She’s alive! Alright, so my last post was a little dramatic but I’m feeling a lot better about the project now. A week later I ended up caving in and buying a replacement motor for the moped after realizing how long my walk to school this summer would be without it (and just how out of shape I am).
The motor came in pretty quick but before I could mount it I had to make a new motor mount because the last one melted. So this time I’m going to weld it out of aluminum to fix those annoying melting problems. It should also help to suck some heat out of the system and give me some more practice on the TIG welder. The main face I cut out at the Frank Innovation Zone using the waterjet to make sure all the holes are precise enough, the rest of the parts were hand cut because they just needed to be close.
The weld came out bad, like really bad. I can’t seem to get a hang of this TIG stuff. But the weld still looks plenty strong so I’m going to use it. I also designed an impeller to combat the heating problem I described in the last post. This one should suck air through the motor and cool off the windings (should be better than stopping every five minutes to blow on the motor).
After grinding a flat spot on the shaft of the motor to hold the sprocket in place(making sure to cover all openings with tape to make sure no metal shavings got inside), I bolted the motor to the mount, attached the fan, and bolted the whole thing to the moped frame. It worked perfectly as a straight drop-in for the old printed mount.
The speed controller I’m using is a cheap $20 E-bike controller from Aliexpress that was recommended on an endless sphere discussion board (great place for finding electric vehicle-related information). These cheap chinesium speed controllers don’t have too much documentation and most don’t even have labeled cables. After browsing a few forums and blogs I found Charles’s website where he has an awesome write up on his experiments with these Chinese controllers and a lot of documentation on some other really cool projects. It was really inspiring to see someone else out there, trying to document their projects like I am. His translated wiring diagram made it really easy to setup this controller although I still don’t know what some of the wires do but I was still able to get it working without them so they must not be important.
He found that the speed controller has both a sensorless mode and a sensored mode. If it detects a hall effect sensor it will use sensored mode for startup torque and then switch to the sensorless mode for higher rpm. If no sensor is detected it will just run in sensorless mode but will have trouble starting and will need a little push to get going. Under this reasoning and with the mopeds’ sensor still not here yet I set up the speed controller in sensorless mode. This did work but was very glitchy. What would happen is when the moped is first powered on I’d give it some throttle and the motor would make a grinding sound as if it were fighting itself for a second and then start spinning. If I let off the throttle and let the motor stop and then tried to give it throttle again the motor would again make a grinding sound but not be able to start spinning again. Sometimes it wouldn’t even start spinning the motor the first time. I found that if I reached down and spun the motor a bit while giving it throttle it would take off. I think this is due to the zero startup torque Charles was talking about with sensorless and because the moped has a clutch on the wheel allowing it to free coast without spinning the motor there was no way to jumpstart the motor by pushing the moped. At this point, I think the hall effect sensor should be all I need to get it working flawlessly.
I ordered the hall effect sensor from Jason at his web store and installed it just as I had done before with the electric bike and expected that to be it to get it working annnnnd I would be wrong. After adding a 3d printed mount to attach the sensor to the motor mount I installed the sensor but noticed the motor behaved oddly. There was definitely plenty of startup torque now but when the motor rpm got to a point where it would normally transition to sensorless it would instead shudder and make a grinding sound until you brought the rpm back down to sensored speed.
This was ridable but not at all fast or comfortable. The shuddering eventually got to the point where it would no longer even start. Initially, I chalked this up to the low voltage protection being too low and trying to stop the motor when the battery voltage sagged under load. After more testing, I found it acts the same with a fully charged battery as a dead battery disproving this theory. Then I thought the shuddering was due to the maximum e-rpm Charles found while trying to run a high KV motor with one of these speed controllers. I figured it got to a point where the motor was spinning too fast for the speed controller to catch up so it was turning on the motor at the wrong time causing the cogging. But after unplugging the sensor the motor returned to pre-sensor behavior. At this point, I was more confused than a Texan in a snowstorm as to what was going on. I hadn’t thought it was the sensor because startup was so smooth but after sliding the sensor over a little bit the cogging lessened. After fiddling with the sensor for a half hour I finally got it in the right spot. Startup torque was strong and it would smoothly transition to sensorless mode. Finally, I soldered the sensor to the speed controller and set it up in lower speed mode to try and avoid the max e-rpm issue. Wahoo all good to go now to see if it’s reliable. Update coming soon!