Hub Motor Design Simulations

Amen. My boosted is constantly scraping the ground (the motor mount is silver now, not black) and I’ve busted a handful of Kevlar reinforced belts. Maybe V2 solves this entirely.

I would like to cut down on the appearance. Belted speeds and efficiency is nice, but the motor assembly can look and feel like a car hitch ball sack thats dragging in the street. Certainly wrinkly and scabbed enough. It get caught in stuff all the time as well that the front truck doesn’t snag.

Exactly, the only solution would be reduce the gearing to have a smaller wheel pulley or reduce the pitch, tried both and them you go to a new set of problems, maybe getting a crazy low kV motor until I get this hub project going

One thing is great with HTD 3M, less belt drag, a lot less, the first time I tried the 5M it felt like something was wrong and braking the board

Seems wasteful to shoot that high and will result in a very heavy motor.

@anon94428844 did you manage to get the motorsolve interface, such as matlab working? If I manage to get my hands on a license my plan is to program a evolutive algorithm to optimize everything on the motor automatically, not needed to spend hours of worked trying to get better results when you can tell your pc to do that

The hard part on this aproach is to define the parameters and maximum and minimum boundaries to be optimized

1 Like

We’re dealing with two separate things.

  1. Design a kickass motor that stays cool under stress. Use MotorSolve (MotorSolve online is just as good tbh probably).

  2. Design a kickass ESC to power said motor. Use Matlab simulink to get this going (considering ditching sinusoidal waveforms because sin is better for AC I’m finding. Trapezoid is gonna be a little louder but it’ll have minimal torque ripples).

I haven’t connected them and wouldn’t want to. There are so many variables and it’s just not worth it. MotorSolve takes 3-5 minute per simulation on my server/desktop and 4-7 minutes of my top of the line MacBook Pro Retina. Don’t try to merge them, you’ve been warned. It’ll suck.

But through simulations will you find magic in another design different than what’s being used? You might get a bit here or there but the general rules of efficiency will apply and the size and weight of the motor will largely determine its output ability and efficiency is right around the corner

Let me know if you get that working, I’ve been wanting something like that for a while.

What kind of mesh settings are you using to get 3-7 min run times?

Thanks for the warning, I still believe that optimization for at least the bulk of the design is worth, I found this article to be enlightening http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.625.4767&rep=rep1&type=pdf

These finite element analysis take a lot into account. More than is really necessary to get into the ballpark.

Just plotting a Speed vs Torque graph takes 3 minutes to calculate every phase angle for each of the speeds (in increments of 100rpm from [1,3000])

Field Oriented Control and Direct Torque Control have historically been too close to prove superiority. However I would rather be able to simulate the heat before I wind it.

The VESC 6.0 will be a game changer with the best FOC setup we can get our hands on for small ESCs and vehicles.

http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.193.6219&rep=rep1&type=pdf

Yes, the gains could be minimal, but a lot of small gains could be significative in the final design, I still have to analyze the cost/benefit of this aproach

With motor not yet, by I’ve done simple problems just as proof of concept in one my courses, if you are interested search for Yang nature book, it has a lot of types of evolutive algorithms that mimics the behavior of groups of animals, crazy ideias, but work beautifully to get a maximum point in problem of non linear behavior and multiple dimensions, and the computational cost in relatively low

Back on topic, did you manage to understand in simple terms what exactly is this hysteresis control? I’ve a few more papers but could not understand in what way it is better than FOC or even BLDC, I talked a lot to the teacher that’s guinding me and we could not wrap out yet heads arround it

In looking for most efficiency include the manufacturing and cost limitation. what can the air gap be, what cost for tiny laminations of ideal steel you’re willing to spend, and if you really want the most efficient maybe add the cost of multiple small segment magnets to cut eddy currents, or the cost of going further with a halbach array or laminations in the flux ring, or skip using wire and instead formed copper sheets.

My point is there are a lot of known ways to get better efficiency but a lot easier and cheaper and more reliable direction might be to just go bigger. Not efficient per se but makes it a non-issue to a large degree

but id love to be more efficient and if I thought I could do much better using simulations I’d be all up and in it. I do respect it and know an efficient motor is effectively a more powerful motor, if only it were more easily achieved and it just needed some adjusting of number but Ido t see it and just follow the tried and true methods within limitations of practicality and expense

It’s direct torque control (DTC). Look at the wave forms and then also look at a DTC algorithm and compare it to an FOC algorithm. It’ll click.

For your consumption about the breakdown of motor control types we can use. The gold would be DTC-SVM.

I’ve put something like 100 hours in Novemeber/December in research for ESCs. The best thing you can do is study up. Google Scholar if you don’t have access to JStor or a comparable academic database collection.

Best mileage is achieved when going slow all the time. Wind resistance is the mileage killer No1. I built a 12S4P (Samsung 3000mAh 18650 cells) Longboard, using our Trampa 118kV Motor, 37/14 gearing. Set the VESC to max. ERPM to match 20km/h. Range was 60 Km. Its impossible to empty the pack in one go - feet start burning after one hour…Belts do wear but you could 3D print a cover for the belt to solve that issue. My belts (Strongbelt premium) last for more than a year.

Current Hub motor issues are in my eyes (compared to a 1 : 2.5 gearing):

  • very low efficiency (only 50-60%)
  • therefore heat issues (losses =heat)
  • therefore a lot less mileage
  • low torque (ca.2,5x less)
  • therefore the need for a twin setup (Two VESCs, two motors)
  • magnets are nearly impossible to cool (insulated by urethane and resting on bearings).
  • Thin urethane on a stiff core and therefore more problems with vibrations and less damping.
  • Thin Urethane will put more strain on the hub motor itself + your feet
  • servicing the wheel bearings is more complicated (customers will struggle to disassemble the motor)
  • not good for a hot summer environment and hills or heavy riders.
  • braking is less strong (no gearing)

Advantages of hubs:

  • better coasting (due to inefficient motor gearing)
  • stealthy look
  • cheap to make

Belt Systems advantages:

  • much much better efficiency (ca. 85%)
  • much more torque (2,5x)
  • ability to change gearing and find out best setup for your ride
  • Wheel bearings are easy to change and maintain
  • you can use different motors and experiment
  • motor is under less strain from vibrations
  • motor sits in the airflow and cools down easily
  • thick urethane tires prevent vibrations (ride comfort)
  • braking is much better (geared brakes)
  • A single, geared motor setup can perform just as good as a dual hub setup (cheaper to build and maintain, 1 VESC only)

Disadvantages:

  • less stealthy and less slick appearance
  • Belts and gears need a service once a while (hub motors will also want some attention btw.)
  • A transmission is more expensive to make (less profit if we compare a twin setup to a twin setup)
  • coasting is less good (I simply use the cruise control button to keep the speed)

Frank

2 Likes

This max 60% I don’t think it’s true, this picture I got from the Mellow engineer, its bad to look but its way more that than we have on the belted sistems, we all talk that belt is great efficiency but I’ve never seen anyone put a board in a dynamometer and run lots of speed and torque combinations to get this kind of data, thinks is something that has to be done, maybe you guys could do it, it will improve your product and make more scientifically to determine the optimum gear ratio

1 Like

3 Likes

http://vedder.se/2014/10/chosing-the-right-bldc-motor-and-battery-setup-for-an-electric-skateboard/

“Now we know that copper losses are proportional to the square of the torque produced by the motor, and at low RPM and high load they are dominant. As RPM increases, other losses start to add up exponentially. In my experience, these losses start to get significant around 60k electrical RPM, which for a 14-pole motor is about 8570 mechanical rpm (most 50mm+ outrunners have 14 poles, some unusual ones have 18). Because of the square relation, it is desirable to run at as high speed and low torque as possible as long as we stay below 8.6k RPM. To express the square relation in some numbers, having double the RPM and half the torque at a certain power output will cause four times less losses. The lesson from this is that: make sure the top speed you design the skateboard for is at around 8.6k rpm on the motor if you are using an 50mm-60mm outrunner.” quote Benjamin Vedder

Don’t get me wrong, I love the idea of hubs - but you have to face the reality of physics.
Hope someone will come up with a clever idea one day.

Frank

1 Like

I’m pretty sure that the first Trillium hub is going to be the most cosmetically controversial hub so far. In fact if everyone on here doesn’t like the cosmetics I’m probably not going to take it into full production. Per motor it will probably peak at 100N to ground (on 7S/30A with no load around 23mph) so certainly not for those people who are looking for that extra dose of adrenaline and don’t mind getting injured. Going for power density, better urethane mechanics, and better thermal performance than what is currently on the market.