Lamination and Springback
A chair back curvature problem leads to a discussion about predicting and minimizing springback in bent wood laminations. December 6, 2009
Question
I'm doing a prototype for a chair using a laminated back. My question is this: my form has a 20" radius but the piece "sprang back" to 22" at the ends. This is where I'll attach support for the arms and approximately 18" from the back of the radius. Would there be any problem bending the piece to the 20" radius even though it's opened up the 2"? The piece is 3/8" thick x 3" high and made up of four pieces about 3/32" thick. Since it's a prototype it's made of three pieces of poplar and one of figured cherry. Since there is little pressure needed to close it up to the original 20" radius I'm wondering if anyone has any experience with this. At this point, I plan to support the end of the laminated piece in an open slot. Thanks in advance.
Forum Responses
(Furniture Making Forum)
From contributor R:
What you describe should be no problem. As long as you can easily bend it to the 20" radius, then you will not be adding undue stress into your chair. Obviously, for later versions, you will want to over-bend your lamination to account for springback. Also, using thinner layers of material will spring back less.
I used to work for a stair company. One of the older men there had been writing notes for over 20 years recording every stringer and handrail he laminated, size, number and thickness of laminations, desired radius, actual radius of the form, and how much springback. He could pretty much predict exactly how any glue up would go and exactly how much spring back there would be. I would love to figure out a way to turn that notebook into a spreadsheet!
From contributor W:
I laminate rocker skids (7 1/4" lams) to a 42" outside radius with less than a 1/16" springback over 48" (cut back to 42"). I agree more lams gives less springback. Being a geophysicist in my previous corporate life, I think it is how many glue surfaces are in the lamination that controls. I have not done any extensive experiments. One question, is your form a solid blank that you pull your lam onto or into, or is it a series of vertical pins? The latter is what I have seen steam bending done on and I use a full form with seven clamps.
From contributor W:
Steaming, if you do that, adds another variable to the mix. It doesn't seem like you need to, but if you take some of the tension out of the wood you will get a truer gluing. My rule of thumb is to bend with steam and get about 50-60% of the curve, then your springback will be minimal. What's even weirder, if you glue the wood up wet (instead of letting it dry a day or two first) as the wood shrinks the curve will tighten. My guess is the shrinkage from drying changes the geometry of the wood while leaving the glue lines steady, tightening the curve.
From contributor I:
Springback can be a problem issue as you’re finding out. Some of my tricks - I start by milling a vertical grain laminate which I use to draw with, if it bends easily for drawing the desired radius then it will work well for pressing (I rip the drawing laminate on my bandsaw then slowly reduce it in thickness with my drum sander until it is the perfect thickness). To meet my desired final stack thickness I reduce from the original thickness of laminates and add laminates.
Also making a sturdy male and female form to press the stack adds a very constant clamping pressure along the entire length of the laminates. I always make my shop drawings large enough to include the clamping/pressing system. Pre-pressing can also help by stretching the material and relieving stress. I'll put the stack in the press and leave for one to a few days (depending on laminate size).
Glue choice is also an important decision. We all develop preferences with experience. West systems and Urea-Bond 800 are usually my first choices.
Finally, if you are the designer, an important element not to be overlooked is material selection and how it’s milled in relation to the pith. All woods will bend, some much easier than others. Test with different scrap. For tricky structural laminating, I've had good results with 1/4rd VG ash. There are many tricks within the few listed here but hopefully this helps. I always keep in mind that designing and building is a dynamic process.
From contributor S:
I have done a limited amount of bent lams, but found the info in the Knowledge Base article below to be accurate.
Calculating Springback
From the original questioner:
Thanks for all of your responses and advice. It looks like I'm going to reduce the radius to 18". I am using a two part form, it's made from 3/4" MDF so what I did was use a trammel point on my plunge router with a 3/8" bit which is the same as the thickness of the piece. Using that bit automatically changed the radius of the male/female parts of the form. I don't have a sander but a shop nearby does so I'll bandsaw/plane the pieces to about 1/8" which is about as thin as I can get and reduce it there until it conforms to the form. I've also thought of spraying the pieces with water and clamping them in the form overnight to "pre-form" them. Anyway, thanks again! Something I did for my mock up of the chair is use 2" Styrofoam. It was easy to cut on the bandsaw and then I sharpened 3/16" dowels to pin it all together. I doubt that's an original idea but it’s much easier to use than wood so I thought I'd pass it along.
