preventing warpage?

[Rema1000]

I have a lightweight stich-and-glue (plywood) rowboat. I thought I'd make some nice thwarts using 1/2-inch poplar.

To make a curved transom thwart, I first joined 5 boards together using tongue/groove (as if for a tabletop), then cut the needed shape, and fiberglassed the underside to reinforce. I left the thwart in my (hot, sometimes humid) shop to cure for a few days; unfortunately, the thwart warped in the heat, such that there is a rise from the center to the sides of about two inches. I'm not sure if having one side of the thwart 'glassed may have contributed to this.

I now have it clamped under modest pressure, and am wondering if I should humidify the shop to help it flatten-out. But especially, I'm wondering if this will be a problem once it is on the boat. If I get it flat enough, and seal it with epoxy, will that prevent it from warping once on the boat?

It will be fastened to the monocoque hull on three sides with silica-thickened epoxy; and I had planned to add one (or two) cross braces underneath, and fiberglass underneath, for added stiffness (since the thwart is only 1/2-inch thick). But I'm not confident that the joint with the cross-braces and hull would keep it flat, if it is determined to warp!

[Bob Smalser]

I'm not sure if having one side of the thwart 'glassed may have contributed to this.
With rare exceptions, whatever you do to one side, you have to do equally on the other side....especially with flatsawn boards. The raw side simply took up moisture at a faster rate that the glassed side and expanded.

You can try drying out the raw side...clamping won't do much and more humidity will make it worse...but your chances of success at this point are small, as the wood has taken a set and there's no way to get moisture under the glass side alone....I'd just go ahead and make another thwart.

Make sure you alternate the cups of the boards on the next one and either glass both sides or glass none.


...fastened to the monocoque hull on three sides with ... epoxy; and I had planned to add one (or two) cross braces underneath, and fiberglass underneath, for added stiffness (since the thwart is only 1/2-inch thick). But I'm not confident that the joint with the cross-braces and hull would keep it flat, if it is determined to warp! It'll also split for sure next dry season if epoxied on three sides.

Either use plywood or screw it down to seat risers conventionally. You don't need crossbracing....laid up 2/4" is sufficiently strong depending on the depth of the seat and lack of support isn't why it is warping.

For more support, cleat the underside and rest the cleat on a post. Screw it from the underside through elongated pilot holes in the cleat so the solid thwart can move side to side seasonally without binding.

[ 08-20-2004, 01:57 AM: Message edited by: Bob Smalser ]

[Dave Carnell]

I have had success flattening warped, wide pieces by wetting them well with ethylene glycol auto antifreeze and clamping or weighting them. In interior service, they stayed flat. The epoxy-fiberglass may complicate things.

[Bruce Hooke]

Yes, the fiberglass on the underside is what almost certainly caused the problem but I really don't think you need to start over. Here is what I would do:

If I am reading your question correctly it sounds like the seat is cupped (i.e., the sides are higher than the center). If that is correct then I think Bob may have misread your question on that point. If it is cupped then laying a damp cloth on the top for a few hours or days should flatten it out (a less radical approach would be to move the whole thing to a cool, damp location for a few days). If it is not cupped but is instead bowed (i.e., the center is higher than the sides) then putting it out in the sun for a few hours should flatten it right out. It may take some juggling to get it flat, especially because you want it to also be dry enough to epoxy, so if you put a damp cloth on it you may need to overdo it a little so that the surface can dry out enough to allow you to epoxy it. While flattening it watch things closely -- it's surprising how fast in can change.

Once you have it flat and dry enough to epoxy then get two or three coats of epoxy onto the top surface ASAP (and then ends if they are not coated already). Ideally you should also put fiberglass on the top just like on the bottom (Bob is right that doing the same thing to both sides of a board is always a good idea), but if you really don't want to fiberglass the top I think you can get away with not doing so as long as you get three good coats of epoxy on it. As long as both sides of the board have a coating that is equally resistant to water movement in and out of the board, there is no reason why it should warp once it's in the boat.

Once you've got all sides epoxied I don't think it will split in the dry season, but it is not completely out of the question. For a wide panel like this the ideal route is either plywood or boards layed with spaces between them so that each board can move around on it's own. Since you are glassing and epoxying everything I probably would have gone with plywood. If you can store the boat in an unheated space then it will be much less likely to split than if it is in, say, a heated shop in the winter. The epoxy will slow down moisture movement in and out of the boards but it will not completely stop it.

As to whether you need cleats on the underside -- this depends on the width of the seat. 1/2" is not that thick so if it is fairly wide then some sort of support underneath may well be necessary. You should be able to judge this fairly well right now based on how stiff the board is. Unfortunately, adding more glass on the underside will not help with keeping the top and bottom in balance with each other. So, if one or more cleats does seem necessary then I would try to put them on with a minimum of extra glass. You could just coat the cleats in epoxy and then glue them in place rather than attaching them with glass cloth. The other risk of using cleats is that they really stop the boards from shrinking and expanding so if the moisture content of the boards does change much from winter to summer then the boards may well split. The sides of the boat will act to some degree in the same way but there may be a little more give there. However, there are two factors that should cut down on the risk of splitting. First off, with everything coated in epoxy water will be pretty slow to move in and out of the boards. Second, if the boat stays in an unheated space in the winter then the change in humidity levels that the boards will be exposed to will be MUCH less than it would be with a piece of furniture that is in a heated space. That said, the vertical post idea that Bob suggested would be a safer way to go when it comes to not constraining the boards from shrinking and expanding.

All of that said, the safest route would be to start over and use plywood, but I'm not at all convinced that the seat you have made cannot be made to work.

[Rema1000]

Originally posted by Bob Smalser:
</font><blockquote>quote:</font><hr /> ...fastened to the monocoque hull on three sides with ... epoxy;It'll also split for sure next dry season if epoxied on three sides.
</font>[/QUOTE]Sorry, I meant that I'd glue it to the hull along three edges (transom, port, starboard). It would be epoxied on all surfaces. Or maybe it would still split.

I think I misrepresented one thing: the thwart is glassed underneath, but the edges now rise upwards. So the unglassed side is concave. It sounds like this means that the unglassed side dried-out; perhaps letting it sit for a week at moderate temperatures will correct the warpage, if the exposed side can recapture some moisture.


You don't need crossbracing....laid up 2/4" is sufficiently strongI had a problem with one of the bench thwarts, where sitting on it caused some deflection, which pried the top edge of the endgrain of the thwart away from the hull on one side, cracking the epoxy fillet which supports the end of the thwart and fastens it to the hull (there are no risers, no keel, etc.; just 1/4" plywood panels for the hull). It won't break; but I need it to be stiff enough for a 3-foot-wide span not to deflect when you sit on it.


For more support, cleat the underside and rest the cleat on a post. Screw it from the underside through elongated pilot holes in the cleat so the solid thwart can move side to side seasonally without binding.Hm, no posts in this hull (actually, no screws either); but I get the idea. It's sounding like if I would do best to pitch the current thwarts and either use marine ply and paint it :( or else laminate some veneer on top. I was using sawn boards to match the wood in the breasthook, knees and rubrails, but obviously this isn't such a good idea.

If I were really set on using boards, it sounds like I would do best to start over, and to alternate the cup of the boards, and to glass the underside, quickly screw each board in the thwart down to a cleat, then glass over the top, and give the whole thing a generous fill-coat or three of epoxy.

Oh, well, the next boat will be perfect, right?
;)

[Bruce Hooke]

Yes, I think the unglassed side dried out. So, getting some moisture into that side should cause it to expand again which will flatten it out.

1/2" thick solid wood is nowhere near thick enough to span that sort of distance without diflection. So, you will either need to use much thicker wood (say 3/4" min.) or use one or more cleats. I assume the grain runs in the long direction, right? (Meaning from one side of the boat to the other in your case, I think.) If the grain is fore-and-aft and the seat spans 3 feet then you would need absurdly thick pieces of wood to span without a cleat. So, all in all, a cleat is probably a good idea, especially if you want to use the thwart you've made. I still think that if you get it flat and epoxy all sides and attach a couple cleats to the bottom that you should be OK.

If you are not inclined towards that route here are some other options:

1. Switch to plywood and either paint the plywood or veneer the top. Remember, even with plywood, if you glass one side you should glass the other side. Similarly, if you veneer one side you should veneer the other side (you can use a cheaper veneer on the underside), but you can sometimes get away with not veneering the bottom, especially if the plywood is fastened to a solid support system.

2. Put a cleat all the way across the space a little aft of where the forward edge of the thwart will be (probably best to orient the cleat on edge rather than flat to make it stronger). Fasten another cleat to the transom. Run individual boards fore-and-aft spanning these two cleats and attach the boards to the cleat with screws (there is no reason why you cannot use a few screws on a boat that is mostly held together with epoxy). This is basically the traditional approach to the issue. If you want to you could cut these boards to fit and then coat them with epoxy before fastening them in place. Note that in this approach you are leaving spaces between the boards so that each board can move on it's own. Fastening down the boards and then coating them with epoxy is about the same as gluing them together first except that without glue between the boards they will be even more likely to split along the joints between the boards, thus splitting the epoxy coating and making for a really messy situation.

3. Find a way to thin down your current thwart to say 3/16" and then glue the whole thing down to a piece of 1/2" plywood (or use thinner plywood and put a cleat underneath) and coat everything in epoxy. This is basically the same as veneering plywood. I would not go any thicker than 3/16" for the solid wood and thinner would be better. Do the plans call for plywood? If so then you can use whatever thickness plywood the plans call for and just treat the solid wood on top as decoration. There is still a chance of some warping issues with this route because the top will be veneered in this case and the bottom won't, however, I think a good cleat fastened to the bottom would be enough to keep things flat.

The basic principles are:

1. Solid wood will always move with moisture changes. You cannot stop this from happening, but you can greatly reduce the degree to which it happens by coating everything in epoxy and thus slowing down the movement of water in and out of the wood and by reducing the moisture swings to which the boat is exposed. The exception to this rule is veneer where the wood is very thin and is attached to a substrate the does not move.

2. If the top and bottom of a piece of solid wood are treated in different ways then when the moisture level changes the board will warp.

[Rema1000]

I assume the grain runs in the long direction, right?Actually, the grain of the transom thwart runs fore-aft. I did this because most of the weight will be at the forward corners of the seat, and didn't want these "fingers" of the seat to break-off (picture below):

http://www.specialgreen.com/~bcarlson/boats/transom-thwart.jpg

(inside of the boat is unfinished, just glassed).

This should lend well to screwing it down to a pair of cleats underneath, as you said, one behind the seat edge and one at the transom. Unfortunately, as you can see, the foreward cleat will be a bit thin towards the edges, where it is needed most.

I think I will have a go at letting the existing thwart straighten, and see how well it recovers.

This is my third plywood stitch-and-glue, but my first use of sawn wood other than sheer clamps or rails, so I guess that I made some interesting decisions along the way.

[Bruce Hooke]

Thanks for the picture. You make a good point that those forward corners would be pretty weak with the grain running the other way. Also, it would look very UNtraditional with the grain running the other way. However, the trade off is that the overall seat will not support anything much at all as it is -- if you sat on it I think there is a very good chance that it would split. The glass on the underside is the key thing that would be acting to prevent that from happening, but that is not enough. So, a couple of cleats across the bottom would be a very good thing in terms of strength. Actually, the key place to cleat is out near the forward edge, the back edge will be well supported by the transom if you attach it well -- a cleat there would mostly be to gain a better connection to the transom. Putting the forward cleat on edge would be very good because it will be much stiffer that way. I'm not sure what you mean when you say that the cleat will get thin at the edges. Are you simply talking about the fact that the curve of the hull will cut into the cleat? If so, I wouldn't worry about that because as long as the bottom face of the cleat rests on and is glued to the hull the load will simply go to the hull, which is what you want. For a cleat I would probably use a piece of wood that is 3/4" thick and maybe 2" high. That is assuming the cleat is on edge. If it is flat (and 3/4" thick) it would need to be 14" wide to be as stiff as a 3/4" x 2" piece on edge!

There will always be some chance of this seat splitting from getting to dry, especially in dry winter air, but if it does then you can deal with it at that time...

Hopefully the wood was pretty dry to start with. If it was on the damp side then you've got more trouble ahead because even if coated with epoxy it will eventually dry out more and shrink...

Another option would be to cut the seat into three pieces with fore-and-aft cuts and then attach each piece to the 2 cleats seperately (leaving the width of the sawcut between each cleat and it's neighbor) and preferably just with screws to allow for a little movement.

[Rema1000]

Yep, I had always intended to support it with two cleats, on edge; although I had been thinking of using the same 1/2-inch stock, in 2-inch-high cleats. I'll have to go hop up and down on one and see how much weight it'll take. I should mention that nobody will be doing jumping jacks on this boat, as the whole boat weighs only about 80 lbs now(!). When you step into it from a dock, it feels quite squirrely in the water, like stepping into a canoe. But maybe it will inspire more abuse once it's weighted-down with gear.

As far as the cleat being "thin", yes, I meant because the hull curves in so far. There should still be plenty of support, but I was thinking that if the cleat is only (say) 3/4-inch tall there, this may make it harder to screw the seat down to the cleat out at the edges to prevent warpage.

Funny you should ask about the wood: I made-up the "tabletop" 2.5 years ago, and just now got around to cutting it up. So the wood has been curing as a piece (unheated) through 2 Minnesota winters. Actually, that makes a case for not starting over, since this wood has been dried a bit.

[ 08-20-2004, 04:16 PM: Message edited by: Rema1000 ]

[Bruce Hooke]

I'd think that 1/2" x 2" cleats would do the job. As you say, it's a lightly built boat. Assuming the wood has nice straight grain and no knots the breaking strength of a 1/2" x 2" x 36" piece of poplar on edge should be near 450 pounds. Diflection may be the bigger issue, but I don't think it will diflect enough to matter.

Realistically the cleat will not do that much to keep the boards flat so I would not worry too much about how much purchase you have for screws at the ends. Keeping the boards flat should mostly be addressed by getting the boards flat now and then coating them with epoxy so that the moisture level of the wood changes evenly from both sides when it does change. In fact, unless you are going to cut the seat up into three pieces (the safest route short of starting over) I would just glue the thwart down to the cleats. With the cleat in there and the thwart glued to the sides of the boat it will not be able to move around anyway so screws will not offer any advantage in terms of flexibility anyway. Of course glueing it down will make it harder to pull it up if it does decide to crack...

Good luck with it...

[Bruce Hooke]

P.S. I just looked again at my calculations and realized that I had assumed a uniformly distributed load on the cleat. A point load is probably more realistic since the maximum load is likely to be someone stepping on the edge of the seat. In that case the breaking load for a 1/2" x 2" piece is 230 pounds -- this is probably more than it's likely to see but not by a whole lot. Going to a 3/4" piece of wood or glueing two 1/2" pieces together to make it 1" would be safer. Of course the thwart itself reinforces the cleat so it would take more than 230 to break it even if you go with just a 1/2" piece...