Steam bending tests

[Bill E]

While I've been doing all this bottom & keelson work, I've been busy as well getting ready to replace all those frames. I had posted earlier about having built my steambox and testing it on some kiln-dried maple scraps. All my oak to this point was also kiln-dried, so I milled some pieces to frame dimensions (1/2" x 1-1/8") and put several pieces in a tub of water for about 3 weeks. I finally got a chance to give it a try, so I steamed some for about 40-45 minutes and did a test-bend over the bow roller on my trailer. It did bend impressively, but it also cracked depressingly.

I figured there were three possible primary causes:
1. I need air-dried white oak
2. I need to steam the wood longer
3. I was bending around a single (rounded) point, over-stressing the wood

I decided to eliminate all three possible causes, and I called up the guy in Gainesville, GA who had advertised on Craigslist about having AD white oak. I made a trip up there, and not only did he have a stack of 6/4 white oak, 9' long by 19" wide, he also had a Wood-Mizer bandsaw mill right there. For an extra $50, he took my three boards, clamped them onto the mill, and neatly sliced off a pile of 5/8" rough-sawn frame stock. That was the best $50 I've spent on this whole project. I was really dreading the task of ripping those boards down to size in my basement at home.

I posted that whole story and pictures here.

Anyway, I got the wood home, planed about 20 pieces down to final size, set up the steam box, let 3 or 4 pieces steam for a good hour to hour & a half. Two issues resolved. For the third issue, I decided to create a crude bending form. I took a scrap of 1/2" rough plywood and traced out one half of the transom onto it. I then took a scrap of 2x4, ripped it lengthwise and cut some blocks about 4" long. These formed a series of points that I would bend the wood around. I pre-drilled them and then used 2" deck screws to create the form. Since the rear frames take the tightest bend, I figured I'd go for broke using the outline of the transom for starters. Some of my blocks had to be tapered to get them close enough together.




The first piece bent smoothly - like a charm! Then I suddenly realized I had not left enough overhang on one end where it would meet the keelson. I quickly straightened the wood back out and shoved it back in the steamer. I did snap and/or crack a couple more, but I got another to take the bend pretty well. I let it cool for a while, but when I removed the clamps, it had a surprising amount of springback.



I then adjusted my blocks to over-bend the next piece. I also realized that I had not subtracted out the thickness of the frame itself, so everything was shifted inward another 1/2". After steaming a few more pieces, I tried again.



It's tougher now to get it to take the bend. For those rear-most frames, I'll probably have to pick through my stock for the straightest grain and steam it for a really long time. I've considered planing those pieces slightly thinner as well. Then there's always laminating, if it comes down to that. Overall, I'd label my efforts a cautious success. I did try bending one frame in place, about 2' forward of the transom, and it snapped. I think I can do them in place farther forward, but toward the rear I'll need to use my handy adjustable bending frame. That frame with the springback will be a pretty good fit about 4-5' forward of the transom, so it's not a total waste.

My next task is rigging a way to take-off the shape for each boat frame so I can set my bending frame to match (plus overbend). I guess I can use the existing frame for part of the shape, but since so many are rotten or broken at the tight bend, I'd be guessing about the shape of the overall new piece. Stay tuned... I have some ideas I haven't tested yet!!!

-wte

[Phill Blank]

Bill,

Try using a strip of metal as a backer on the outside of the bend to hlep support the wood while bending. You need it thick enough to support the wood yet thin enough to allow it to follow the curve of the bend. Us clamps to pull the metal band against the strip you are bending and pull both of them against you form.

I talked with a pro boat restorer in the Madison, Wisconsin area sometime ago and he told me that they soak there wood in a heated tank for a week or so before steaming. The have a tall tank with a electric water heater coil in it to keep the water hot and at the temp they want and a float valve to keep the water level where it needs to be for that length of time. Even with all that he mentioned they still break ribs when bending them.

Good luck,

Phill

[Michael J. Seiber]

Bill, That's how I made the new stem for my boat only I used three thinner pieces of green white oak and used the bow of the boat to form them with. I screwed them to the bow with spacers in between. Left them that way with a small fan blowing on them for three weeks. When I laminated them together I used Gorilla glue and a jig like you used to form the ribs. It worked great and only sprung back a little bit.

[JoeCB]

My $0.02 worth, any mind you, I'm no bending expert but I have picked up on a couple of helpful tid bits.
1. it's the heat more than the water that facilitates a successful bend. yes you do need both but I think the moisture helps most by conducting the heat into the core of the wood.
2. yes the metal strap helps a lot. But understand what it's really contributing. when you bend anything with any significant thickness the outer fibers are put in tension, the inner fibers are in compression... somewhere in between there is no stress at all, this is called the neutral axis. Wood fibers fail easiest when in tension... pretty hard to cause a failure in compression. When you add a securely attached metal strap on the outside of the bend you shift the neutral axis out toward the strap, hence less wood in the zone of tension and more in compression... no cracking, a good thing!
3. keeping the neutral axis concept in mind. be mindful of the cross section of the piece of wood you intend to bend. The best example of an emminent failure would be a outer stem that one has already shaped into the trapazoid shape before bending... the neutral axis is way down near the wide edge and the bulk of the wood is in the tension zone... breakage for sure.

Joe

[Bill E]

Phill - That sounds like a good idea, because the cracking always occurs on the outside face, not on the inside. One material that comes to mind is pallet strap bands. Being in construction, I come across these all the time. The next thing that occurred to me is that a lot of lumber nowadays is bundled using nylon webbing/strapping. Some of it is 3/4" - 1" wide, but for major loads they use ~1.5" material. Seems like a length of that with a couple of D-rings as a ratchet might be worth a shot. I do have one of those big 2" ratchet tie-down straps, but that's probably overkill.

Michael, thanks for the input, too. Even if your pieces were still a little moist after 3 weeks, the G. Glue loved it, as it cures in combination with water.

Joe -- I appreciate the perspective you bring. I think Danenberg talks about the moisture being more of a vehicle for heat transfer than a softening agent in itself. He also says that most of the bending occurs through compression of the cells on the inside of the curve, with very little elongation of the outside cells occurring. I don't have the percentages handy, but I think you get maybe 10% max elongation, and a lot higher % compression.

Thanks especially for the observation on the outer stem bending. I'll be replacing mine, and I had thought I might pre-bevel the edges, but now I won't!

Great input, everyone! Keep it coming! This is what makes this forum so valuable, and not just for me. Many others (Peter S.) will be walking this same path before too long, and the more collective information we have readily available out there, the easier that walk will be. Heck, I may just have so much fun that I'll walk it again someday!

-wte