Although I don’t update this site very frequently, I usually *eventually* get around to doing a write-up for my more interesting projects. I still aspire to doing that, but my free time has been at more of a premium since the arrival of el bebe, so the documentation has suffered more than anything else. On that note, I’ve decided to start lightly documenting some of my past projects as quick blog posts. Some of them might get a more in-depth writeup at some point.
The Curlometer
For several years my wife and I were enthusiastic curlers. It’s a fantastic sport for grown-ups, and if you have access to a local curling club, I highly recommend it. Once you know how to build electronics, it tends to burrow its way into other aspects of your life, and curling was no exception.
The first project was the “Curlometer,” a simple sensor brick that could be attached to a 40lb curling stone to monitor it as it traverses the ice. The entire device consisted of a Teensy 3.0 microcontroller, a 9dof sensor board that provided a magnetometer, gyro and accelerometer, and a AAA battery pack to power it all. It was a simple data logger application, so the Teensy just sampled all of the sensors for a minute or so as the stone made its way down to the house. I roped my friend John (and fellow curling enthusiast) into writing some analysis software for it, but the whole project turned out to be less interesting than we hoped.
First, the accelerometer data turned out to be nearly useless. For those unfamiliar with this fantastic winter sport from our friends up north, one player crouches at one end of a hundred foot long sheet of ice and slides a 40-lb piece of granite towards the ‘house’ at the other end, imparting a slight rotation to the stone as it leaves their hand. The ‘skip’ stands in the house and provides a target to aim at, along with calling out when the other teammates should ‘sweep’ the stone. The two aforementioned teammates run down the ice slightly ahead of the stone with brooms and, when directed, furiously sweep the ice in front of the stone with their brooms. This has the interesting effect of decreasing the friction between the stone and the ice, causing it to both decelerate and rotate more slowly, thus causing it to ‘curl’ less than it otherwise would towards one side of the sheet.
The intent with the curlometer was to see if we could instrument the stone as a means of measuring the performance of the sweepers. The gyros (and the magnetometer, effectively working as a gyro) actually worked reasonably well, but the accelerometer on the board was mostly intended to sense things like an object being dropped or hit, and wasn’t nearly accurate enough to detect the gentle change in deceleration of a stone as it was being swept. After two evenings of playing around with it, we decided to retire the curlometer. It was a fun experiment, but it wasn’t really interesting enough to warrant further effort (or detailed documentation).
The Super Broom
This project went on for far longer than the Curlometer, and really deserves a full write-up at some point. As I mentioned above, sweeping the ice in front of the stone as it moves down the sheet is a key part of the sport of curling. The broom (typically a synthetic pad with a carbon fiber broom handle these days, although some old school people still rock actual straw brooms occasionally). When I say ‘sweeping’, one might have the gentle side to side motions associated with cleaning up crumbs or dust come to mind, although that really doesn’t do it justice. You’re basically trying to melt the ice via friction, so you’re applying as much downward pressure as humanly possible, while sweeping your broom back and forth perpendicular to the motion of the stone (while you run down the ice with a teammate who is similarly sweeping, while your skip yells “SWEEEEEEP HAAAAARD” at the top of their lungs, and the bottom of one of your shoes is covered in teflon . . . like I said, it’s a great sport).
The Super Broom started out as a casual summer project in the off-season – I decided I would build my own broom with a handle that would light up based on how hard you were sweeping (or maybe some combination of force and acceleration/frequency). I rigged up a quick prototype by re-purposing the previously mentioned curlometer and attaching it to my existing broom along with a meter of RGB LED tape. The accelerometer was fairly well suited to the rapid movements of sweeping, and all in all, it was a pretty satisfying effect – the only downside being that 1) there was no force measurement, which is really the most fun metric, and 2) having LED tape and a small pile of electronics dangling from my broom actually made it pretty hard to sweep effectively.
The need to actually preserve its functionality as a broom proved to be the downfall of this project. A broom is a deceptively simple device, but 1) it typically has a doubly-hinged head mechanism, allowing for a more or less arbitrary angle and rotation between the head and handle, and 2) even a mediocre curler can apply in excess of 100 lbs of force (and it turns out that anyone, presented with a light-up broom that measures their strength, will instantly apply as much force as they can to test it).
To handle the grip issue, I decided to go all out and build a transparent broom handle with the RGB LED tape embedded in it. I needed a material that was both transparent and strong (imagine applying a hundred pounds of force to the middle of a stick that’s braced on both ends to get an idea of the problem). , as well as light enough to run with for two hours. I settled on some ~1″ diameter, schedule-80 ‘clear PVC’ pipe from the good folks at U.S. Plastics (which happens to have a factory with attached store in my hometown). The outer diameter and internal volume were both suitable for the project, but it turned out to be far too flexible to actually use directly as a curling broom. Throwing the full might of my fabrication skills at it, I took a meter long piece of 15mm square aluminum extrusion, and machined it into a suitable ‘core’ to reinforce the plastic pipe. The aluminum had a flat surface to run the LED tape on, as well as a ‘routing channel’ for the power and control wires. Even with that, it was still way more flexible than my regular broom handle, but it was at least adequate for novelty purposes.
So I had a broom handle, but what I really wanted was to be able to measure the force I was applying while sweeping. I acquired a sturdy load cell from sparkfun, and then attempted to construct a 3D-printed ‘adapter’ to attach my normal broom pad to the handle, with the load cell positioned such that it could measure the strain applied to the pad. If the aluminum-reinforced plastic handle qualified as ‘mostly inadequate’, the 3D-printed adapter mechanism turned out to be ‘mostly useless.’ The first attempt made it as far as me saying “Hey, check it out!” to my spouse, before promptly snapping in half.
For the next attempt, I narrowed things down a little and made it an extremely heavy-duty, single-hinge adapter – still 3D-printed, but out of ABS with ~85% in-fill and quadruple-thick walls. This version actually worked well enough that I managed to play half of a drop-in league game with it, but it was heavy and didn’t work terrible well.
I finally re-designed things with an omni-directional adapter for the load cell, but after half a dozen iterations that were *all* promptly snapped by my fellow curlers when put to ice, I just gave up.
This was one of those projects that sounds extremely simple on the face of it, but to make it both look good and be very functional turned out to be beyond my fabrication skills. I couldn’t figure out a way to make a rigid, lightweight handle to hold the LED tape, I couldn’t figure out a way to adequately attach a load cell in a way that preserved the needed flexibility and could actually stand up to the forces even an amateur curler can readily apply, and I couldn’t even figure out a good way to attach the battery and electronics that didn’t apply an awkward weight to the end of the handle. This project just required a mechanical engineering and fabrication skillset that was beyond me, and I will, without shame, say that it was an utter failure. Lessons learned? Curling is harder than it looks, and apparently, so is broom design.