Ebike week 8: Mixed success

In classic engineering fashion, I finally got all the bits together for a test of the initial prototype, and ran into:

  • a minor problem I should have thought of but didn't

  • a bigger problem I always suspected would happen

  • a completely new problem I didn't anticipate at all

So, the good news:

It works!

Yes, I know, portrait video once again - it fits the content and it's hard videoing while simultaneously riding a bike

All of the electricals are in place (except the pedal rotation sensor - shh, don't tell the rozzers) so I can press the button and it will turn the motor.

The bad news:

Problem 1: underpowered

When testing the motor on the bench with a mains PSU that can only deliver about 50W, I had the VESC current limit set to 3 amps to stop it tripping out from undervoltage. I went back into the settings and upped that to 10A normal, 30A peak, which let it use a proper amount of grunt from the battery. THe second test ride showed:

Problem 2: it's bendy

As I suspected, the thin 3d printed PLA motor bracket bends a lot under the side load put on it by the chain line. As soon as the motor starts pulling hard, it pulls the bracket out of square and the chain would derail almost immediately. I didn't get any video of this, but you can imagine. I can't reproduce it now because...

Problem 3: it's broke, yo

While mulling over how to support the jackshaft to fix problem #2, I took the bike for a ride with the motor attached to make sure the freewheel worked and wasn't too draggy. It started off fine, but over the 8 miles I rode it started sounding increasingly rattly. I finally diagnosed the problem about half a mile from home:

no bueno

Yep - 3mm PLA walls with sharp 90 degree corners and 20% infill don't stand up to a lot of road vibration with 1.5kg of metal hanging cantilevered off a couple of bolts. The middle of the frame cracked at a layer line, with the whole thing luckily staying just about held together by a support rib bridging to one of the not-yet-broken sides. Well, it was always a prototype. I've got a few ideas which put together should make the second revision a bit stronger:

  • thinner layer height (0.2mm vs. 0.3mm)
  • thicker walls (5mm vs. 3mm)
  • more infill (>50% vs. 20%)
  • radiused internal corners to avoid stress risers (fillet vs. no fillet)
  • more cross-bracing between different parts
  • newer, less crunchy filament that I've kept in a Ziploc bag full of dessicant since opening

If all of that doesn't work, I'll have to try printing one in stinky-but-less-brittle ABS. If this works well enough to want to use long term, I should really find a way to make it in aluminium, but that sounds like Proper Metal Work.

Now I think I'll go spend the next week reminding myself how Fusion 360 works, because the idea of trying to do radiused internal corners in Sketchup gives me the screaming horrors.