View Full Version : Waterproof Glue Testing Plan

02-03-2009, 11:10 PM
Due for my Statics & Strengths II Lab tomorrow. Our assignment this week was to create a testing plan for something that we could test in class. Here is my testing plan.

Waterproof Glue Testing Plan

I. Introduction

a. This test is to test the tensile strength of the bonding capacity of three different glues in both dry and wet conditions. The test will be conducting by gluing up three identical samples (one for each glue). These samples will than be halved with one half left dry and the other half soaked in water. By tension loading the wood samples we should be able to determine how each glue reacts to being exposed to water for extended periods of time.

II. Materials

a. Three different types of adhesives will be used in this test. These three types are common adhesives and all are advertised as being 100% waterproof.

i. Two-Part Epoxy. When used as a glue two-part epoxies needed to be “thickened” with a thickening agent. The most common (and what we will use for this test) is wood flour. Wood flour can either be purchased or you can use the dust generated by a belt sander. Epoxy thickened with wood flour needs to have about the same consistency of peanut butter. Both pieces of wood will need to be coated with unthickened epoxy prior to bonding. This will ensure that the wood does not soak up all the resin from the epoxy and thus create a epoxy starved glue joint. The pieces should be lightly clamped so as not to squeeze the epoxy out of the joint.

ii. Gorilla Glue. Gorilla glue is a polyurethane adhesive which uses water as a catalyst and is advertised as “the toughest glue on planet Earth.” To use Gorilla glue both surfaces need to be dampened with water and than tightly clamped for 1-2 hours or for best results overnight.

iii. Titebond III. Titebond III is a relatively new comer to the adhesives market and is marketed as the “ultimate” wood glue. Titebond III is an advanced proprietary polymer and requires clean surface areas and high clamping pressure. It has an advertised strength of 4,000 psi.

b. Wood. For this test we will use standard off the shelf construction grade lumber (1 x 6).


a. Glue up. The 1 x 6 will be cut into 6 equal length pieces for our three test pieces. One pair will be glued up with thickened epoxy, one with Gorilla glue and one with Titebond III. Each glue up will use proper gluing and clamping procedures as outlined for the specific product and be allowed to cure in the clamps for 24 hours. An additional 24 hours will be allowed to insure that the glues have had the proper amount of time to cure.

b. Soak. Each piece will than be cut in half to create two equal pieces. One piece from each glue up sample will be sit aside, while the other half is placed in a water tank and allowed to soak for a week.

c. Test preparation. Once the test pieces have been allowed to soak for one week they will be removed from the water and than each of the 6 samples will be cut into three equal test samples. This will give us a total of 6 samples of each glue (3 wet and 3 dry) for testing purposes and will help ensure that we achieve a good test range. Each piece will need to appropriately labeled to ensure that a proper record is kept of each piece.

d. Moisture Check. A moisture meter will be used at this point (optional) to determine the moisture content of each piece.

e. Testing. Each piece will than be individually tested in a tensile testing machine to determine the ultimate strength of the glue joint. For each incidence there will be 3 test samples and each test sample results should be recorded and graphed for analysis.

IV. Conclusion.

a. From this test we should be able to determine which glue works best in dry and wet conditions. This would allow us to make determinations on whether a glue is acceptable for use below the waterline or if it should only be used above the waterline.


02-04-2009, 01:55 AM
Sounds like you have it figured out. What testing rig will you recommend using?

02-04-2009, 01:59 AM
Gorrila glue is a foam and a joke, and should not be used. If you want to test a polyurethane, use P.L. premium construction glue.(not subfloor glue) But there is plenty of info on glues in the building forum. Don't reinvent the wheel..

Captain Blight
02-04-2009, 02:34 AM
Yeah, I was going to suggest that myself. A few guys, myself included, would really like to see actual lab testing of PL premium.

Now, Mr Smalser's already done some testing of his own and one thing that struck me about his results was that after weeks and weeks in the water, the wood wasn't anywhere near "saturated." So if you want to test the glue on actual wet wood, the wood samples will have to be sized so that they can actually be considered "wet." In other words, fairly small.

Food for thought.

Bill Lowe
02-04-2009, 06:06 AM

Bill Lowe
02-04-2009, 06:11 AM
PL Premium and Cherry

02-04-2009, 06:30 AM
I will be using something similar to the Instron testing machine shown below.


Not that I'm re-inventing the wheel. For a lab assignment we were asked to come up with a testing plan of something we wanted to test and this was something that I thought would be interesting. BTW these are the glues that I'm interested in. I might throw in the other if we actually get to do this test.


Ron Williamson
02-04-2009, 06:32 AM
Methinks Bill needs a longer tenon,ie.another 1/2"deeper than the panel groove.

Bill Lowe
02-04-2009, 07:07 AM
That would improve its strength, but it took a lot pf pressure to break it, more than enough for cabinet doors.http://picasaweb.google.com/lh/photo/qsGSV0l5gi3s4-0RXmYIsw?authkey=Y888E5WeE1E&feat=directlink

Gary E
02-04-2009, 08:08 AM

Did you plan that to fail or just pick out a weak board knowing it will fail because you created the less than substantial joint?

02-04-2009, 08:12 AM
My understanding is that Bill did the test in the lead up or part of building cabinet doors. I doubt that there was any intent to hang a keel to them and go to sea. So in his case, its probably not a real issue. ;) Heck biscuits are used all the time in butt seams and have little to no structual characteristics either.

Gary E
02-04-2009, 08:17 AM
My understanding is that Bill did the test in the lead up or part of building cabinet doors. I doubt that there was any intent to hang a keel to them and go to sea. So in his case, its probably not a real issue. ;) Heck biscuits are used all the time in butt seams and have little to no structual characteristics either.

IF the idea was to test a corner joint for cabinet doors, why load the joint in a way that the joint wil never be loaded when in use? Those tests are useless information.

02-04-2009, 08:23 AM
IF the idea was to test a corner joint for cabinet doors, why load the joint in a way that the joint wil never be leaded when in use? Those tests are useless information.
I think many do test the limits in products that they themselves have not used in new applications just for a mental excersize. This is probably the case in his own stress test. Meaningless/ Thats always up for ones opinion. In several previous disussions in some of the PL products that I have personally used, I have also tested the flexibilty in the sealant brands which also has a greater stretch factor while still have a pretty good adhesion than say even the 5200 brands. Thats my take on it anyway. The reason I say this is that on one of the threads using the product this actually came up.

Gary E
02-04-2009, 08:35 AM
IV. Conclusion.

a. From this test we should be able to determine which glue works best in dry and wet conditions. This would allow us to make determinations on whether a glue is acceptable for use below the waterline or if it should only be used above the waterline.


For a basic test, for this class that should do ok and is most likely what your teacher is looking for... BUT, for a real world test that industry would use, so much more is involved.

For example, the type of wood being glued, the connection between the wood and the testing machine, the rate of loading, the list gets rather large and it all boils down to, Make the glue joint bigger so it wont fail and use bolts along with glue.

Good luck

02-04-2009, 10:39 AM
The pieces will be edged glued together and the jaws of the Instron tensile testing machine will grasp the 1x. We saw a demonstration of this machine the other week. It has self tightening jaws for grasping the material. We did a tension test on plastic and on steel.

I will probably reduce the material to 1x4's so that the jaws are closer to the glue line and so I am testing the glue joint and not the wood. One of my concerns is that with any of these glues I will probably get a stronger glue joint than the material being bonded.

If I have enough material I may do stages of wet soaking (i.e. some for 1 week some for 2 weeks and some for 3). Depends on weather, material and time left in class.

At the moment I am using construction grade lumber because it is cheap and easy to get. The only other material on hand is white oak and 1) I don't want to waste my good white oak on a school test and 2) Construction grade pine is cheap enough to waste on a test. I do have some western red cedar but I think that the pine might be a better choice.


02-04-2009, 10:58 AM
The other option is to go and spend twenty or thirty bucks and by a chunk of good lumber. What would you recommend using? Keep in mind that epoxy is supposed to not bond good with white oak. Maybe this would be a good time to kill two birds with one stone and do the waterproof test and the epoxy/white oak test at the same time.


02-04-2009, 11:50 AM
I'll think about that. Let me rummage around the shop and see what is laying about.


Bob Adams
02-04-2009, 12:08 PM
I was gonna make an intellegent reply, but since this is the bilge here's a good test.

Apply the test adhesive to Woody's fingers and see how long before his inane threads resume.:D

Captain Blight
02-04-2009, 12:50 PM

I'd like to put in a request for white oak. I wish I'd known about the test sooner, I would have even sent you a plank! What are ya gonna need, like three or four board feet all told?

Anyway: We know anecdotally that epoxy doesn't work well with white oak. Here's a chance to quantify that. I'm thinking if you apply thickened epoxy to a freshly-planed surface, we might all be surprised (for good or for ill, I have no idea). But I've never seen the numbers on epoxy-on-oak.

02-04-2009, 12:59 PM
I'm going to look around the shop and see what I have. The biggest obstacle I see right now is it warming up enough outside to do the glue ups. This weekend might give me a chance to do that.

As far as length I'm reckoning that each final test piece needs to be about 3". If I have 6 pieces (3 wet and 3 dry) of each glue and 3 different glues that is 54" or 6' by the time that you figure in waste. For every test range added (for more test samples in longer saturation periods) you need to add 27" or around 2-1/2" with waste.

Like I said I will look around and see what I have, I may go down and buy a piece just for this.


02-04-2009, 07:26 PM
I wonder if tension testing is an accurate indicator of real world performance. (just thinking aloud here).

I think having a glue joint in tension probably reflects bad joint design.

02-04-2009, 07:34 PM
Shear testing may be the better option. I need to see if we have a static loading shear tester. But if not I will do the tension test. I"m looking for the ultimate strength.


02-05-2009, 05:18 AM
You could probably use that type of rig anyway for shear if you oriented your joints vertically. It is a pretty impressive looking piece of equipment.

02-05-2009, 07:46 AM
Ours at school is not as impressive looking as the one I showed, but it is impressive nonetheless. They have other machines for doing shear testing, but not sure what. They do have one which is a pendallum that swings a weight down through the object. This is more of a dynamic tester though and what I need is a static tester for testing shear. I'll check with the instructor today and see if we have such.


02-05-2009, 08:11 AM
Chad, good for you for doing this. There is far too much anecdotal "knowledge" and far too little actual testing done by us woodworkers in general.

It sounds as if you are starting to think about how the test is to be set up regarding grain direction, etc. This is important, since the glues you're testing are potentially much stronger than the wood -- you could easily wind up just testing the wood and getting no meaningful results for the glue. Bill's photo demonstrates this pretty well -- it takes practically nothing to make cherry split like that; mucilage glue would give identical results.

I think you're on the right track to be considering testing the joints in shear. Takes a little design thought to get away from peeling or twisting the joint. What if you glued the pieces end-to-end using a long half-lap joint? So you're pulling the piece in tension along the grain, then it's cut to half thickness and glued face-to-face with its mate, so the joint is centered relative to the machine's pull.

Also, the bigger the joint area, the more accurately you can quantify results, and the less the effect of some small local defect in the wood or joint.

Plenty to think about.

02-05-2009, 09:25 AM
Just an observation, but, would an individual test on White Oak not be fairly meaningless? i.e what are you comparing it to?
To produce a relevant result would you not have to also test pine, cedar, mahogany, teak etc. And compare those.

Dunno of this is relevant but is clamping pressure not a variable that needs to be taken into concideration? If it is I would think it would be pretty easy to attach a torque wrench to a clamp.


02-06-2009, 12:44 PM
If this test was specifically to test the bonding capacity of epoxy to white oak than I would design the test around that premises. But is not, it is to test the strength of various glues. With that you need the same type of wood throughout. Throwing in white oak gives us the chance to at least glance at that end of the spectrum. We know that in most circumstances that epoxy is the superior choice.

Back to the test, after much thinking and talking I will probably do lap joints instead of butt joints. Lap joints lets you use the same type machine and do shear test.


Here are the three types of basic test that are done on glues.



02-06-2009, 03:28 PM
Thankyou Chad
for sharing this kind of knowelege with us and I look forward to your tests. since I am a bolger box/pete spector quick and dirty boat build kind of guy I would have tended toward ac x plywood. but guess that puts me in the minority here ( grin)

a comment a few months back I belive it was fine wood working had an artical on glue strength and they picked tightbond as far stronger than the PL premium glue which came as a suprize as the advertising for gorilla and similar glues has been very effective. plan to use the pl for this springs boat(s)