Adhesive Choice for Bent Laminations
Lessons learned from trying various glues to lay up beefy members. June 26, 2009
Question
I ran out of the West Epoxy I was using to make some 2" x 10.5" cherry stair stringers built of 7-1.5" laminations (FAS grade). Turns out it will be about ten days until I can get some more. The stringers are not heavily loaded since their size is more so based upon looks. I didn’t want to use PVA to avoid long term creep, but now I am re-thinking this. Does anyone have any suggestions?
Forum Responses
(Architectural Woodworking Forum)
From contributor J:
Why epoxy in the first place? Have you considered yellow glue - the regular yellow stuff or maybe the super yellow stuff? If you just can't bring yourself to trust "old yeller" then you should probably go for the purple-grape Resorcinol. I think it's the same glue in those glue-lams holding up the roof of your building.
From the original questioner:
I suppose it is overkill. I was aiming for a glue line that was a bit more rigid than yellow glue. I worked with a PHD fellow years ago that did a lot of research reinforcing glue lams and he started out with yellow type glues. It turned out to be too elastic. We reinforced a bunch of under-designed school glue lam beams using an epoxy he recommended and I had some kicking around.
From contributor Z:
This is a subject I always find interesting. If what you’re gluing will fail before the glue then what’s the point of the "super glue"? I have been gluing butcherblock type stringers for many years and always used Titebond 1 or 2. A glue lam beam cannot be compared to a stair stringer. The forces applied to each are different and in different directions. Use the yellow glue. Creep? It will be for all intents and purposes un-measurable.
From contributor A:
I would likewise use yellow glue (TiteBond 1).
From the original questioner:
Out of curiosity I calculated the max longitudinal shear stress. Its only 45psi, so PVA will be absolutely fine as suggested by all.
From contributor R:
Where are you in Canada? I'm in Vancouver and there are many sources - any marine supplier, Lee Valley, I think even Windsor Plywood carries it.
From the original questioner:
The Pro-set 275/175 in cartridges. I called direct to the Canadian distributor (Paynes in Victoria) and they passed me on to a company in Edmonton. Not a big deal, I will be using PVA, but thanks for reminding me Lee Valley carries the other West Epoxy brands. The supplier did tell me about a new product that West came out with that they are quite excited about - the 610 epoxy. It is a two part epoxy that comes in a caulking tube. It mixes in the tube before application. Sounds quite reasonably priced for an epoxy.
From contributor N:
I have been using Titebond III, even with projects in my vacuum press the Titebond III works great and I have not had a failure since using the glue when it first came out. I guess in 20 years I'll know more because the III is not been around very long, but I bet the wood fails before the Titebond III. I love the color it dries to matches loads of lumber.
From contributor L:
Quite a while back I did a series of glue tests on bent laminations. The wood used was hard maple (one of the more difficult from both the glue adhesion and bending stiffness). 10 layers, 1/8" at about 5' radius - after the glue had cured for 24 hours on the form the parts were laid aside for at least a week. Then put against the original form. Each part was then machined smooth and mortised; screws were inserted into the edge of the laminations using a pilot hole that was a snug fit to the screw.
Resorcinol: no creep, no delamination from screws, no glue line failure from flexing the part. Urea formaldehyde: no creep, screws caused some failure in glue lines, flexing the part caused some delamination. TiteBond 1: slight creep, no problem with screws, no problem with flexing delamination, West system epoxy: more creep than yellow glue, no screw or flex delamination, White glue: significant creep, no other problems.
Has anyone else run similar tests? Since I only made one sample for each glue, there could have been variables that are not accounted for in the results. A different wood may have resulted in better results, especially for the UF.
From contributor J:
With respect for your efforts, I don't think 1/8 inch stock, bent around a 5' radius is much of a test. Try asking the guys who use bender-rail about glue-line creep, spring-back, unwind and delamination. I was once boomeranged off the second floor while singlehandedly trying to bend a bundle. I might as well have been trying to bend the bow of Ulysses. I hit the floor just slightly ahead of the entire contents of my nail bags and (long before my hardhat). In this case the Bender failed before the glue even had a chance.
From contributor R:
Fine Woodworking #192 August 2007. They tested Type1 PVA (Titebond III), PVA (Elmers Carpenter Glue), slow set epoxy, liquid hide glue, hot hide glue, and polyurethane (Gorilla). Total of 162 samples - three samples of each of maple, oak and ipe, each with a tight, snug and loose bridle joint. There were some variations due to wood type and joint fit, but overall winner was Titebond III. Epoxy came in at 99% of Titebond, regular PVA at 95%. The surprising and controversial results were for polyurethane glue - bottom of the pack at 58%. There's quite a bit of data in the results; if you're interested should try to get your hands on a copy of the article.
From contributor L:
The test was the only one that I ever did as a direct comparison between glues; the curve was what was going to be fabricated after the testing. We currently bend to 12" radius and run the results through an arch shaper with no problem. One oddity that we've come across is that if you make the laminations very thin for the radius and use Titebond the curve will actually tighten after removing from the form rather than what you might expect in spring back.
From contributor J:
Contributor L - that is interesting, Do you have a theory or any idea why (the spring-in)? I've never heard of that before.
Contributor R - I will check that out, that sounds pretty definitive. One thing I've learned about glue, adhesion is one thing and flexibility is another. A rubbery type glue may actually have more sticking power under stress but only because it's more flexible when cured. This flexibility can contribute to more glue-line creep while still staying stuck.
From contributor L:
Don't know why the spring-in but suspect it has something to do with the amount of water absorbed (from the glue) into the thin laminations.
From the original questioner:
I did fess up and admit that the epoxy was overkill for this project. The stringers are all done and final sized. I would say that the epoxy was a bit better for gapping any stock imperfection.
