Templates for Cutting Gears from Wood
woodgears.ca
woodgears.ca
Funny how my old stuff shows up on hacker news from time to time, though it was less than a year ago that I ported it to javascript, used to be flash, cause when I created it more than ten years ago, javascript canvas wasn't mature enough yet.
Big fan of your videos. I hope your tendinitis improves and you can make it back into the shop soon to make more power tools out of old furniture other people have thrown out.
Cheers!
> Funny how my old stuff shows up on hacker news from time to time
I really like this feature from Hacker News. People sharing periodically the same projects that have inspired or awe them. These kind of projects are not "news", and that makes them better, they've standed the test of time and keep amazing people year after year.
I really appreciate your style of videos as well. Thanks for not "chasing the algorithm", and for staying true to your fanbase (as demanding as we can be).
I've learned so much from your general approach to problem-solving, as well. So, thanks for not editing out those parts.
I used this tool to make a PVC front rack for my bike: http://dougkoellmer.com/front_rack/attached_side.jpg
I tried to do all the calculations by hand, but eventually realized I needed to make a little custom CAD program like this to help me. I tried more general-purpose CAD programs but couldn't figure out how to set up the constraints I needed.
Anyway I think there's a lot of usefulness in purpose-built online tools like this for generating templates and plans. Such programs are pretty easy to whip up too using the Canvas API and a basic geometric primitives library.
They are both basically "CAD/CAM for programmers" - and allow for parametric models to be designed. In fact, there are more than a few models like that on Thingiverse, so if you needed to resize something to fit your needs, you can (plus tons of gear and other similar custom power-transmission systems).
Of course, that's all for design and modeling - it wouldn't help for other problems, where some form of calculation and checking of a "model" (in the sense of a simulation or idea - not necessarily a complete physical object rendering) is needed.
You see a ton of these kinds of tools in the electronics and hobbyist realms; quick calculators to tell you how much resistance to use for various LED configurations, or 555 timer customizers for common circuits - things like that. I know of one out there which gives you torque/power curves for a variety of different RC hobby electric motors. There are of course several for physics simulation and testing, too...
I once had a problem with OpenSCAD when I needed to sort - within model - objects of different kinds. The example sorting algorithm (a quicksort) is shown in the documentation, but it was working only with a particular kind of objects.
Long story short, OpenSCAD turned out to be powerful enough to have a Lisp written in it. Lisp manipulated strings and allowed factoring out the sorting algorithm and applying it with different comparison algorithms, supplied as parameters. Given Lisp, I think it's possible that a lot of things can be done in an OpenSCAD script.
Check out https://www.youtube.com/watch?v=Mn0m104NVAY for how I tied and tightened all the knots.
If I did use holes though, assuming I kept the lincoln log joints as well, then I reckon that would weaken the joints too much, e.g. crack the pipe on hard impacts. Also all the weight of all the bolts would add up quick. Hmm I suppose I could use plastic bolts of some kind, but then they'd have to be thicker, thus making the holes bigger, thus weakening structure further. Just thinking out loud. There could be a way, but yea I was proving the lashing concept first and foremost.
Thanks for the question. :)
As a parallel case, consider multi-ply composite construction, where sheets of fabric, sometimes of sheets with a single fiber orientation, are laid at angles to each other to spread the loads in exactly this way. They both better approximate an isotropic material, and are stronger than an equivalent thickness of singly-oriented plies all in the same orientation.
Gears are subject to predictable stresses, and so being able to orient a material to best handle those stresses will make a stronger gear. Plywood, like anything else, is a compromise that, as you say, has a more isotropic character. Obviously, there is a wide range of strength in wood, and so plywood. I think it's reasonable to say that one could construct a stronger gear using teeth from readily available solid wood, then readily available plywood. Good luck finding hickory plywood :).
To your first sentence though, if you take the same size and quality solid board and piece of plywood at your local hardware store (making a reasonable attempt at an apples-to-apples comparison), the board will be almost certainly be stiffer and stronger with the grain, than the plywood will: https://www.fpl.fs.fed.us/documnts/fplgtr/fplgtr190/chapter_...
He uses the gears in many heavier-than-you-would-think-duty applications.
Mathias was an engineer for BlackBerry Ltd. before he retired young.
The visualization there is pretty cool though!
In wood, I've cut gears with my CNC Router, that works great for wood precision. In metal, I get a gear cutter set and cut them out of blanks with a lathe/mill. Metal gears tend to be smaller and require much more precision. There's no reason why the shape of these gears wouldn't work, though.
But of course if your metal gears are large enough to cut from the flat side then this template will work fine.