Bending Douglas Fir Dry

After asking for advice, a woodworker successfully laminates a curved 5-inch by 7-inch beam on a tight radius by laying up and bending 3/4-inch-thick plies. October 17, 2012

I need to bend and laminate some douglas fir boards for a current project. Sadly, details are my business so I seek insights on the below areas. All the information that I have read says I'm on the right track, but I only have one shot and need to do all I can to ensure success.

Here are the particulars:

Chord length: 65".
Rise height 6".
Air dry boards.
Goal is to bend a 4/4 plank.
7 plys to yield a 7" wide beam.
Boards are 5.5" wide (beam to be 5.5x7 when done).
Glue will be West System 207 epoxy.

Question 1: The boards are all quarter sawn. Is this trouble? I think flat sawn is better but I don't really know.

Question 2: Should I have a form to support them? Will a slight dampening help the wood, but hurt the epoxy? Regarding question one - I can make flat sawn stock, but prefer the quarter due to sawing conditions (we are sawing beams up for this).

Question 3: Could we go thicker to 6/4 plys?

Forum Responses
(WOODnetWORK Forum)
From contributor S:
Is it a support beam? If it is, I am not sure you can build it without an engineer’s stamp. I designed and built my house and there are a lot of exposed fir beams in the main living area. When I submitted the drawings to the building permit people the only thing they wanted was an engineer’s approval for those beams.

From the original questioner:
Perhaps I was unclear on question two. Would a solid support structure or form mold so as to guide and support the fibers during bending be a requirement?

From contributor D:
Why the 4/4 and 7 ply deal? The radius is 91.021", and length is 66.467 at that radius. I would make the pies 1/4" (22 plies) or maybe .3055 (18 plies). If you need a beam 5-1/2 x 7, then make enough plies to get to 5-1/2", and make them 7-1/4" wide. Then make a form for the inside edge of the beam, clamp it up. Plane it to 7" when cured. Water is not needed, but an engineer's stamp will be if it is structural.

From contributor B:
I doubt you can bend a 4/4 plank of doug fir on a 91'' radius (especially short pieces). A stack of seven pieces would really require a big gorilla (steam and a heated hydraulic iron press). I also question the use of epoxy glue. Good laminating requires extreme clamping and that will positively squeeze out all your epoxy. Commercial glue-lams are bonded with Resorcinal.

Reduce the thickness of your stock until it becomes a manageable cold bend. No on the 1/4 sawn boards. They won't bend (or resist bending) along their vertical grain. Yes, either way you will require a very substantial bending form and clamping method. Why not glue-up your solid beams to form your radius pieces? You can simply band-saw them out.

From contributor H:
Contributor B has pretty much covered the key points here. If I understand correctly you would like to laminate seven 1" x 5 1/2" boards into this 5 1/2" x 7" beam. This is a massive bend that I believe would require something on the order of hydraulics to pull off.

You stated that you only had one shot at getting this right. That actually isn't correct. I would never attempt an extreme bend without doing a dry clamp up first. I believe that when you try to do this without glue you will discover that it isn't going to happen. At best I think you might pull it off with 1/2" thick material, but even this would be a massive undertaking. Using 1/4" thick material would do the job though. And, with that thickness you could use West Systems, although there is certainly nothing wrong with going to Resorcinal.

There is one other issue that has not been brought up by anyone as yet. That is the uniformity of the bend. If you take a 4' long 1/4" dowel and hold it by the ends in front of you, you'll discover that when you bend it, it bends nicely in the center but remains straight at the ends. That is because there is no wood pulling it into a curve beyond your hands. The ends of a stick will always be straight when you bend without a form. So, in answer to your question: yes, you will need a form to make the bend successfully.

From contributor Z:
Another point on quarter sawn - quarter sawn defects ruin wood for bending. Picture a knot in the quartersawn direction and then think about bending that. It looks nice in rustic flooring, but will crack when bent. Also, if you manage to bend 4/4 DFir at all, it'll be under so much stress that the beam will want to relax its shape if it can. Water and steam won’t help doug fir but will spoil the glue joint to some degree. Consider hardwoods if you want to bend something thick, like ash, oak, and elm.

From contributor E:
Being a one shot deal, I'd use 1/8" thick plys referred above. We do this every day for arched casing with no problems. 4/4 could be bent successfully by steam bending, as traditional wooden boat builders do this all the time, but it is a process that requires lots of experience to perform properly. Quartersawn is the best to use in doug fir as it won't separate as badly as flat sawn when bent. With such a flat radius that you're making though, I'm sure it wouldn't matter either way. I've gotten 1/8" vertical grain doug fir veneers to bend all the way down to a 12" radius without cracking.

From the original questioner:
I've completed my task with only minor difficulty. So I'll try to respond the all the insights. Yes we did bend quartersawn doug fir straight grained, with few knots, and cold without steam. The boat building community seems to think that grain orientation is not such a big deal - just grain run-out. Reading up on the subject I found the fiber should be able to handle the radius, but to be conservative we thinned the board down. Laminations were 3/4" thick, right off the band saw. No planning - 11 plys in the beam. The glue line is about .020 in thickness - about the same as a commercial Glue-lam scrap we had on hand.

Yes it was a 5.5" x 7" beam. Planks were 7' long to yield a 5.5' curve. Yes we used a form to ensure even fiber stress on the convex side.

We did not do a dry bend - slipping the surfaces under such friction without lube (glue) would be a risk. Resorcinol is a tough glue to use because it requires 125 PSI of clamping pressure to make for the best results. My hand clamps cannot produce this with predictable results so epoxy was the adhesive of choice. Yes it took 1.5 tons of force (best guess) to bend the 11 plys in the stack, which a car jack did a fine job with. I recommend this as the control was excellent.