Troubleshooting Glue Joint Performance with Steamed and Bent Wood Glue-Ups
Lots of advice on the fine points of achieving solid glue joints when laminating steamed and bent wood. October 30, 2010
Question
I'm making a steam bent lamination red oak handrail for a custom staircase. Laminations are 5/8" square and 16 feet long. The laminations are bent, clamped up, drying and awaiting glue-up. I took several short test (~8") pieces (not steamed, just milled) and did a test glue-up using DAP plastic resin glue. Did everything by the book (mix ratios, temperature, clamping, set time). Good coat both sides, good squeeze-out under pressure. Laminations have straight, jointed edges. Once cured, I took the test pieces, placed a chisel on one end, and the whole thing fell apart with one gentle chisel tap. The glue failed with no damage to the wood. What happened?
I'm considering using another glue for this project. I have only one chance to glue this up and I'll be on suicide watch if this does not work the first time.
Forum Responses
(Adhesive Forum)
From contributor C:
It might be too wet. If you don't have a pinned moisture meter, look up the microwave drying method on this site and dry out a small sample to bone dry (be sure to follow the directions for a low power setting). You'll need a small digital kitchen scale (about $20) that reads to 1 gram or better, which will get you close enough for your needs. If it's over 15% MC, you can air dry the rail before gluing to the low teens. Increase the heat a bit if you have a ways to go, and it'll dry faster. It'll eventually equilibrate to ~7% indoors, but it should take glue well in the low teens.
From contributor J:
Since you did not steam the test pieces, I doubt if moisture is the issue, and since you followed all the instructions carefully, I suspect the glue itself is no good. Does it have a printed shelf life somewhere on the can? That was a good call to test the glue first.
I hate to recommend Gorilla Glue since I hate that stuff, but it is actually activated by moisture and may be just the stuff on your steamed sticks. I typically wet the wood before I spread the glue. This purple goo will foam all over everything and stick to everything, but it will work on wet and damp stock.
From contributor H:
Forget the resin. Use Titebond II or III and be done with it.
5/8" squares for rail? Steaming? How tight is this radius?
From the original questioner:
I just love yellow glue. Love it. Having said that... One problem is open time. It's going to take me a decent amount of time to get all the sticks glued and clamped up. I'm pessimistic about being fast enough with old yeller before it skins. Polyurethane glue would certainly do the job, but the open time is not very good either.
I looked at the can. Good point - no expiration date, no "born on" date.
From contributor S:
I've used Titebond Extend successfully.
From contributor W:
I have heard from a glue rep that if your planer cut is too smooth, any glue can fail.
From Professor Gene Wengert, forum technical advisor:
Are you trying to glue the bent wood that has not been resurfaced after bending? The surface after bending will not be active for gluing due to the heat. The surface must be refreshed.
From the original questioner:
Interesting responses - thanks. I checked in with the folks at Titebond; they weren't afraid of straight jointed edges; said their glues would perform well with good clamping. I have a little experience with (successful) steam bending and I've never heard of needing to resurface the laminations before glue-up. That doesn't mean it ain't true, of course, and I've never done a handrail before in this manner.
Well, I used Titebond III (open/assembly time is actually comparable to their Extend product) and with a helper we moved like jackrabbits gluing and clamping. The room looked like a glue-fight broke out when it was all done. Thank goodness for sheet plastic. Leaving it clamped up for a few days. If this doesn't work, this will be my final transmission and you'll never hear from me again...
From Professor Gene Wengert, Sawing and Drying Forum technical advisor:
For best results, all wood surfaces should be freshly prepared. This is even more critical with steaming, as the heat destroys the ability of the wood to bond. Further, after steaming, the two pieces will not fit perfectly together; most adhesives can tolerate a gap no bigger than 0.006 inches before they lose strength.
One non-steaming example: A cabinet manufacturer was getting about 20% failure of a glue bond. That is, about 20% of the panels had one or more failures. I suggested that they glue within 15 minutes after the surfaces were ripped, and their failures were 2%.
From contributor S:
Gene, I've only done pre-manufactured handrail glue-ups and nothing steam bent, but I'd like to know what exactly you mean by freshly preparing the surface.
From contributor G:
There was a time when I applied as much pressure as possible, trying for maximum squeeze out, believing that the intimate contact of the wood was the only real factor. Then I read about glue starvation. Have you left enough glue in the joint? The clean break apart that you describe sounds a bit like glue starvation. Try leaving more glue.
The steamed wood with a bit of raised grain, aided possibly by the pores being opened up by the steaming (as well as the surface drying effect when the wood leaves the steamer and the surface temperature flash dries the very outer layer), may retain glue better than your test.
From Professor Gene Wengert, forum technical advisor:
The wood should be between 0.002 to 0.006 inches apart. Too much pressure and you will be too close; too little and you will be too far apart. So, the two pieces being joined must, if curved, have identical profiles and be smooth. This will not happen with bent wood. Hence, you will not be close enough and you will see glue failure with no wood failure. Similarly, the wood surface, after heating, will be inactive (the glue cannot chemically attach to the wood). Again, glue failure. Finally, the surface can be inactive if it was prepared weeks or months ago; again, glue failure and not wood.
If the surfaces are prepared today, but then sit around and change MC, which means they will shrink or swell and not mate well, the same event - glue failure and not wood failure. So, even short storage (hours or days) can lead to poor glue joint strength.
From contributor R:
I've had trouble with starved joints gluing up red oak in the past. I think it's so open grained that the glue just gets sucked up by capillary action. I've seen it ooze out of end grain several inches away from where glue was applied. I've had success with sizing the joints with Titebond first to seal the pores, then block sanding the dried glue surface or lightly re-jointing the edge if need be, then gluing up. White oak soaks up less glue also.
From Professor Gene Wengert, Sawing and Drying Forum technical advisor:
Contributor R, next time try increasing the spread rate.
From contributor R:
Gene, I've laid on as much glue as I could get on some joints and put on several coats, as well as given as much open time as possible, and still ended up with a starved joint. Some pieces of wood are just voracious.
From contributor G:
Contributor R, since you have plenty of glue, would it be reasonable to reduce the clamping pressure? If you have it oozing out the end grain, it would seem that either the pressure was too high or the glue too thin. (That was a question, not a statement.)
From Professor Gene Wengert, Sawing and Drying Forum technical advisor:
I have the same question as contributor G. With as much glue as you are using, the pressure must be excessive to get a starved joint. There are hundreds of people gluing red oak every day without difficulty (that is, without excessive absorption).
From contributor R:
I'm referring to the glue wicking away from the glue surface before it's clamped, so it's not a pressure issue. The glue being too thin, on the other hand, might very well be at issue. Any suggestions on thickening Titebond or using Franklins' cold press formulation?
From Professor Gene Wengert, Sawing and Drying Forum technical advisor:
Try wiping the surface with a wet (not dripping wet, however) rag prior to spreading the adhesive.
From contributor G:
I agree with Gene. If you were trying to laminate fifty year old Brazilian rosewood to titanium composite for an aircraft interior, we might have an adhesive question, but red oak to red oak? How many million glued up panels on their way to making kitchen doors must be made each week across the country? And many, if not most, done with PVA.
The dampening might work, but if done with more than a trifle of water, just thins the glue at the surface and makes it easier to squeeze out of the joint. You get a starved joint without the bother of really cranking the clamps down.
As an experiment, put glue on two blocks and just rub them together lightly - no clamping. Do not disturb for a day or so, then see if the joint fails. The pieces need to be very flat and use enough glue to fill the joint, but not stingy, nor sloppy. If you get a good joint, then you are using too much pressure. The pressure only holds the pieces in alignment.
The last thought is that your glue is bad. Old or frozen. The store might not have rotated its stock, or it was frozen in transit. Buy a new bottle from a different source. Suggest Titebond 2 or 3.
From Professor Gene Wengert, Sawing and Drying Forum technical advisor:
Incidentally, the glue molecule is so large that it cannot easily wick into wood. So even if the wood is exceptionally dry and absorbs the water before you can get the pressure on, the adhesive will still be on the surface unless you have excessive pressure. So I think you must look for another reason why your joints are not doing well.
From Jeff Pitcher, forum technical advisor:
A few thoughts regarding your adhesive choice... First, if the Dap adhesive mixes okay with water, it will work. However, the temperature needs to be at or above 70 F. Much less and you risk dry out. This results when the moisture leaves the glue line, but the temperature is too low to allow the chemical reaction to occur. It looks like the glue has set up properly, but there's no real strength. Check this first. Next, look at the surfaces being glued. As Gene pointed out, if it's not freshly machined, it won't glue well.
From contributor R:
Okay, enough with the red oak. I've found sizing works for me. It also works for edging particleboard, as does using polyurethane. So let me digress and ask Gene about gluing woods that persist in perplexing me - oily tropicals. I usually only use them for trim or accents, so don't have a large surface area for glue. I've tried epoxy, poly, Titebond, wiping with alcohol and not, all with limited success.
From the original questioner:
Okay, back to the oak handrail. I used Titebond III with brilliant success. Sanded it, scraped it, shaped it. Finishing next. Thanks for your help. Maybe the plastic resin would have worked on a macro piece, but my test sticks just weren't giving me a warm and fuzzy.
