I think the problem was that since the chain tensioner was fixed relative to the body of the scooter, rather than the rear wheel, when the wheel moved when going over bumps, the chain tensioner could easily come out of alignment with the back sprocket, causing the chain to slip off. Also, the chain tensioner was made from a long piece of thin aluminum, making it even more susceptible to flexing out of alignment. Solution: Build a compact, physically robust chain tensioner mounted to the suspension swingarm.
With that (questionably rationalized) solution in mind, I made this:
First, instead of being made from modified 1/8" aluminum angle, this new tensioner is made from short, wide pieces of 1/4" aluminum. I got the metal from some iMac stands I happened to have lying around, which is why the larger surfaces of the metal are anodized.
The bolt that holds the right side of the spring was screwed directly into the 1/4" edge of the aluminum. You can just barely see the peaks of the screw's threads on the side of the metal, since the screw only has about 1/64" of metal on each side.
I would have preferred to have the smaller sprocket at the end of the swinging part of the chain tensioner, as it would respond to bumps more quickly due to it's smaller mass, but the geometry did not work out.
A second improvement over previous chain tensioner versions is that the pivot point actually pivots around a bearing(!). Since I don't have the proper tools for either boring a press fit hole for the bearing or actually pressing the bearing in place, I used the same method I did for the tensioner's small sprocket: I drilled out a 1/2" hole, which is the largest drill bit size I have, and then carefully filed the whole to the 7/8" O.D. of the bearing. I could not apply enough force to press the bearing in place with my wooden vice table, so I carefully hammered the bearing in, with an anvil (yes, I have an anvil but don't have a proper vice) on one side, and a block of aluminum on the other, so that I wouldn't destroy the bearing's shielding.
The combination of actually using bearings and using a lot more metal makes the whole assembly feel extremely solid, especially compared to the previous version. To mount the chain tensioner to the swingarm, I bolted it through two 1.5" x 1.5" squares of 1/4" aluminum. Where it's mounted, there's a full inch of solid aluminum.
I kept the part of the previous chain tensioner assembly that stopped the chain from coming off the drive sprocket, as it did it's job properly.
I added an assortment of stickers to the steering column. From top to bottom, the are:
- Alot (is a lot better than you at everything)
- Atlanta Mini Maker Faire
This chain tensioner was a drastic improvement over the previous one. I managed to ride two lengths of the trail that I had previously never been able to ride all of without losing the chain. However, riding on the field by my old elementary school, the chain tensioner managed to move more than it had been designed to, so it horribly overstretched its spring. When the tensioner stretched past horizontal, the spring became caught on the corner of the tensioner, which is why only about a quarter of the spring was actually ruined. Fortunately, this is an easy problem to fix. All I need to do is cut off and round the corner, so that the chain tensioner can bend farther without ruining the spring. I know that if the entire spring rather than just a short section of it is stretched, the spring can extend plenty far enough to accommodate the extra chain length.
This latest failure was also a testament to the strength of the rebuilt brake. When the spring was overstretched, the chain got wedged bewtween the brake housing and the sprocket on the rear wheel, instantly locking up the wheel. If the brake housing can withstand forces high enough to lock up the wheel while on damp grass moving 15-20 mph, it should be sturdy enough to slow me down in a more controlled situation.
Finally, here's five minutes of me tooling around on the scooter. I attached the camera to the steering column, rather than my helmet, so the video is pretty bumpy, even after a bunch of post filming software stabilization.