[unclematt]

Hi All,
There was another thread where we have discussed the need for an open source wheel motor. The wheel motors that are currently in the media are not available to the public unless you want to spend obscene amounts of money, if even then. So I think it would be great for all of us to collaborate on creating an open source set of plans to enable EVers to create their own wheel motors.

So lets start the show:

1) We need a list of objectives to meet with our wheel motor.

2) We need sources for raw materials and components.

3) We also need to find or design controllers to run our wheel motor.

So please offer links and expertise to accomplish this.

I personally think we need a brushless motor, and it needs to put out at least 40kW of power. It also needs regenerative capability, and if possible I would like to do away with friction brakes entirely.

Let the creativity begin!

:-D

 

[JRK5150]

I agree, open source is the best way to go for us all! Designing wheel motors should be fairly straight forward. The real issue is in designing an appropriate controller that can act as a controller as well as keeping all of the motors moving EXACTLY in accordance with the others. I suggest we expand this to an open source wheel motor drive system with emphasis on a safe controller. Since we would be designing a controller from scratch anyway, I feel AC would meet the goals better than DC. With AC we could have a brushless system with regenerative braking, ABS, traction control and it would be more efficient. May I further suggest that we focus on writing the control program first (test it with existing hardware), then focus on building our open source hardware. I think this is a great idea and would like to help out any way I can!

-John

 

[CNCRouterman]

I like the open source concept, though I think making it more universal is more appealing.
Wheel Hub motors are, by definition, in the wheel... So, that would mean that the weight of the wheel will be greater than a conventional wheel design. The higher the capacity of the motor, the more it is likely to weigh.

Given the above, a wheel hub motor would seem more applicable to low speed use, or applications where suspensions are a secondary consideration, due to the inherently higher un-sprung weight.

I agree than defining the objectives of the motor are important at the start.

I think a wheel hub motor would be well suited for applications such as garden tractors, material haulers, electric mule, forklift, etc. 40kw would be overkill for most of these applications. However, I have read on the website of one manufacturer of wheel hub motors, that they can be mounted inboard of the wheel, and couple to the wheel though a drive axle (one motor, one wheel). That makes sense to me in terms of keeping unsprung weight down, even using one to drive a conventional auto RWD axle, keeping the differential, or two for 4WD.

I suggest keeping conventional mechanical brakes because at some rpm point, it will take energy to slow it anymore, and it will definitely require energy to hold position against an outside load, so for full stop or parking, conventional brake are hard to beat.

With regards to the controller. I am very interested in learning more about design and building a controller for my future EV project. I have found that there are many electrical CAD packages out there, I found one called Dipfree 1.5 which has a freeware version. As I am pretty ignorant about integrated circuits at this point, I don't know if it will help in design and understanding. I am quite familiar with CAD/CAM software for use in CNC routing, so I know that mere possession of a tool does not equate to proficiency with the tool. I do hope, however, that possession of this electric design tool will help me learn. I have been watching another thread on this board regarding an AC 3 phase controller, but it is still mostly over my head technically.

 

[thesilverring]

Does anybody have an IEEE membership? If so you could view this article. I actually went to Bath Uni and went to this guy's lectures - but at the time I was not so interested. Going to uni when you are 18 years old to actually learn the material in a degree course is not the best idea.

This is the sort of thing you are talking about right?
http://ieeexplore.ieee.org/xpl/freeabs_all.jsp?tp=&arnumber=497248&isnumber=10700

A couple of the guys working on this stuff were decent chaps - Prof Dave Rodger (Scottish guy who looks more like Billy Connolly than Billy Connolly). Paul (?) Leonard and some other chap who's name escapes me right now. Anyway, they developed some Finite Element Analysis software to model electro-magnetics - specifically for motors.

About all I remember from the Motor design courses are Gaussian beer cans, the model Doc Rogers used to teach field theory.

 

[Manntis]

Does anybody have an IEEE membership? If so you could view this article. I actually went to Bath Uni and went to this guy's lectures - but at the time I was not so interested. Going to uni when you are 18 years old to actually learn the material in a degree course is not the best idea.

This is the sort of thing you are talking about right?
http://ieeexplore.ieee.org/xpl/freeabs_all.jsp?tp=&arnumber=497248&isnumber=10700

Good lead - and an interesting paper Unfortunately such an axial flux motor, IMHO, adds sprung weight in the form of the greater diameter of fixed magnets and the diagrams seem to underestimate the amount of travel required by the rotating mass in suspension compression and suspension extension conditions. Covering the complete range of motion would necessitate placing the bottom of the (rather expensive) magnet housing assembly very close to the road and because of its hard mounting position it would not be able to softly deflect from road debris encountered while in suspension compression. The potential for damage could be very high indeed.

 

[unclematt]

So how will we decide which basic design is best to follow? What power output is being targeted would be good to figure out as well.

If people could do a mini-presentation on the design they are promoting, that would help the rest of us get up to speed on their proposal. Links to photos/articles and text should do fine.

Question: should the controller be able to fit in the stator of the motor? If space is left for this in the center of the motor, will that cause problem with flux distribution or saturation?

Some informative links:

http://groups.yahoo.com/group/lrk-torquemax/

http://www.wavecrestlabs.com/automotive-products-wheelmotor.php

http://www.youtube.com/watch?v=8tLQ2-yKT4Y

 

[judebert]

I always thought it would be interesting to put magnets in the rotor, and outfit the calipers with windings. You could get propulsion, regen, and friction braking all in the same place.

 

[Manntis]

I always thought it would be interesting to put magnets in the rotor, and outfit the calipers with windings. You could get propulsion, regen, and friction braking all in the same place.

brake dust would be a problem, as would rotor wear and the necessity of keeping the rotors in an open airstream for cooling.

Easier to make a conventional wheel motor with a separate brake rotor for mechanical braking. Or make the rim itself the motor spinning on a fixed hub.

 

[John]

The hub/wheel motor would be very difficult to do and do right. Getting it to fit with suspension and brakes would possibly require adaption of those designs as well as the wheel and motor. Each different vehicle it was adapted to would require a redesign. While the hub motor fits with the ideal of minimum mechanical complexity it doesn't fit with other ideals such as minimising unsprung weight. It will also change the dive/rise characteristic of the suspension geometry by applying the motor torque to the suspension rather than the vehicle body, and require building the motor to with stand high shock loading while being light weight. For a retro fit conversion unit I think a dual motor direct drive diff replacement for an AWD donor vehicle would be easier to achieve and wouldn't have the unsprung weight issue. An electrically commutated permanent magnet brushless DC outrunner with a large number of poles would be the configuration I would choose. The permanent magnets (on the rotor) would be arranged in a halbach array to minimise external magnetism, intensify internal flux density, and negate the flux ring. I wonder if an iron less core would be feasible to prevent cogging. Would this make the motor more or less efficient?

 

[JRK5150]

I think we should first look at the current cutting edge of technology in wheel motor design. If you don't already know about PML and their Hi-Pa drive wheel motors, check them out here.http://www.pmlflightlink.com/motors/hipa_drive.html

Next, I think we should make sure that we do NOT infringe on any patents. To this end, I suggest keeping this as simple as possible. We can get fancy later. Getting sued developing a free drive system would severely suck!

For simplicities sake, we should probably decide on a standard size and basic mounting bracket arrangement. I think unclematt's suggestion of 40kws per motor seems about right.

On the topic of controller design, I suggest we split duties between a main controller and separate wheel controllers.
-The main controller would rectify dc to ac (vice versa for regen. braking) and send basic commands to the wheels.
-The wheel motor controllers would determine the final frequency and amplitude.

The driver input, ie. throttle control, steering, braking, would go to the main controller which would send out the basic signal to the wheel motor controllers. The wheel motor controller would take the main signal and modify it based on input it gathers from it's wheel and the other wheels on the vehicle.

Finally, what sort of open source programs are currently available to accomplish our goal? And, what sort of hardware would we run it on?

Well, what do you think about this for nailing down the basics?

 

[John]

This could be very cool. Independent torque controlled high powered all wheel drive and extreme mechanical simplicity. It could be a huge leap forward in terms of vehicle dynamics. It could do full regen with proper brake bias not just fore and aft but also left and right if braking while cornering. It could do active yaw control with the drive assisting the turning of the vehicle possibly even using a combination of regenerative braking and drive on the inside and outside wheels to aid vehicle turning. The computer could dial down the torque to any wheel when excessive slip is detected. All manner of new things could be done that aren't currently possible with today’s cars. It could put the electric car firmly out in front in terms of drivability and nimbleness and surefooted safety.

 

[JRK5150]

Yeah, this is cool! Traditionally, converting a car to electricity has been a bit of a compromise, but this could actually make the cars perform better in every way. Also, with no engine or transmission, vehicles will be left with a large, empty engine bay that could be filled with a generator... extra batteries... fuel cell... the options are endless!

 

[mattW]

It would also make a conversion to a series hybrid ridiculously easy. Just swap the wheels and add a battery pack/controller while keeping the existing engine (or downgrading the size). It could be fitted pretty easily to just about any vehicle with the space for batteries.

The only problem would be designing the motors for such low rpm and sensing/controlling the 4 wheels at once.

 

[Dennis]

To make such a motor you would need the following:

• Vacuum Pressure Impregnation System (VPI)

• Oven

• High temp Resin

• Metal press with dies needed to cut out the stator and rotor steel sheet laminates

• Thin steel sheets on large rolls

• Enamel coated copper wire

• Furnace for melting copper or Aluminum

• Special metal casting machine to cast the rotor bars of either aluminum or Copper into the rotor...

• Nomex paper

• Coil winding machine (unless you want to do it by hand)

• Induction heating device to heat the inner race of the bearings such that it expands to fit over shafts.

• Test equipment such as insulation testers...

This undertaking is not exactly what I call easy. Maybe you can outsource the rotor design and stator design?

 

[JRK5150]

Anyone else in the central Florida area? I think I'd like to start developing this into a working concept. I don't have a problem with fabrication, so my plan is to get a small, fairly aerodynamic piece of junk and mount conventional AC motors at each wheel. We could use this to develop the controller setup. After we have the controller designed and programmed, we could post free plans and software online for everyone to use. Then we could work on developing the actual motor. I'm looking around locally as well, to try and find others to help.

 

[JRK5150]

Would developing this in stages be the best idea? Rather than jumping head first into the entire project with all of it's complexities, maybe the best idea would be to break it down into steps.
For instance, step one could be to mount two conventional AC motors to the rear wheels of a front wheel drive car. This could be used to develop the control system without the extra complexity of four wheels or the need for complex algorithms for steering wheel motors. It would allow for the main controller to be developed as well as the basic programing for the wheel motor controller.
Step two could be designing and building the actual wheel motors. With the basic control system in place and wheel motors built, we would have a simple 2wd conversion for light cars. This would also be nearly a bolt on hybrid setup for most front wheel drive cars.
Step three could be to add the front wheel motors and work out the added complexities of four motors working together as well as steering inputs. This would open the door for heavier, all electric conversions as well as true hybridization.
This seems to be the quickest way to having this technology on our cars. Does this make sense or am I just over thinking it?

 

[Wirecutter]

The only problem would be designing the motors for such low rpm and sensing/controlling the 4 wheels at once.

I have to agree with Manntis about the unsprung weight issue. A wheel-hub motor would be fine for something like a forklift or some kind of industrial vehicle where weight isn't as much an issue, and the "road" surface is very smooth. What I always thought of when I considered a "4 wheel - 4 motor" vehicle is having the motors mounted inboard and shafted to the wheels.

Back to mattw's comment, though - I read that this sensing and controlling of motor speed is one of the things that makes the Wrightspeed X1 so stinkin' fast. The AC Propulsion controller used in the X1 is modified to provide traction control. It will sense wheel speed of the driven and undriven wheels, and allow the driven wheels to spin only slightly faster than the undriven wheels. The effects of limiting the amount of slip allowed between the tire and the road surface is already well proven with ABS. I would imagine the algorithm gets slightly more complicated when dealing with turning and attempting to power slide, because at such times, front and rear wheels need to turn at different speeds. But I would also imagine that the performance of such a car would feel absolutely marvelous.

Ok, well it's a cool idea, but I admit a little problematic for a vehicle with all wheels being driven. You'd have to come up with another way of measuring the ground speed of the vehicle.

-Mark

 

[mattW]

Have you guys seen this:
Siemens VDO eCorner Project
That wheel does just about everything!

 

[unclematt]

I have to agree with Manntis about the unsprung weight issue. A wheel-hub motor would be fine for something like a forklift or some kind of industrial vehicle where weight isn't as much an issue, and the "road" surface is very smooth. What I always thought of when I considered a "4 wheel - 4 motor" vehicle is having the motors mounted inboard and shafted to the wheels.

Back to mattw's comment, though - I read that this sensing and controlling of motor speed is one of the things that makes the Wrightspeed X1 so stinkin' fast. The AC Propulsion controller used in the X1 is modified to provide traction control. It will sense wheel speed of the driven and undriven wheels, and allow the driven wheels to spin only slightly faster than the undriven wheels. The effects of limiting the amount of slip allowed between the tire and the road surface is already well proven with ABS. I would imagine the algorithm gets slightly more complicated when dealing with turning and attempting to power slide, because at such times, front and rear wheels need to turn at different speeds. But I would also imagine that the performance of such a car would feel absolutely marvelous.

Ok, well it's a cool idea, but I admit a little problematic for a vehicle with all wheels being driven. You'd have to come up with another way of measuring the ground speed of the vehicle.

-Mark

The whole "unsprung weight" issue is overblown in my opinion. Yes, we should design them as light as possible, but in my research I have found that most manufacturers found that when much of the weight is removed that is associated with the regualr drivetrain that the wheel motor ends up weighing only slightly more by comparison. I am NOT interested in mounting motors to shafts. That defeats part of the purpose of doing wheel motors in the first place.

 

[unclematt]

Have you guys seen this:
Siemens VDO eCorner Project
That wheel does just about everything!

Their concept is a good one, but they still use friction brakes, which is unneccesary and weighty. They also do not offer them to the public at any price, so they are vaporware for the moment.