I have occasionally tested different types of wood joints. My test apparatus consists of an L-shaped jig, with a bottle jack to apply the force, on a bathroom scale to measure it. The main drawback of this system is that the act of pumping the jack affects the readings on the scale.
A continuously increasing force would be much better. Thinking about this, I figured a screw type mechanism would do the trick. Going through my junk, I found the necessary materials.
My jack will push a threaded rod up, by turning the nut (C). This nut pushes against a washer below it, forcing the threaded rod to move up. The threaded rod is guided by a bronze bushing (B). To keep the threaded rod from turning with the nut, another nut, with a shaft through it, is attached to the bottom. The shaft is inserted through a 1/4" hole that I drilled in a coupler nut. I cut the coupler nut in half to make it less bulky.
I made a big gear to turn the nut with, with a smaller gear to turn the big one at a right angle.
I designed these gears with my gear generator program. What I'd really need is bevel gears, but such gears can't be made with flat paper templates, so my program won't do them.
But if you bevel the sides of the teeth by the same value as the pressure angle of the gears, they work a bit more like bevel gears and mesh fairly well at a right angle.
I tilted the bandsaw table 20-degrees, which is the same as the pressure angle of the gears. I then attached a piece of melamine coated hardboard to the table. I pushed this piece of hardboard partly into the blade so it acts as a sort of "zero clearance insert" for the bandsaw. This cuts down on tearout where the blade exits the wood.
Making the right-side cut on every tooth on both gears.
To make the left-side cuts, I needed to tilt the table in the other direction. And like most bandsaws, mine doesn't tilt to the left. So I rigged an angled piece of wood onto the table to get that tilt.
The catch with cutting the teeth at an angle is that they get very narrow towards the back. I should widen the teeth in the design so that they are the correct thickness at half-depth in the plywood, and they'd be perfectly snug. But I was too lazy, and just cut slightly outside the line to give them more width. But that wasn't adding as much width as I should have.
When I made the gears for my tilting router lift I imported them into SketchUp and edited the design to add the extra width.
At least, with these gears loose like this, there's less risk of them jamming. I may have to re-make the small gear if the teeth break once I use it, but I think the big one will be ok regardless.
A quick check on how smooth the gears run against each other. No problem!
I then drilled a hole, just slightly smaller than the nut and pounded the nut into the hole. I drove it in using a long piece of wood to make sure I was driving it in square. It would be too easy to get it crooked if I hit it directly with a hammer.
Next I need to mount the smaller gear to run at a right angle to the big gear.
I drilled a hole for the bushing, and hammered the bushing into a piece of wood. Snug fit!
Two pieces of wood will hold the small gear like this.
I attached the small gear to a piece of 1/4" shaft by drilling a 15/64" hole (about 0.4 mm smaller than the shaft), tapering the end of the shaft, and hammering the shaft through the undersized hole. I also glued a second layer to the front of the gear to get better rigidity (I glued that layer on after driving the shaft through the first layer)
The gears have a 20-degree pressure angle. This makes the teeth pointy, and makes them mesh better. But it also means that under load, they are also pushing each other apart. To support the side-load on the big gear, I mounted a rollerskate ball bearing in the piece of wood. The ball bearing fits tightly on a square piece of wood, which is screwed into a cross-shaped cavity that I hogged out on the drill press.
I made two blocks with dadoes cut out of them to guide the coupler nut on the bottom side.
These are just screwed to the top from above. To get the screw holes aligned precisely, I first lined up a block, clamped it in place, then drilled a pilot hole though both parts. Then the same with the other block.
Then a piece of plywood on the bottom. The plywood is oversized to also act as a base for the screw-jack, so it won't tip over.
The small gear is to be turned by a drill. But I didn't want to have chuck up a drill to the shaft every time. That, and the weight of the drill would throw off the scale's readings.
So I made a coupling, to act sort of like a U-joint. the fork, on the end of a dowel, fits over the bar. The metal shaft fits into a hole in the middle of the fork to help keep it aligned as it spins.
Testing it out. It's a bit loud, with the drill running and the gears grinding against each other. But it works adequately.
Screw-jack on the testing apparatus.
I put two dowels under the screw-jack base to allow it to roll side-to-side. If the joint deflects as it breaks, the jack can just roll to the side, even under load. That way, I can be sure the applied force is always straight up.
First test of the tester. With a pocket hole joint! (I'm no fan of pocket holes, but they sure are quick to make). The apparatus worked smoothly, with slowly increasing force until the joint broke. I think this will speed up running tests in the future.
The joint broke at just 65 pounds (290 N) applied 20 cm (8") from the joint. Not very strong. Hopefully, the jack will also be strong enough to break more traditional joints too.
Using this jack to test pocket hole joints