Hub Motor Design Simulations

Just building the market up for us. lol

80mm wheel or? I wouldn’t expect many complaints.

I saw that something was not adding up, the graph was too different than the previous one, if in the future after the release of Mellow you could release your hub efficiency graph, that would be great

Mellow should not be left behind a lot according to this data, maybe even pass when speed catches up

@Trillium 80mm wheel.

We’ve used the dyno to do for pre-builts what @anon94428844 is doing with you all here. We’ve done a full workup on efficiency, torque, power, etc for Carbon GT, Boosted Dual+, Onan X-1, Stary and an Evolve Pintail, and are very happy where we stand. @Pedrodemio As you have surmised: we do catch up to the Dual+ eventually. Our torque curves cross each other at 13 km/h. Hubs can and will compete with the belt drives performance-wise. We’ll release the charts in the near future.


@Mellow Thank you for the support. In a product oriented community, its not often that we get support in development/pioneering a project, since its easier for people to criticize and throw money at problems.

So guys, its clear that we need a half decent bench testing solution we can all reference. Laser Cut aluminum or CNCed aluminum frame dynamometer. We need a community effort to build one for BLDC testing.

Boosted’s in house Dynamometer from their early days:

Preliminary VESC Based Dynamometer (DATED, but useful):

With the VESC 6.0, we can have higher resolution data from the motors we use in testing. While we wait, lets consider the best way to test motors by considering how to BUILD a rig for hub motors. We could build a custom can/rotor that connects to the shaft, but thats additional milling for every motor. A Wheel-to-Wheel test might work, but we have mechanical losses. I’m pretty good at AutoCAD and Solidworks, but I need someone to do the mechanical engineering here to reduce losses. Anyone with me? If so, PM me. Otherwise, I’ll create a new thread about a community driven Dynamometer.


I’m in but let’s keep all open, and I can help on the mechanical part

In my opinion the better aproach to be cheap and easy and to make is to use friction, a bike disc brake and caliper, a load cell taken out os a cheap scale or bought, an arduino and we are done The vesc send info to the arduino, same for the load cell, with that info we can make a simple mechanism to apply brake, that can be manual or servo driven, on the latter we can make a routine to automatically generate a torque/current/efficiency curve

1 Like

That’s a great start. I like the idea of using a disk break instead of another BLDC.

I’m going to get some basic information together and start a thread tonight.

1 Like

Nice, we continue the discussion there to not go off topic

I think a generator would be better. It will be hard to use a disc brake as sort of standard. You want a well defined and measurable resistance. You need to know the output power to measure the efficiency. If you know your generator it’s easy to straights out the numbers. Define a cheap generator everyone has access to and standard is set. Mechanics should be easy to sort out. Some alloy and wood…


The programming will be quite simple, seen in the image bellow, we just have to test how stable is the torque using the brake

The test will be completely automatic and will generate in one run torque vs speed vs efficiency

I’m just a bit worried if the brake disc will end the test on a acceptable temperature, I think a big fan pointed at it will be needed and incorporate cooling time and stabilization time, or in another aproach, make each measurement so quick that it don’t have time to heat up, Just have to check if the data is stable in short periods or if we will have to round and filter it over a period of time

The disc brake doesn’t matter, what we need is the torque that it generates, so the load cell matters, but that can be easily calibrated using a portable dynamometer, the ones used to tighten bolts

Would you mind to share the excel data for the competitors to you and not yours? It’s not critical to my development but would be nice to have.

This is a community driven process. We’ll have some form of database I think.

Heads up, I have created the thread regarding Dynamometer to properly split the topic. Look for it under innovations.

Is the stator on your website the same one that’s in your motor or is that just a dummy stator with unrealistic pole count and such you used to camouflage your design?

We’ll share studies on all we tested, including ours, in the near future.

The stator is our actual one. Check out the Mellow subforum here, or the MellowBoards subreddit on Reddit for more.

What is the steady state temp of your motor cover / hanger?

guys, kudos for this thread

In my opinion this is one of the most informative, valuable and civilized discussions on this forum.



It’s a study of science by engineers and design practitioners. With any luck, we can pioneer something great.

1 Like

Progress Update

This is how a hub motor should behave. It’s only taken a few months of simulation to get here. With a 4WD setup, less heat is generated per motor when starting so the temperatures tend to stay pretty cool. Practically, the thermal results indicate you wouldn’t want to push more than 1.5 hours of riding time and definitely half that if you’re going up hills constantly. Two motors aren’t strong enough and wouldn’t stay cool under those conditions with 100kg of human… Looking at the data, I wonder how close these results are to the specs from the Action Blink Quatro.

Easy to replace stators/inexpensive recycling or rewinding would be a victory.

1 Like