwise naturally fall into required shape with minimum amount of constrains. That is the part I'd like to learn. I wonder if anyone here except Gary knows how to do it? I mean, I would pay good money for someone to teach me this

What you said makes sense, I just didn't quite catch what you meant by that narrow 40mm wide panel at the outboard face of the wing. Did you mean that the Moth with wooden wings had the wing frame naked (no top/bottom ply panels), except for a narrow strip along the outer edge of the wing? In my case, I do plan to completely enclosed the wing box structure (again, like in PaperJet design, as can be seen in this pic), so I think it should be good from ergonomics side.

As for the depth, yeah, I could do 60mm deep wing box, or even more, weight gain is very minimal anyway, but stiffness would improve a lot. I suspect the best way to go about it is to have the box tapered (thickest where it joins with the hull, narrow towards the edge).

That is a great suggestion. I had an idea of doing the outside with strip planking and the inside with plywood, but even doing the outside with a mix of both methods is truly thinking outside the box. I threw together a quick and rough design in CAD to check the numbers. Very rough. Instead of pics I made a gif to make it easier for me to show you and you could tell me whether this is what you had in mind. Just an important distinction: I don't have Okoume at 3mm, the minimum sold here is 4mm. So I used that. The bottom striped part is the strip planking, the light yellow is the plywood. XPS foam is grey.



It's quite ugly, I couldn't get that flare plywood panel in any nicer shape without introducing compound curvature, and even as it is, I suspect it would be quite difficult to bend into such shape without more bulkheads. The more natural shape would be just one continous smooth curve from transom to the bow, like in Moth Magnums, but it looks totally ugly to my eye.

Now for the numbers (this is for both sides of the boat):

1. The underwater part is 3 m^3. With Canadian Cedar at 390 kg/m^3, that is 7kg, plus 200 gsm fiberglass on both sides (double that for epoxy), that's 0.200 g/m^2 x 2 x 2 x 3 m^2 = 2.4 kg, so a total of 7 + 2.4 = 9.4 kg.
2. The total plywood area of the flare, cockpit and deck is 7.6 m^2. With Okoume plywood of 4 mm at 550 kg/m^3, that is 16.8 kg. Plus fiberglass, same calc as before, 0.200 g/m^2 x 2 x 2 x 7.6 m^2 = 6.1 kg, so a total of 16.8 + 6.1 = 22.9 kg.
3. The framing as depicted in the gif above, is 1.6 kg of plywood, assuming no fiberglass.
4. XPS foam to completely fill the volume in between everything is 0.42 m^3, at 50 kg/m^3, that's 21kg of foam.
5. I didn't design daggerboard case, mast step/partner reinforcements and doghouse for the rudder, but let's assume 2kg for that.
6. Probably some 1kg more for epoxy fillets and fiberglass to join the frames and the keel.
7. Probably some 3kg more for fairing and paint, and that's being optimistic.

So in total we're looking at 60.9kg. Without the foam it would be 39.9 kg, but then I'd need substantially more reinforcement for the floor, deck, and especially the wings. Way over the target...

Rumars, can you confirm if this was what you had in mind? I feel like I made a mistake somewhere, the weight is insane, but I can't seem to find any errors in the math so far, and my CAD is telling me the same thing.

 

Yes, that's what I was thinking, now let's make everything as light as we can.
1. Underwater, 3m2: cedar core 7kg, balsa core 2.7kg. 200g fiberglass outside (for abrasion) but 100g carbon inside. Plus epoxy = 400x3 + 200x3 = 1.8kg. So cedar core 7+1.8= 8.8, balsa 2.7+1.8=4.5kg. We can even go to 8mm balsa for better stiffness and the core would be just 3.6kg.
If you can promise not to hit the bottom or drag the boat on shore we can think of 100g carbon outside also, maybe with an additional fiberglass patch in the bow area.
2. Plywood, 7.6sqm. We must divide this into sections, because fiberglass isn't needed everywhere. We put 200g on the outside of the vertical forward part for scrapes. We put 100g carbon on the underside of the wing, because that's where the loading will be. Inside we just paint the plywood with epoxy, no glass at all. Maybe 20kg all in. We can also play a little with the deck and inside part of the wings if necessary.
Even with a cedar core lower hull, it's under 30kg.
3. Foam: we most certainly don't fill the entire volume with it, that would be crazy. To determine how much foam is needed you take a piece of the 4mm ply and cut a square that's just a little bigger then your shoe size (30x30cm or 50x50cm depending on how big your feet are). Then you glue 5cm thick xps strips around the edge, place it on the ground and step on it with one foot. If it doesn't crack you repeat with a bigger piece of ply. If it cracks you add another strip in the center and see if it cracks. Then you sit down on the test piece, but usually people's buts have a bigger area then their feet, so it shouldn't crack from sitting. The determined spacing will be used to make a xps grid between the two layers of plywood. This is just under the cockpit and between the wings, the forward part of the boat remains hollow.

I will write a little about how I think Gary did his boat later tonight.

 

I think the carbon will be in compression under the wing and the wrong choice.

Otherwise, this plan sounds like getting close..

 

My earlier post regarding the comfort of sitting on the wing may not have been particularly clear.I hope the attached rough sketch will be helpful.The transverse wing frame will obviously need a small batten attached to the top edge for the deck to bond to.

 

You are right, there is compression in the concave part where the wing cantilevers out. There is also tension in the flat where the sailor sits, this is part I wanted to reinforce with carbon. Especially since I'm thinking that in the end we will have to tinker with the wing and go from a two ply box beam solution to a single ply with added stiffeners and some padding in order to save weight. Might require carbon on both sides of the plywood.

 

I would just use a different glass on the bottom side. Afa the glassing of the entire box; the inside and top sides in tension can be carbon. I’m not certain about the best choice for the bottom, but something between 6oz woven and 12oz biax ~ it all depends on the engineering and it might be worth some testing