Reality-Check Requested: Glue Clamping Pressure Gone Wild???

[Jim Neeley]

Please explain to me what I'm missing on determining the number of clamps based on the clamping pressure required, because I must be missing something significant. My reference is "Get Serious About Clamping" from the November / December 2007 issue of FWW, which I found on their web site.

In the article, Roman Rabiej gives a table of recommended clamping pressures for gluing up different woods using PVA glue. These values range from 150 psi for flatsawn glue face ponderosa pine to 1,200 psi for flatsawn glue face sugar maple.

They also test a range of clamps for clamping force. These include the parallel jaw clamp (i.e. Jorgy Cabinetmaster) @ 370#, the 3/4" "Pony" pipe clamp at 1,050# and the "I"-Beam clamp the highest at 1,350#.

Now, if I were to make a 3" thick x 8' (96") long work bench from laminated flatsawn glue face sugar maple, the recommended clamping pressure would be 3" x 96" x 1,200#/in2 + 345,600# = 172.8 tons.

To get this much force you'd need 934 parallel jaw clamps or 329 3/4" pipe clamps or 256 I-Beam clamps.

If you were to put a clamp every 1" of length and one on each face of the board, you'd still only get 192 clamps.

With Bessey K-Body's lined up touchung each other you'd only get about 48 in a 96" length and, at Bessey's claim of 1,500# force you'd still only have about 20% of the recommended clamping force. What am I missing? It's gotta be something simple...

[Andy Margeson]

It is kind of hilarious, isn't it. I was particularly curious when I read statements like this since I had used Titebond III in rub joints with success, meaning no clamping pressure at all. So, I emailed the manufacturer and here is the response I got:

"Rub joints can yield good results, but a clamped joint will be thinner and stronger. The loss in strength is literally a function of glueline thickness, which is optimal at 5 mils or less. Clamp pressure is limited only by the compressive strength of the wood and we have tested oak at over 400 PSI. Assuming perfectly fitted joints, lower pressures can produce nearly identical results since the purpose of clamping is to bring pieces into closer proximity if there are irregularities."

So, how much clamping pressure you need is a function of how much strength you need in the joint, how well the joint fits and how thin a glue layer you apply (which avoids the need for squeezing out the excess). In the boundary case, you don't need clamps at all. I have experimented with rub joints and verified that you can in fact get an incredibly strong joint without clamps. However, whenever possible, it is best to use as much clamping pressure as you can reasonably apply, I think.

I'm not an expert. What do the rest of you think?

[george wilson]

I don't believe in having to clamp terribly hard. If your joints are so bad that you need to squeeze them flat,they need to be made straighter before gluing,else they will probably pop apart at some later time.

[john brenton]

This study came up here about a year ago, with the poster mentioning the several bajillion tons of pressure needed to join cherry. I also noted the time tested success of rubbed joints and other "crude" methods of clamiping used in certain musical instruments that have a much heavier and consistent load of pressure from the strings than you'd ever get in any piece of furniture. I think this guy got a payoff from Jorgenson.

[Pam Niedermayer]

...I'm not an expert. What do the rest of you think?

I think FWW hasn't displayed much expertise either during the last few years; and that they're recommending the wrong glues and apparently the wrong clamps. :) Rubbing, assuming one's done a good jointing job, must produce piles of pressure that can't be approximated by off-the-shelf clamps used with off-the-shelf glues.

Pam

[Adam Cherubini]

The science is right and we are wrong.

PVA is NOT a stiff glue. To some extent, that is its chief benefit to woodworkers. It allows wood to move slightly and remains roughly intact. For this same reason, it is not good for parts under constant stress like the bridge of a guitar or parts in a piano. I glued the abutment of a wooden plane with PVA once and found the stress from the wedge drove the broken up out of the plane.

Because PVA is not stiff, it cannot transfer load over a gap very well. Epoxy and Hot hide glue, both being significantly stiffer, DO transfer load across gaps. In engineering we would call that a "gap filling" glue. The "filling" we're talking about is not cosmetic, it's structural. Epoxy can fill structural gaps. Hide glue can fill structural gaps.

So imagine clamping 2 boards together with a sponge between them. If you apply a single clamp in the middle of the boards' length, you can squeeze the sponge there locally, but the sponge pushes back such that at the free ends of the board, The sponge is not squeezed much at all. This is because the wood itself is deflecting. Now, the wider/stiffer the wood, the less deflection you get. Adding more clamps will create a "festooned" edge. The boards may touch under the clamp heads, and get pushed apart between them. So they clamps aren't forcing a bad joint together. They are squeezing the glue out.

This is how a PVA glue up is. So the number of clamps you need, depends on the stiffness of the wood. If the boards were infinitely stiff, a single clamp (placed anywhere along the bond) would suffice. But boards are NOT infinitely stiff.

So....When you use PVA glue, you need a bunch of clamps to develop the full capability of the glue/joint. In the example with the workbench, the area is so high that you would literally need tons to get to desired bondline pressure. Obviously, that's not going to happen. What is going to happen instead is that you will get spots like spot welds that hold the boards together.

If it were me...when it was me...Years ago, I was gluing up a beech workbench out of 5/4 boards 4" wide (glued face to face). I glued a single lamination each night and I positioned the clamps where I felt I needed the best bond strength (top surface and at the ends). 24 laminations. 24 nights. I used every clamp I had (maybe 16 or 20). That was the best I could do.

Now the last thing to mention is "what do we really need?" "We're building workbenches, not foot bridges". I don't know the answer. I know we're not getting anything near the bond strength advertised. We might be getting a tenth of it. I think Franklin advertises something like 2500-3000 psi. My guess is realistically we're probably getting 300psi, which isn't bad.

Woodworkers need to know a couple things about PVA:
1) It's hard to repair and therefore shouldn't be used for heirloom quality items or any surface or joint that may need to be reversed or repaired. I think it's a terrible product for veneer for example. I wouldn't use it for chair joints or dining table joints. I use it only for long joints like table tops.

2) It is not stiff and cannot transfer load across gaps. Therefore, it is inappropirate for a wide range of traditional joints where the joint is unclampable. (like a mortise and tenon) Here is it a double whammy. Not particularly strong, but also difficult to reverse/repair.

3) For butt joints, PVA can produce a near invisible bondline with excellent strength and just enough give to account for seasonal movement. I think this is where PVA really shines. But to achieve this joint, woodworkers need to clamp the heck out of it. The clamps are used to squeeze out the glue. They should use as many clamps as they have and concentrate their clamping at the ends where seasonal movement is most dramatic (especially when end grain is exposed). This is the rationale behind sprung joints. The spring helps increase bond line pressure at the ends of the boards. The preload at the ends also remains there, such that as the boards shrink the preload is releived first, and only after it's gone will teh bond line go into tension.

4) The strength values the manufacturers of glues claim are not wholly relevant to woodworkers. Many many woodworkers (including some here) do what they want with PVA and everything turns out alright. The reason for this is that their products see little real stress, either from seasonal movement or usage.

5) Woodworkers need to start thinking about glues in terms of which glue makes sense for what specific circumstances and not seek a single glue that works for every situation.

[george wilson]

By stiff glue,do you mean brittle glue? I am not opposed to using several clamps along a wide pair of boards being glued together. I just do not see the need to squeeze the daylights out of the joint. Over clamping results in a dry joint.

There is a difference in applying pressure all the way across a surface,and applying huge amounts of pressure.

[David Weaver]

That FWW article was a little goofy. The key is whether or not you want maximum glue strength from PVA, which is a level of strength you'll never need for gluing boards. I'll bet a lot of people on here have never had a furniture glue joint fail. I haven't built a whole lot of furniture, but the first two years, I used nothing but gorilla glue and PVA. I've never had a joint fail.

If a PVA joint fails, it'll probably be because the application was wrong, and not because the joint wasn't tight enough. As far as I know, you can't starve a PVA joint, but you can sure starve other types of glue joints.

I'd rather just avoid PVA in general, unless cheap is concern #1, and have managed to do that. Regardless of the glue type I've used, it's never been a large part of the budget for any project.

Articles like that one are why I no longer subscribe to FWW, that and the cost of the magazine for what you get out of it. Their website is OK, but unless you need a specific article on it, also something that was a once and done thing for me. I don't like that their CS won't answer email from their form (at least they didn't answer mine) and expects you to call them, and I find all of their offer cards and post-subscription follow-up junk mail to be really really offputting, both in the way it's worded ("your skills are diminishing already since you've let your subscription lapse" or "this is your seventh notice that your subscription has lapsed, we have a special offer ...you can renew for full price") and the fact that all of their "special offers" are to re-subscribe at regular price.

[george wilson]

I'm letting my subscription to a certain other woodworking magazine run out. I don't enjoy paying to read fallacious articles.

[Dave Anderson NH]

Andy's post contains a very important line about glue line thickness. From my experience both in woodworking and as a student at a 5 day adhesives course at 3M, glue line thickness can not be overstated in its importance.

A glue line should only be thick enough to completely wet out the mating surfaces. Obviously the appropriate thickness varies depending on the substrates that are being bonded and the type of adhesive being used. Glue lines that are thicker than necessary produce joints of greatly reduced strength because adhesives themselves do not have very much cohesive strength. Part of what this means is that the adhesive itself decreases in shear strength as thickness increases. A glue like PVA needs a thin glue line because creep is both inherent in the polymer and as thickness increases it is more subject to shear forces which allow it to move more.

[Joseph Klosek]

I had always used Titebond Extend to glue up bending rail for curved stairs. During my time as a finish carpenter I probably installed over 2000 lineal feet of bending rail. I would generally clamp every 6" with handscrews and use small bar clamps as filler. Given that the rail is in both tension outside and compression inside, rarely did I ever have a glue line failure. Although I had only gone back to see one railing I installed after about 7 years. It still was solid with no delamination.

The biggest problem I noticed with the rail were some areas had a raised glue line. I assume this is due to an uneven glue joint where the glue line was too heavy and actually squeezed out over the years. Or a combination of wood movement and glue creep.

I understand the science and fascination with such studies, but whatever happened to empirical evidence and common sense?

Articles such as these seem to raise more questions than answers. I assume most folks will just continue with their use of PVA glues and forget such studies even exist.

Although, it does help to flush out the internet "experts" and "know it alls" who tend to liven up the whole net forum atmosphere.

J.P.

[David Weaver]

I understand the science and fascination with such studies, but whatever happened to empirical evidence and common sense?

Articles such as these seem to raise more questions than answers. I assume most folks will just continue with their use of PVA glues and forget such studies even exist.

J.P.

This is the point that I didn't make that well. If the joint holds, good. If it fails, it'll probably fail not because of the gap, but because PVA was used where it shouldn't have been.

The only benefit from all of this stuff that I've seen is:
1) debunking the myth that gorilla glue is a strong gap filler
2) making people aware of what glues you should use where you expect future repairs

[Chris Fournier]

I'm not sure why folks are slagging PVA in this thread. This is an economical and solid performing glue when applied correctly in the right circumstances which is almost all of the ones you'd find in a general purpose woodworking shop.

I agree with George that it is very easy to starve a glue joint. This is somethng that newer woodworkers tend to do - crank the clamps until you have carpal tunnel syndrom! Edge glueing is a situation where glue joint starvation is very easy to achieve, the thinner the edge, the easier it is to starve it. PSI.

There are not many joints that I would consider needing to be "reversible" or repairable with the exception of some musical instrument applications and these are really limited but not exclusive to the violin family for the most part.

Avoiding PVA is a pretty foolish position and I can't see any logic behind it. I fully agree that there are different horses for different courses and that's why I have about a dozen plus different adhesives in my shop. Far and away I reach for the PVA most often. I have not experienced too many glue joint failures in my time and in most cases the failure was craftsmanship, not the glue.

As with finishing its important to read the mfgrs instructions, do some testing (destructive is very informative at times) and applythe right glue to the right job.

[george wilson]

One of my posts vanished. Maybe I caused it? I have made a few starved joints myself in my long woodworking career. I think that the question really isn't 100% black and white,but depends upon the type of wood as well as the type of glue,as to wether you can starve a joint.

[David Weaver]

You can starve an epoxy joint pretty easily. I think Franklin (titebond) claims two things:
1) you can't starve a glue joint with normal clamping
2) a surface doesn't need to be roughed to glue well with PVA

Someone on wood central had extensive dealings with franklin (how they did, I don't know, if they just asked questions of the engineers there or if they were involved on testing something), but I can't remember who it was.

Their guidelines have an interesting condition in the statements, if you look at their suggested clamping pressure, the verbiage right before it is "enough to bring a joint tightly together".

http://www.titebond.com/IntroPageTB.ASP?UserType=1&ProdSel=ProductLineTB.asp?prodline=2?prodcat=1

I suspect the commentary about 150-300 psi pressure has a lot more to do with making the joint very thin than it does with smashing the wood together. I don't know, though, and a question to them can clear it up. BUT, I'll bet that the critical issue is the thickness of the glue line, that's just my hunch.

Still, with 4 bessey clamps along a 48 inch glue line sort of glue up and an *accurately* jointed edge that comes together easily, i've never seen any kind of failure and I doubt any of us ever will. That's not much in PSI compared to the guidelines.

Repairability, expansion and unexpected issues in the finish are much more likely to be a problem.

[Andy Margeson]

I am curious to know how many of you use epoxy for mortise and tenon joints if in fact hot hide glue and epoxy are the only two appropriate choices.

[Jonas Baker]

Epoxy and Hot hide glue, both being significantly stiffer, DO transfer load across gaps. In engineering we would call that a "gap filling" glue. The "filling" we're talking about is not cosmetic, it's structural. Epoxy can fill structural gaps. Hide glue can fill structural gaps.


I'm sorry to nit pick your post, but I just want to mention that Hide glue is not a good gap filling glue. Hide glue has excellent resistance to cold creep, which can be a problem with PVA in certain high tension glue joints (like in a guitar bridge), but hide glue does not have excellent gap filling properties. It requires a tight fitting joint to have a strong glue joint. Just wanted to get that out there before someone gets the wrong idea.

I only use hot hide glue myself, and use rub joints in certain applications, and clamp in others, such as a larger glue joint. Rub joints can work really well, but when your gluing a large surface area joint, I prefer to use clamps.

[Adam Cherubini]

You can starve an epoxy joint pretty easily. I think Franklin (titebond) claims two things:
1) you can't starve a glue joint with normal clamping
2) a surface doesn't need to be roughed to glue well with PVA

Someone on wood central had extensive dealings with franklin (how they did, I don't know, if they just asked questions of the engineers there or if they were involved on testing something), but I can't remember who it was.

Their guidelines have an interesting condition in the statements, if you look at their suggested clamping pressure, the verbiage right before it is "enough to bring a joint tightly together".

http://www.titebond.com/IntroPageTB.ASP?UserType=1&ProdSel=ProductLineTB.asp?prodline=2?prodcat=1

I suspect the commentary about 150-300 psi pressure has a lot more to do with making the joint very thin than it does with smashing the wood together. I don't know, though, and a question to them can clear it up. BUT, I'll bet that the critical issue is the thickness of the glue line, that's just my hunch.

Still, with 4 bessey clamps along a 48 inch glue line sort of glue up and an *accurately* jointed edge that comes together easily, i've never seen any kind of failure and I doubt any of us ever will. That's not much in PSI compared to the guidelines.

Repairability, expansion and unexpected issues in the finish are much more likely to be a problem.

You've got it right. I was at least one of the people who had a chat with the engineering folks at Franklin around 2001-2002 and wrote about it on WC. At the time I was using glue in an industrial application and wanted to understand the difference between what I was doing and what Franklin recommended.

I caution woodworkers about relying on their experience in this issue. PVA is known to last about 50 years or so, depending on conditions. Can anybody honestly say they are supporting a large number of their 50+ year old PVA glued projects? The fact that nothing fails in our shops or in 5 years is not evidence that we are using the glue correctly. The conservators at the Philadelphia Museum of Art and currently repairing modern furniture put together with PVA from the 1950's. They have found it nightmarish. Regluing previously glued joints requires complete removal of old glue. Veneers are popping off and are virtually unrepairable.

I don't think PVA is terrible. But we should use it either carefully or knowingly. I vote knowingly! Frankly, I hate everything I made 5 years ago. I'd be happy if the glue lines failed so I'd have a reason to rebuild!

Adam

[David Weaver]

I caution woodworkers about relying on their experience in this issue. PVA is known to last about 50 years or so, depending on conditions. Can anybody honestly say they are supporting a large number of their 50+ year old PVA glued projects?

I guess I don't care if it lasts 15 (if not applied incorrectly) or 50 (if applied correctly). I generally would probably be in the "applied correctly" category with very clean joints with very thin and tight glue lines, and in a lot of cases earlier on I put screws in everything that didn't show in case the glue failed. I don't do that any longer, but I really don't use PVA, either.

I guess I live in a little bit of a black and white world, but I either want something to last indefinitely or I don't. If glue failure will occur at 50 years in something I want to last indefinitely, then that is not OK unless repair will be easy and it's known. For people using PVA, I don't think they care, or they feel that the glue won't fail if they use it correctly.

That's my point, anyway, know what you have and follow the instructions on using it. It seems to me that going to franklin or whoever makes the glue is a better place to go for advice than an FWW review.

I see franklin has some tips for wood that will be stained (which I don't do, but I would assume would translate not exactly, but close enough to people who use finishes that show surprise PVA anyway, like rubbed out lacquer over shellac). Generally, tape something you don't want to have glue on.

And i'm not dissing franklin, I use their liquid hide and like it OK as long as it doesn't show up past date at the local rockler (which it does sometimes, and it looks like rubber when that happens). I can deal with the expense of having to replace it fairly often for the convenience of not having to dodge my wife to make it in the kitchen.