Page 1 of 1

1st Post, 2nd Conversion: Coleman youth ATV

Posted: Aug 19 2022 7:49am
by mtbkrdave
Hi E-S! Looking forward to joining this community and to the projects to come…

My first conversion was electrifying an old Trek mountain bike with a 24-inch wheel kit, controller, and battery off eBay. Worked well as a commuter for a couple seasons, but the already-failing headset was getting too sketchy to ride.

Recently my older son (9) has gotten interested in ATVing. Wanting it to be electric, and for him to be able to work on it, I offered to build one with him. We visited a local Tractor Supply to take measurements from a Coleman AT125-UT ( youth ATV they had in stock. The manager stopped over and told us this particular one had never run; we ended up buying it for half price and free delivery. The final cost was less than I would’ve spent on aliexpress for the front/rear suspension, steering, wheels, and a pile of aluminum extrusion to build a frame.
3D149140-05B8-4475-92CD-7AE027335B28.jpeg (1.71 MiB) Viewed 363 times
After diagnosing and bypassing a bad ignition switch, the 120cc engine started on first-crank. So he’s having fun with it for now while I spec-out and price-out an electric drivetrain upgrade.

So far, I’m looking at various combinations of motor and controller, and the ATV-conversion threads on here have been super helpful!
  • The 120cc ICE engine is spec’d at 8Nm of torque, no power spec given
  • The vehice’s dry weight with ICE engine is 126kg. Kiddos are in the 18-25kg range now, but mom and I would like to take it for a spin now and then too.
  • Looking at various BLDC motors in the 2-5kW power range with torque ratings from 7-10Nm
  • Looking at VESC controllers and some of the other kits mentioned here: QS, Golden motor, Aliexpress unbranded…
  • Trying to figure out how to mate up with the existing driveline - it’s an integrated motor/gearbox, so I’m thinking a gear motor (like QS 3kW with integral gear reduction) would be easiest to direct-drive, or fall back to a similar-sized BLDC with an off-the-shelf jack shaft.
Suggestions on motor/controller choice more than welcome! Budget it not super-constrained, but I also don’t want to get way upside-down on value of electric drivetrain vs. the overall vehicle (noting, of course, that he and his little brother will eventually outgrow this, and I’ll be able to repurpose the bits for something else).

I’ll post some more before pics/details here soon, and once I (/we?) choose and source some parts, I’ll post tons of build detail… looking forward to diving in!


Re: 1st Post, 2nd Conversion: Coleman youth ATV

Posted: Sep 05 2022 5:36pm
by mtbkrdave
Update: We've done some benchmarking, and I think I'm narrowing down on a component set...

I mounted a GoPro to the quad, watching the dashboard. We did full-throttle-to-top-speed runs with my son (30kgs) and me (90kgs) riding. Knowing mass and [piecewise average] acceleration, I could back out the force, then extract torque given wheel radius. I was obv the worst-case:
120cc Engine Performance (1).png
120cc Engine Performance (1).png (71.28 KiB) Viewed 229 times
I stepped through the video frame-by-frame and grabbed the timestamps each time the speedometer changed. While delta-V and delta-T got me piecewise acceleration, I suspect the speedo has a fixed >=0.5s update rate, resulting in the choppy output - we didn't really reach those MPH thresholds right when the speedo updated. I figured a polynomial curve fit would smooth that out reasonably.

We got a max force of about 1200N (276 lbf) for me, which seems reasonable to accelerate 215ish-kg of machine and rider. That translates to a max wheel torque of 281Nm. We worked back through the ratios: first wheel-size-to-sprocket-size, then driveshaft-sprocket-to-gearbox-sprocket, then gearbox-to-engine.

We got the last one by also noting the engine tach RPMs at each speedo update. That got us a gearbox ratio of 0.12 to 0.13 (neglecting lower revs when the centrifugal clutch was [presumably] still slipping). Working our way through, we got a max of 123Nm at the rear sprocket, 60Nm at the gearbox output, and 7.62Nm at the engine output. Lines up well with the motor's 7Nm spec.

I'm looking at using a QSMotor 138 3 phase internally-reduced BLDC, which gets us about 0.65Nm/Amp of torque:

We're limited to 100A peak by our batteries (Headway 38120 x 20, 8 of which a coworker graciously donated), so the max motor torque is about 65Nm. That works well with the 60Nm max gearbox torque we extrapolated (lets say 65Nm assuming some loss to the chain, tire deformation, and slip), but what about speed range?

With a 1:1 gearing, the motor's 2955RPM redline would put us at.......158MPH!! That won't do. Switching the motor to an 11t sprocket and fitting the axle with a 50t sprocket, we'd end up with a top speed of 34MPH (likely slower with lower 20S voltage) and an insane wheel torque of 678Nm (only 2.5x the max we measured in testing). So we'll need a current limit on the motor to avoid flipping over backwards, but that's better than being underpowered, and lower peak discharge rates means longer battery pack cycle life.

I was glad to see my recollection of Physics 1 got us to a point where observed operation roughly correlated with the ICE engine's specs, gives me a bit more confidence selecting the electric motor (even if I'm erring toward over-powered - who knows, this powertrain might end up repurposed in a Vespa for my wife, or a motorcycle for me someday).