We’ll be testing a dozen different glues and epoxies to see how they hold up for knife making.
My friend Steve Sando (great looking knives Steve) started mumbling about doing a glue test awhile ago and I thought it was a dandy idea. I said I would do one also. Well he did his and now I have to step up and do mine. That will teach me to speak up. Anyway, it’s a good idea so here is the start of it. I have about a dozen different glues and epoxies, a sheet of metal and some WSSI stabilized wood. I’ll start the glue up tonight and get at least part way and finish the glue up this week and let them sit a bit to cure.
3-7-05 Here is the start. Using approximately 1” squares of wood, each will be glued and clamped to the metal per the instructions. ha ha ha I crack myself up.
Instructions? Instructions?
We don’t need no damn instructions.
We glued the crap up and let it set.
Here is some whacked out glue. It is a two-part epoxy that is billed as waterproof marine epoxy. It says it will even cure underwater. I didn’t realize the stuff was blue until I cut open the tube getting ready to mix it up. The blue color basically takes it out of the hunt for being a suitable knife epoxy but I’ll test it anyway. It also says it is flexible and Steve’s tests show that it is a plus. We’ll see. Besides being an awful color of blue, it stinks to high hell. More pictures later as the great glue off continues.
3-12-05 The adhesives have all cured for a couple of days. Time to tap test. I’ll use three progressively larger hammers and tap the edges of the woodblocks to simulate dropping a knife onto a hard surface giving a sudden, sharp shearing force. 3 to 4 taps each piece.
3-12-05 To the right, the middle tap and below the big tap most commonly known as the BFH. Everyone made it through the little tap. Our first failure came from the middle tap, the contact cement just popped right off. I believe the little tap loosened it, the middle tap failed it. All passed the big tap test. I didn’t smash like a mad man but if you look at the picture below you can see where the wood was damaged on the edge by the big guy. I gave everyone an extra bump or two with the BFH and they all still held on – for the moment.
3-12-05 I stuck the test material into the freezer for a couple of hours and then took it to my shop to repeat the tests while everything was still sub-zero. I went right to the BFH and gave everyone several good knocks. The Devcon 2 failed immediately. The surface of the epoxy looked glazed and it appears the metal was smooth. Some of the epoxy sheared off and stayed on in the micro-grooves in the metal, but not all that much. I think a combination of things failed this epoxy, cold temp, minimal surface prep and it doesn’t like sudden-shock shearing force. Everyone else did a great job holding up.
3-12-05 Chuck Bybee kicked in some testing materials and a couple of test adhesives. Here I’ve taken some of the wood and cut it into blocks and then I’ve sand blasted half of the metal strips and will rough surface grind the other half to see if one preparation method provides for a better purchase than the other.
My sandblast cabinet. I cheap one from Harbor Freight but good enough. I used 70 grit AO to roughen up the surface on the test coupons.
3-12-05 Here is the start of the glue for round two of testing. Each block is glued up by the same glue on either end of the metal. The difference is one end was sandblasted, the other just roughed up from a quick grind from a fresh 36 grit belt. It’s going to take a week or so to get through all the glue up for round two and to allow a good cure. Some of these epoxies don’t reach full strength for 3 or 4 days.
3-16-05 To the left see some test coupons on top of 25lb bags of lead shot. 3 sets of which were glued up using PC-7 epoxy (which is now verified as amazing stuff) to test what effect surface preparation had on how effective epoxy held. 1 set was sandblasted, 1 set was roughed up with a fresh 36 grit belt, another set had no preparation done at all. Each of the 2 pieces was glued at a ninety-degree angle and allowed to cure for a couple of days. They were then placed in a vise and a load applied until they failed.
3-16-05 To the right you can see a load being applied to the test pieces. Here you can see a 50 lbs load held by a simple lap joint of approximately 1” square of PC-7 epoxy. I was stunned at how much they held. Each test piece held 75lbs! They all failed when I added another 25lbs so I never could measure accurately which surface held more but we can examine the failed joints and come to some conclusions. Note the metal is starting to bend from the leverage and weight.
3-16-05 The joint to the left is 36 grit, the middle sand blasted surfaces and to the right no surface prep at all. Starting from right note one piece ended up with most of the epoxy and the other piece had maybe 50% of the surface with epoxy residue. In all three cases, the epoxy failed, not the material, but in the test pair on the right, the adhesion partially failed by not sticking well to the metal.
The left pair shows fairly even epoxy failure with 10% or so adhesion failure. The center test shows very even epoxy failure with no adhesion failure visible. The residue coverage was very even on both pieces. The conclusion we can draw is that sandblasting a surface is clearly the best method of surface preparation followed closely by a good, even ‘scratched up’ surface from a new, sharp 36 grit belt. I’ve been using a 36 grit belt but going forward I will be using the sandblaster to surface prep anywhere I can for better performance.
3-19-05 Last week I left the test piece in my heat treat oven for 1 hour at 175F and there it sat all week. It’s warming up here in the spring so the garage has been in the 20’s and 30’s. I brought the piece in to whack it. The Devcon 5 minute epoxy half failed. You can see 1/2 of it broke off, the other half is still hanging in there. All the others took the beatings. Then I threw the whole thing in the dishwasher.
This is where I had to explain what I was doing to my wife. Even with all the goofy things I tend to do, this one still forced her to ask why I was doing this. To see which one holds the best was my answer. After the dishwasher cycle, I took the piece down and whacked it some more. Keep in mind, this is all with a large framing hammer and I have long since stopped tapping these things I am giving them 3 or 4 good thumps each time. Two more failures to report, the Permatex and T-88 epoxies both fail. The Permatex came loose in the dishwasher, the T-88 got tired of getting beaten.
The T-88 failed by 3/4ths. A quarter of the test piece is still attached and I’ll leave it that way just to see how long it goes. The leaders at this point are Acra Glass, Golfsmith Shafting epoxy and oddly metal super glue. The polyurethane glues, I believe will withstand all of the tests better than any of the others but it appears to be a dreadful product to work with on a knife – it’s fine for woodworking.
I also tested 324 Loctite Speedbonder with two metal pieces with sandblasted surfaces like the previous tests with PC-7.
The Loctite held 100lbs (!) until the metal stated to bend and I started moving things around to stop the metal from bending completely over. I was torn between running for my camera or stopping the metal from bending. I didn’t run for the camera, I should have. Once I moved it around the bond popped but it was amazing. I immediately glued up another test set and will try to get a picture of this time.
3-21-05 Here is what the test piece looks after it has been cooked, frozen, dishwashed, pounded, and now boiled for an hour and a half. The body count rises on glue failures. After boiled, we lost JB Weld, Super Glue Metal epoxy and the clinging partial pieces of T-88 and Devcon 2 ton 5 Minute epoxy. The two metal pieces you see were glued up with Loctite 324 at a 90-degree angle and then load tested.
This little 1” square lap joint held 100lbs for 45 seconds before it bent over and the joint peeled to failure. I still couldn’t get a picture of it holding the load so you will have to settle for the broken pieces.