My next project is going to involve a couple of bent laminations, with quite a few layers (16). Each piece is going to be about 2.25″ wide, and 78″ long, and the will be bent into an approximately 5′ radius. I’m using padauk, and my plies will probably be about 1/8″, and I need a finished thickness of 2″.
After digging through all my old FW magazines, I see that polyurethane glue is recommended for bent laminations, so I picked some up. Now the articles also mentioned that you don’t need as much polyurethane as you do your standard yellow PVA. So, my question, do I glue both sides of each ply, or just one side?
Thanks!
John Friesen
Replies
Hey, John—
One side should do it. You have plenty of opportunity for the glue to transfer to the other piece, thanks to clamping pressure. Try using a 3-in. grey or yellow foam roller to apply an even coat. The only snag with using polyurethane is that you’ll have to dampen each face of each ply, since the glue needs extra moisture to activate. That could become quite a time-wasting chore in such a big glue-up as yours.
Yellow or white PVA glues work OK for laminations, but two things they don’t handle particularly well is springback and creep. If there’s enough tension on the laminates due to the curvature, you’ll get more springback using a softer PVA. And your gluelines may exhibit protrusions over time, which is the glue moving beyond the surface of the wood, a term some call creep. For many lamination applications these aren’t serious drawbacks. But for flat surfaces such as broad tabletops, where the surface really shows, or wherever a hand, leg or arm caress the wood (think: chair arms) you’ll see or feel the raised glue, and it will disappoint.
Polyurethane cures to a much harder consistency than white or yellow glues, but still I’ve never used it for laminations. I find it fussy to work with, and the fact that it expands during curing is an annoyance—and sometimes a disaster if you don’t plan for it.
My favorite lamination glue, especially for severely curved work, is urea formaldehyde, otherwise known as plastic resin. I use the powder/resin variety (one brand is Unibond 800) instead of the hardware-store powder/water mix, as it doesn’t introduce moisture into the wood. Plastic resin glue is easy to work with, and dries super hard, which lessens springback and doesn’t permit the adhesive to creep. You might want to try it on your next curvaceous project.
Good luck,
Andy
Thanks for the tips Andy. I'll try urea formaldehyde on my next project, but I'm going to have to stick to polyurethane on this one, as I'm doing the project next week (I've booked a weeks' holidays to be able to get some serious shop time), and I haven't seen any of my local suppliers carrying urea formaldehyde. I've looked, since it was another alternative I was considering. I'll have to order some to have on hand for next time.
Thanks again!
John
Hi, John,
I have used polyurethane glue for bent laminations and think it works pretty well. It is messy, but that can be handled. Use a release layer like wax paper or just liberally wax every potential contact surface. I applied the glue to one side and wiped the mating surface with a damp sponge to provide some moisture. I like Andy's suggestion of a roller. I used a putty knife to apply the glue and then removed about half of it after the entire surface was wet with glue. That helped minimize the foaming.
You might consider using somewhat thicker and fewer plies. At a 5' radius, you could probably get by using 10 or 12 plies of appropriate thickness and still have a reasonably small springback. Springback is inversely proportional to the square of the number of laminants, so 10 laminants is a good number for small percentage springback.
Plastic resin glue is also a good choice in my opinion. However, it is temperature sensitive, and at 70 degrees F, it takes about 12 hours to set. At 90 degrees F, it takes about 3 hours. My shop is in a basement at 65 to 70 degrees F, and that is an issue with me.
Good luck, Tom.
Tom— That's good advice about the poly glue, and a good perspective from someone who's used it. I want to clear up a couple misconceptions, though.One thing to watch for with laminate bending is the thickness of the laminates. I'm not sure what you're implying about using less laminates, with each laminate being thicker, but in my experience, using more laminates means having less springback, and using thinner laminates also means less springback. I hope this was your intention; I just wanted to clarify. John's nominal 1/8-in. thickness sounds perfect for the job, assuming springback is an issue for him.Also, a word about the resin/powder form of plastic resin, such as Unibond 800. It has a comparatively small clamping time compared to conventional (powder/water) plastic resin. Depending on your mix ratio (which can be adjusted to suit), at 75 degrees F, there's an incredibly long (in woodworking terms) open time of 20 minutes— the time you have to spread the glue before applying clamping pressure. Clamping time is about three hours, not the agonizingly long 12 hours for the conventional powder/water adhesive. In your basement at 65 degrees F, clamping time would be five hours. At 95 degrees F, clamp time drops to one hour. In fact, when I use the glue in the vacuum press for flatwork (non-stressed work) during warmer weather, I usually release the bag in about an hour.These attributes, along with water clean-up, make Unibond 800 very friendly to use.—Andy
Andy,
You are exactly right about Unibond 800. It's properties are quite a bit better than the plastic resin powder glue. I have used both and always put the unused glue near the glue up to tell me when the glue is set. In my basement shop, I put a box of styrofoam around my vac press and heat the inside with an electric space heater. The temp reaches 95 to 100 degrees f, and I can get very fast cycle times.
On the subject of springback: Many years ago, a reader derived the theoretical relationship between springback and design factors like number and thickness of laminates, and material properties. He found that springback is inversely proportional to the square of the number of laminates. Laminate thickness dropped out and was not a factor. So, for a constant overall thickness, 20 laminates of .10" thickness will have one fourth of the springback after the glue sets of 10 laminates of .20" thickness. Both laminations will end up 2" thick. My point on the question is that 10 laminates of 1/4" thickness will also have small springback, even though it will be larger than if he used 20. Is the extra work of creating twice as many laminates worth the effort? Only experience and an understanding of the design requirements can answer that.
Thanks for the chance to exchange thoughts on this subject.
Best regards, Tom.
This forum post is now archived. Commenting has been disabled