The problem is that it's sort of hard to tell how close I got. Everything was plugged in and in place, but nothing actually turned on, went beep or spun round. It's that sort of 'back to the drawing board' feel that I imagine makes scientific research so depressing - what bit of the large supply of things isn't working, and how do I deduce which one it is?
I'd spent the week preparing for it. I got the motor fitted with a tyre (from the excellent TJ's Tyres in Braddon), where it was opined that the fact that the spokes were bent where they attached to the wheel was a cause for concern. I got the awesome Laurie who runs Suspension Smith in Fyshwick to make me up some spacers and find an axle that would make sure that the motor fit in the middle of the rear suspension arms. I borrowed eight large UPS batteries in order to supply the 96 volts that is my aim for the battery pack's minimum potential. I spent the rest of Saturday going through my wiring and control box trying to get all the things that were supposed to connect together to do so correctly.
That bit is the real challenge. You see, there are five large and twenty-eight small connections from the motor controller; the large ones are the battery terminals in and the three motor coils out, the smaller ones do everything from turn a light on when the controller's 'on' to detect the position of the throttle to connections like 'boost' that aren't connected to a wire on the break-out cables and aren't documented in the manual. The motor itself has ten connectors (power for the three coils, hall effect sensors for each coil, ground and five volts, and the temperature sensor to and from). Some of the controller's outputs attach to external things - LEDs, variable resistors for throttle and brake, power supply, etc. which have to be put somewhere. So it's a bit bewildering working out how to connect it all together.
So on the day Rob and I fitted the brake and torque arm, fitted the spacers and axle (which were perfect!) and ran the cable. Then I lugged one of the batteries, the controller and the huge splay of leads out and started connecting things. I'd built a little kit box, used two D-15 ('VGA') connectors for the controller's inputs, and soldered all the wires to a circuit board. It's all reasonably neat but nothing like waterproof or robust enough to run - it's only really there to give me a chance to work out how the electrics all go together before I start actually finalising where everything goes. And the familiarity with its less-than-professional construction made me tentative as to how it was going to go. Would connecting it spin the motor up to maximum RPM? Had I something round the wrong way that would have its magic smoke escape?
With Rob holding on the front brake, I gingerly connected the battery to the controller. Nothing. I connected the mock-up box to the controller. Nothing. I switched the switch on. Nothing. Hmmmm.
I remembered that the controller required separate power and that was the only jack I hadn't soldered up. I did a quick soldering job and checked the polarity, and attached a spare motorcycle battery to the mock-up box's power input.
Nothing. Double hmmmm.
And that was more or less it for the day. We had to go out to lunch, Rob had to go and play hockey, and the next big task was to try and work out why all my clever soldering and wiring had produced a dud. The important lesson was that instead of one big box that everything plugged into, which was a pain to solder and required special connectors, I should have had each little subsystem in a separate box (or a different part of the same box) with its own connector. That way, when something breaks you can narrow it down fairly easily, and you don't have so much that's interconnected and jammed together. The Kelly jumper leads bundle the leads for the same functions - throttle, brake, LEDs, etc. - together, so you can have a separate box for each.
Also, buying an extra set of connectors isn't necessary, since Kelly supply you with a set and they're push-on pull-off 'quick' connectors so they can be taken apart again, but it is very helpful as it means you've got a test set if you botch something completely.
All posts licensed under the CC-BY-NC license. Author Paul Wayper.