Recently I made the power unit for my new longboard. This board is my fourth skateboard, Version 9 between Aaron and I. This blog post spans almost a year and you can see my engineering skills improve, lol
The completed drive unit on the board:
The power unit:
This board was the result of almost a year of ideas. The general desire moving from the Version 7 hubmotor longboard was to have something lighter.
While doing motor control experiments this summer, I really got a fundamental understanding of how motors work and how to use them. One fact I grasped was that mechanical power is speed*torque, but that most motors don’t really care how fast they are spinning. This is a very important fact to understand- it turns out that by just spinning the motor 2x as fast you effectively double the power of the motor without making it any less happy. What makes a motor unhappy is when you feed it too many amps, as heat is generated by I^2*R. Notice that this equation does not depend at all on volts. More amps = motor heats up and gets unhappy, but more volts = motor spins faster, no extra charge.
I realized that most longboards spin their motors way too slow to extract full mechanical power from the motor, instead using a huge motor to compensate. Big motors are heavy. The extreme example of this is the hubmotor, spinning at a 1:1 gear ratio with the wheel, just about the worst possible setup with respect to mechanical power output. On Kimmy, the hoverboard motor weighed at least 5 lbs, almost half the weight of the vehicle. On the version 7 hubmotor longboard the hubmotor weighed about 2 pounds, a lot lighter but still pretty heavy. Belted longboards are a lot better, but still cannot achieve enough reduction to fully harness the power of the motor. With this problem at the forefront of my mind, I began to design the electric longboard version 9.
V9.1: This was designed in September 2016, back to the roots, A simple belt drive. I got lazy and decided to just try out something simple. But this one is so overplayed… Also you can really only get a 3:1 ratio with just a simple belt drive in this form factor, and thus you really need a big motor with lots of torque.
I wrote a python script to calculate the board speed and motor speeds. On a simple belt drive like this, the motor would need to spin about 9000 rpm to hit 20mph. I tentatively estimated that my NTM motor was good for about 20,000 rpm, meaning this setup would utilize under 50% of the available motor power.
V9.3: how about a direct gearing setup?
This one was playing with fire. Rocks would probably immediately destroy these gears as they are open, and I still only got a 3.1:1 gear ratio. Could have gotten better by decreasing the pitch but then the gears would be even more fragile.
I decided to ask “how bad can 3:1 be?” on V9.4:
I went for maximum laziness and 3d printed this bracket.
Taped on a 4s and an airplane ESC:
Time for a test run. It was able to move and took me around the block. However it was pretty underpowered and the motor got super hot. Unsurprisingly the 3d printed bracket exploded after about 3 minutes.
Back to the drawing board.
I designed this mount with the intention of CNCing it. It is made of two adjustable pieces with bolts sticking up from the top and bottom.
I made extensive use of 3d printing during the design of this truck. I 3d printed all of the parts to see if they would fit. The parts were a little complicated because they had to line up with the weird cast features of the truck. The cast features prevent rotation.
And in aluminum!
I then 3d printed the other part and tried to see if it would fit, which it did:
Of course it was all put together and I couldn’t resist a good test ride. Unsurprisingly the plastic bit exploded after about 20 minutes. Also, it was still incredibly underpowered. At the end of March 2017, I put this project on the shelf for a while while I played with motor controllers.
End of summer 2017: Time to restart this project.
The summer bought me plenty of time to think. I realized a nice form factor would be a small gear reduction and then a belt drive. Ben had some leftover 0.8 module 12 and 24 tooth gear stock from his mini-cheetah which he gave to me to use. I cadded up a gear drive using these parts.
The gears are selected.
The 1/4″ shaft which holds the gears is double supported with two sealed bearings. The 12t pulley will mount on this shaft.
The motor mounts to a second plate on the back, which seals the mechanism.
Time to build.
The gears were parted off and attached to the shafts with loctite.
I CNCed the parts on the dyna.
I CNCed a 12 tooth HTD pulley as well. Looks nice!
Gears greased up:
Electronics attached with tape. Sort of ghetto, but good for a test run.
Results: It rides good! A good amount of power. And a nice gear noise too. Up next: Building a controller for it!