Glue Wars - testing ideas please

[Sando]

Well the last testing method did not provide the results I was looking for. So hat in hand I'm looking for ideas.

First let me explain better what I'm looking for.....

We all use some kinda epoxy on full tang knives. Why? to hold it together? Partly. But mostly to keep the slab/tang seam sealed. Welding is out for wooden handles (That I didn't have to test).

First I need to ask myself, what does it mean for the epoxy to fail? Picture your knife 100 years from now. The epoxy has cracked, shrivelled, something. Well, I can think of two things. One, it lets water (moisture, chemical gasses, hand oils, .....) in under the slabs. Second, it alows the slabs to slide around a bit even if bolted.

So in evaluating epoxies how do you know it failed? If I bolt the samples on it's really hard to tell. The only thing I can think of is a slight jar. If the epoxy holds when it's new, but things fall apart after a trip thru the dishwasher then I know it failed the dishwasher test.

My last test had a big problem, the 'jarring' effect was way to strong - nothing held on long enough to determine the effects of the 'environments'.

I'm thinking now to drop the sample down a 2 foot pipe on to a block of wood. That doesn't sound too extreme does it?

----------------------------
Part two:

I did learn that some epoxies shrink and the epoxy pockets gave an advantage to those epoxies that don't shrink. So here's my new attachment method (based on comments already made).

Dymond wood (It's more consistent than the natural stuff)
80 grit finish on the steel
36 grit finish on the Dymond wood
Light clamping (I don't have enough spring clamps. so I used hand squeeze clamps, but light enough the samples can easily be slid around).

This method will create thicker glue lines than anyone would probably want, but things will probably hold longer.

-------------------------------------------
Back to my question.

Do you have any ideas on what to subject the samples to (dishwasher cycles is already on the list)? You know ways to simulate a hunting knife left in the snow or on the dashboard of a truck or forgotten in the fishgut barrel or my wife using the butt end to break frozen chicken breasts......

Do you have any other ideas to test if the epoxy has given up?

Steve

 

[galadduin]

I thought, the impact test must be standardised. I saw impact a test equipment in a radio factory when I was a child. It was just like a catapult but not releasing the piece just hitting to a block, a simple design but every piece recieved the same force. Also, the pieces must be seperated, not on a long bar IMHO. That way you can treat the samples seperately and equally (remember the comment on vibration).

I would make the test not to wood but brass, copper, horn, micarta, corian, etc at the same time. I would love to make a test here but I have not enough products to compare.

Best wishes
Emre

ps: I began a test here but a simple one. Just comparing the pipe sealant vs. metal-epoxy. This epoxy is a Turkish brand, having a two day curing time. I was informed by the seller that this is the stongest adhesive amongst epoxy products. That seemed to me a very bold expression, but I will see.

The test pieces is D2 and brass on D2 (making four seperate pieces). These are prepared yesterday and tomorrow curing time finishes. Then shock test and dishwasher.

 

[DaveH]

I wouldn't polish the steel as much, Perfectly smooth steel means 100% of the forces are shear. Left a little rougher and all those ridges convert some of tthe forces to compression and tension.

It would be interesting to perform the same test on and old file.

 

[Kim Breed]

I have used 2 pieces of 1/16 steel two feet long by 1 1/4 wide. ground to 50 grit 2 inches in from the end. apply epoxy, overlap on grind marks. lock in vice for 24 hours. you can flex like a bow and you can also stress horizontal.

 

[Wulf]

If you're up for it, it might be interesting to see how they behave in different "climates." The dishwasher test will certainly help answer the question of how they hold up in a hot, wet environment. Those interested to learn how they might hold up in a cold environment, however, might like to see a test whereby the samples are stuck in your freezer for a night and are subjected to another impact test when they're nice and cold.

Another interesting test would be to affix the handle material to a longer tang (using just one pin or bolt), then after curing, clamp just the handle material in a vise and use a torque wrench to twist the tang loose, measuring the amount of torque required to break the bond.

These are just ideas, not necessarily suggestions. I'm just happy to be able to observe your tests because they've been very informative so far. Thanks for doing this!

By the way, the problem with using dymondwood is that you'll have to actually buy dymondwood!

 

[pso]

Steve

Thanks for doing all this work and sharing your results. I wish that my co-wrokers here were as thoughful, thorough and meticulous as you are.

When dropping the test piece itself, it would be difficult to ensure that it hit in the same orientation all the time, leading to inconsistent results. I would suggest taking a page from UL's test procedures for the impact testing. They have a series of impact tests for the plastic parts of enclosures or cabinets of electrical equipment. They drop a steel ball onto the part of the enclosure to be tested from some predetermined height. If it is a small target, they drop the ball down a tube. I don't have the UL standard that tells you how to do it handy right now but I could look it up if you wish. The European version of the test uses pendulum made of a steel ball, roughly 2 inches in diameter, on the end of 1 meter of cord. the pendulum is pulled back to 45 or 90 degrees before being released. Either of these methods would probably give more consistent results than dropping the piece itself. You can direct the impact so that it produces a shearing force or so that the impact is perpendicular to the bonding plane. As Emre has suggested, it would probably be best to use separate pieces of steel for each sample to prevent interactions between the tests on the various samples.

If I am not mistaken, many of the adhesives reach maximum strenght after some period of time like 24 or 48 hours. It would probably be a good to just make up the test samples and let them sit for a week. You are then not constrained by the number of clamps available either since you can do them in batches and an extra day or two would not really make a difference. I suspect that most customers would not get a chance to abuse a knife until at least 24 or 48 hours after the handle was attached to the knife.

A temperature cycling test may be of value. The samples would be placed, alternately, in a freezer or in a warm (150 - 200 degree F) oven. Let the samples soak for an hour in each environment. Transfer the samples directly between the two environments to maximize the thermal shock effects. Do something like 50 to 100 cycles. Yes, it would take a while and be rather tedious. I had access to such a test chamber that did this automatically many jobs ago and I am pretty sure that it cost over $100k. We used it to check for microscopic leaks and cracks IC (chip) packages destined for military applications where the part absolutely had to work just once and for just a few minutes.

Lastly, if you want to see if the impact or temperature cycling tests have compromised the seal between the two pieces, soak the sample in some water with dye in it. Choose a color for the dye that would contrast with the color of the metal and handle/bloster material. Alternatively, an air conditioning repair place would have some flourescent dye that they inject into the A/C system. Leaks can be detected by shining a UV (I think) light on the area. I am not sure if the A/C dye is water soluble. It may be soluble in oil though.

Thanks again for all your work.

Phil

 

[Rob!]

All great suggestions and it's a difficult matter to find the best product because every application is different. I think it's a great idea to test something relatively stable like dymondwood first - eliminate the extraneous variable.

For impact testing, the Canadian military used to use a cannon to fire chickens at jet fighter windshields - simulating birdstrike in flight. Why do I mention this? Because instructions and procedures have to be carefully described. No one told them to thaw the chickens first.

 

[FlaMtnBkr]

Like everyone else, thank you very much for the tests.

You are wanting to find out the best adhesive to use on knives so I would follow that form. Maybe not put pins, because that is a mechanical bond that would make it harder to tell when the epoxy fails, but perhaps drill a hole thru the metal so you have an epoxy 'pin'.

Also, no one wants a thick glue line on their knife so I would suggest not using one as that might increase the strength exponentially but would not happen in actual use.

For the impact test you had everything mounted to bar stock which can flex quite easily in one direction. What if you mounted it to a piece of square tubing that will not flex as much in either direction? Or better yet a piece of square tubing that has a slightly longer round tube running through the middle and the gaps filled up with 'great stuff' foam to dampen vibration. Then the whole thing could be slid over a smaller diameter pipe and dropped from a marked height onto a piece of wood where the inner pipe would hit first and there would be less vibration transmitted to the test pieces because of the foam and the fact the sqaure tubing is more stiff in all directions. I would think this would give all pieces a level playing field but flex in itself might be a good test. Someone who makes a 'sharpened prybar' might have thier knife exposed to flexing if the knife is used hard. It might be worthwhile to do a flex test if it can be applied to each piece uniformly.

As for other tests I guess just testing normal knife conditions is all that is needed. Some here make kitchen knives so the dishwasher test is very informative as it also would represent a very extreme wet and hot enviroment. Another environment is hot and dry so maybe the oven on a low setting (wouldnt think it gets much hotter than 150 deg in a car on a hot summer day). Another condition would be cold so the freezer would be a good test. Lastly, soaking in water would be a good one as I am sure many knives see there fair share of blood and sink time.

And the last thing that pops into my head is about the shrinking part. Have you considered that the epoxy was maybe too thick to fill the holes in the wood all the way and it appeared that the epoxy shrunk when it was never full? I can see an adhesive that works by evaporation shrinking (such as the gorilla glue) but a two part epoxy that 'cures' I would not think would shrink much. If you want to test that maybe get some thong hole tubing and fill it to the top that way when it dries you can see if it caves in or remains level?

Just some thoughts I had reading through the glue wars posts and that popped into my head while typing this. Hope I didnt over look anything...Good luck with round two as we all are looking forward to it!!!

Ryan

 

[Sando]

Dang those at good, well thought out ideas!

I really like the square tube/slide down a pipe idea. I was so focused on using stainless that options such as that didn't occur to me.

PSO, standardized tests should also occured to me. As you guys seem to fully understand, it isn't so much which epoxy is the strongest or exactly quantifing attributes. Rather to run a test to see if the epoxy has given up.

-------------------------------------------------------
Re: glue lines

Good point. Who cares if it holds and is ugly.

I was really struck by the whole issue of glue line thickness. E-120HP and DEVCON seem to want a consistent thickness. Not too thick not too thin. The engineering docs on E-120HP suggests a thickness of 3 to 9 mils. That's visible, but not too bad.

For me the epoxy pocket concept is dead (a hole in the tang to form an epoxy pin appears worthwhile). However, I am going to really rough up the center of the tangs from now on. Then I can squeeze a thin glue line but the bulk of the surfaces will have the right thickness. Does that make sense?

-----------------------------------------------------
Some changes:

I've excluded Gorilla Glue from the test. Looks like an excellent product and would work well, but the mess as that foam expands over the ricasso.... I don't want to deal with it.

I've excluded super glues. They have a respectable shearstrength, but I don't believe they have the sealing qualities and I don't trust them for aging.

----------------------------------------------------
Impact test:

How about placing the sample in a 2' pipe vertical on the concrete floor. Then drop a small weight (1 pound). Flip the sample over and do it once more.

The vibrations this causes won't be perfectly even, but we're just trying to see if the epoxy has let go, not what's strongest.
---------------------------------------------------
Environment:

More great ideas there!

OK I'll leave the sample alone to cure for a few days (I already made this one, so sorry I can't include those construction ideas this go-around.)

Then a gentle impact test.

Freezer overnight, warm to room temp

Impact test

Freeze overnight, impact test while cold

Dishwasher (several cycles)

Impact test

heat to 125 degrees overnight, warm to room temp

Impact test

Heat to 125 degrees overnight, impact test while hot

-----------------------------------------
Then the flex test as mentioned with the vice

Then boiled in water

.....??

Thanks again guys you're renewing my interest in this project!

Steve