Development of the

Electric Hubcap Motor Family

and BLDC Motor Controller

The Concept
95% Peak Efficiency!
Evolving Mechanical Design
The Electric Hubcap Motor Family: Electric Caik -  Electric Hubcap - Electric Weel -  Bicycle Wheel Motor
The Turquoise Motor Controller


   The Electric Hubcap drive systems were born with the concept that supermagnets had such strong magnetic fields that surely it should be possible to make a simple, lightweight "pancake" motor that could turn a car wheel directly to propel the car. This would bypass the regular inefficient automotive transmission and drive train with their typical 30-40% losses of power between the engine and the wheels, so it would take 2/3 as much power to propel the car, 2/3 as much battery substance to drive the same distance, and 2/3 of the energy from the electrical grid to recharge the car.
   Then I thought of mounting this motor on the outside of the wheel, where it would merely be an "add-on" accessory, and the car could still be run on its regular fuel. That way, there only had to be enough battery power for "typical" driving, rather than for a maximum range limit, "untypical" further distance being covered with the original engine rather the electric motor. Thus the concept of "hybridizing" a regular car came into being.
   This was a great idea in theory as far as the above reasoning went, but a critical part doesn't work. The supermagnet fields act against the electromagnet fields, and the electromagnet fields are the same in any motor, supermagnets or not. A pancake motor with enough torque to turn a car wheel directly would be too big in diameter to fit on the wheel. A motor of smaller diameter that wouldn't hit the ground would have to be ten times more powerful than it would otherwise need to be in order to get sufficient torque for direct drive, and hence it would be too long and too heavy to fit on the wheel - and it would need more power to run.

   But everything was rather vague and such details weren't clear to me when I began. So I hopefully made a "motor" with the supermagnets mounted directly on a plate on the car wheel, thus using the wheel and axle as the motor's rotor and axle. The stator with the coils mounted in front of that to make a "BLDC" motor -- which term I hadn't run across yet, making web searches for information unfruitful. The motor stator simply bounced around. I finally worked out that 12 alternating magnets with 9 coils had zero net thrust: every forward force was matched by a backward force. It couldn't be made the same as the generator! It turns out that for a three-phase motor there must be two magnet poles for every three coils. Period. I did some gluing and ungluing of magnets to rotors in those times, soaking the rotors in a solution that eroded away magnets as well as epoxy.
   Once it had a working axle to couple the stator more solidly to the wheel - a "lazy Susan" bearing, and the magnets were rearranged to a workable configuration, it turned. But if the load got too high for the frequency drive - I had made the wrong type of motor controller for this motor - it simply lost all power and stalled.
   After creating a more suitable stator and inventing for myself a workable BLDC motor controller system that turned out to be the way everybody else already does it, I had a wheel motor with some power behind it.
But an unexpected characteristic of the motors was that they worked mch better with a gap of over 1/2 an inch between the coils and the supermagnets, which meant the powerful fields of the supermagnets acting with a small gap weren't available, a major factor in substantially decreasing the torque over what I had originally guestimated in my mind. Later I found other axial flux motors of similar configurations have similar large gaps.
   In October 2008 I actually got the car to crawl across a parking area with who knows how many hundred amps going into the coils. I soon burned out the power MOSFET transistors in the motor controller. In fact, I went through dozens of them. I tried improving the motor design, but at the same time I started limiting the currents the controllers would supply to non-destructive values, and the new system wouldn't budge the car at all.

Car driven - barely - by 2nd prototype Electric HubcapTM driving a wheel directly, October 20th 2008.

Third motor, with improved coils, was then tried.

However, torque from such directly connected motors proved insufficient for practical driving.

   I started to realize there had to be some sort of gearing between the motor and the wheel, and I conceived that a variable torque converter would be the thing to keep the motor operating in its most efficient range from stopped to highway speeds. Here the gap between knowing what I wanted and and knowing how to get there was much larger. I had electronics and motor theory, and had wound and even made a motor or two before ever tackling this project. I had never been a student of mechanics... and except for some types which seemed unworkable for an automotive on-wheel configuration, nobody else appeared to have done it either.

   After spending a couple of years trying various futile things, I diverted to something less frustrating and put the stator that had moved the car into an outboard motor, along with its 9" magnet rotor. They just fit in in an open arrangement under the hood. Working on the motors again led to their further evolution.

First Electric Outboard Motor Project, October 2010

Testing the 'Electric Hubcap' outboard, Nov. 6 2010


   Soon I started testing the motors and found they weren't efficient - in fact only around 50%. At first I wondered about losses in the bearings, but it turned out the trailer wheel bearings had in fact been an excellent choice. Then I noticed that a lot of the energy went into heating up the steel stator plate disk, even though it was over 1.5 inches from the spinning magnets. Even the mounting bolts got warm. And the nail gun finishing nail strips I was making into motor laminates were thick compared to manufactured die cut laminates and so would have extra iron eddy current losses.
   I made a composite plastic plate of polypropylene and epoxy
with a bearing race moulded into it, and remounted the original stator #1 on it. The difference was unbelievable! It soon spun up to a dangerous RPM with very little power applied.
   At almost the same time, someone posted up a link to '', who made iron powder cores, mostly used for electronic switching devices. In their catalog I found some iron powder toroid cores 2" diameter by 1" tall. Wow - exactly my size! They sent me 9 samples and I wound the best motor yet.
   I had been trying to create a better 'nanocrystalline' core that would have very low losses, and found some interesting materials (as well as creating an ewn glaze for solar collector cover glass). I tried painting the coils with paramagnetic rutile (titanium dioxide mineral) in sodium silicate ("water glass") to improve the magnetic circuit. The no load currents dropped a further 15-25% or so. Then I tried ilmenite (titanium-iron oxide mineral, again in sodium silicate) on a second motor and it was even better - 25-40% lower currents.

   With these improvements, I now had a motor that was perhaps 95% efficient! Yet it couldn't be said it was mechanically sound.


   Occasional magnets flying off rotors at higher speeds convinced me (a) to adopt a better system for attaching them and (b) that the motors had to have a solid rim around the rotor compartment for safety. Covering the magnets with polypropylene strapping allows the motors to spin up to a couple of thousand RPM fine. They'll still self-destruct if permitted to go up to somewhere between perhaps 3500 and 5000 RPM, or maybe even less. The original thin solid composite rim (image below) still proved inadequate when such a speed was attained and magnets broke off, and other deficiencies were shown from the same failure, so I really beefed it up to 1/2" thick and otherwise improved it.

First closed body Electric Hubcap motor with polypropylene-epoxy body,
ilmenite painted donut coils with toroidal iron powder cores. (2011)
Paint is polyurethane with left side here still unpainted.

   I then conceived that the coils with their hollow center cores could be mounted held securely centered on "buttons" moulded into the polypropylene-epoxy body rings, and improving molds continued the process of mechanically improving the design of the motor with each new one I made. By 2012 I had good, efficient, safe motors.

2012 version Electric Hubcap

   The ilmenite coil coatings didn't stick well and flaked off, and it wasn't until late in 2013, as I developed the Electric Caik size motor, that I found a rubberized primer it would stick to well.

   I'd still like to improve on the magnet mounting on the rotors. The motors could handle higher RPMs if not for the possibility of having the magnets fly off the rotor from centrifugal force, which (for the Hubcap size) seems to occur at perhaps around 3500-4500 RPM or less, which leads me to limit RPM to about 2000. (My basic BLDC motor controller has no "smarts" to set a definite speed limit.) On the other hand, at higher speeds vibration can get pretty intense and energy is being wasted. In practice, such magnet mounting improvement will probably consist simply of a beefier layup of PP strapping with more epoxy.

THE ELECTRIC HUBCAP MOTOR FAMILY: Electric Caik - Electric Hubcap - Electric Weel - Electric Cycle Wheel

   The original size Electric Hubcap Motor has 9 coils and 6 magnet poles. The magnets are doubled-up and hence there are 12 supermagnets on the rotor. But the coils and the magnets are separate units which may be employed in various configurations. The coils can be wound as 63 turns of #14 wire for 36 volts, or 21 turns of #11 wire, for 12 volts. If the 12V coils are placed in series, three coils per phase makes 36 volts, so it can be done either way.

The potential list of practical choices reads:

6 coils and 8 magnets, 24 or 36 volts, OD 9.25", torque '2/3'. This is the Electric Caik.
9 coils and 12 magnets, 36 volts, OD 11.25", torque '1'. This is the Electric Hubcap.
12 coils and 16 magnets, 36 or 48 volts. This would be a 'large Electric Hubcap' and isn't in present plans.
18 coils and 24 magnets, 36 volts (72v is possible) This would be a 'double Electric Hubcap', not in present plans.
24 coils and 32 magnets, 36 or 48 volts, OD 28", torque '8'. This is the Electric Weel, now under development.

   In addition, one can envision magnets arranged around a bicycle wheel and attached to the rim or to the spokes near the rim, but with only an arc of coils instead of a complete circle stator, eg 6 coils, driving some of the magnets in turn as they pass by the coils. This is 'incomplete' utilization of the magnets, but it would have the same power as the Electric Caik, and should have sufficient torque to direct-drive a bicycle to good effect with no gears. It would be a silent direct drive motor with 'extra' magnets -- instead of gears. (If you can't direct-drive a car, you can still use the concept!) At age 59, my motivation for building this one is limited compared to when I was an avid cyclist in my teens and 20s, but it's in my mind to do if I find the time.

   The torque of an Electric Hubcap motor at low speed is around 1.5 foot-pounds per 10 amps of DC current from the batteries. As the number of driving elements increases, so must the motor diameter to fit them. The torque increases both with the diameter and the number of elements, so almost as the square of the size. Hence the Weel motor, with 2.66 times the driving elements and triple the effective driven diameter, is to have about 8 times the torque of the Hubcap motor. That might actually drive a vehicle wheel directly with no gearing. Two of them for double torque on left and right wheels, certainly should.
   The Caik motor was developed at the end of 2012. has somewhat more torque per amp than I expected.
   This is probably a result of using 3/8" thick magnets instead of 1/2" thick. Contrary to simple logic that says they're weaker magnets so the torque will be lower, this results in a smaller flux gap so the motor coils may be better coupled to the magnets. I'll be trying the 3/8" magnet thickness on all the motors, and probably even 1/4". Also three thin magnets (3/16" to 1/4") per pole in place of two thicker ones to even out the flux, would be a good experiment. More even flux may result in still thinner gaps. And that still leaves the possibility of using larger and thicker but less powerful ceramic magnets instead of the more costly supermagnets.
   One concern of thin gaps is that I took advantage of the thick gaps to put a wall between the "stator compartment" and the "rotor compartment" so that trouble in one (eg, magnets flying off from over-revving, or burned coils from excess power) will leave the other undamaged. If I actually get thin enough gaps, there won't be room for the wall and the whole design would have to change. I don't expect this to happen with supermagnets, tho I may have to thin this inner wall and reinforce the coil mountings via the outer wall if results are really good with thin magnets.

Electric Caik Magnet Rotor

   On another note, motors have to have air flow to cool the coils. In the Electric Hubcap family, the magnets themselves act as the blades of a surprisingly effective centrifugal fan. Air is drawn in all around the rim of the stator compartment through a 'furnace filter' strip across the coils to the center, where it goes through the central hole into the rotor compartment, is flung to the outside by the magnets on the rotor, goes around the outer edge of the rotor, and exits through holes on the rotor end.
   The high efficiency not only gives a bit more drive, but it reduces heat generated and hence cooling requiements. For example, 95% efficiency has just a bit more thrust than 90%, but it has just half the internal heat generation.

   The Electric Hubcap and Electric Caik motors have been created. The molds and jigs exist and they can be produced in small quantities. Furthermore, the designs for the molds and jigs exist as CNC files and more molds and jigs can easily be made. It is intended that they are available to anyone who wants to build these motors, and I'm willing to teach motor making.

Pieces for Electric Hubcap Motor Kit - 2011
(Case items have been improved)

   The Hubcap motor is still intended for vehicle transport, but still, after 5 years, awaits a successful torque converter to enable turning cars into plug-in hybrids. The Caik motor was originally intended for electric motorbikes and boats. The first one is in use as a converted electric outboard motor.

Electric Caik motor mounted in Honda 7.5 HP outboard motor body

Concept & Development
...reinventing the Weel

   It is intended that the Electric Weel molds and jigs be created next. The bike rim motor is planned but if I'll have time for it I'm not sure.

   In designing motors for electric transport, I had the thought that we all, even axial flux and brushless motor designers, have been thinking inside a very cramped box about what a motor is or should be, selecting dimensions and proportions that aren't appropriate for vehicle drives, and then having to gear them down to get sufficient torque to budge a car, incurring the efficiency losses and additional expenses and problems inherent in gear and transmission systems.

   This motor bursts that box in a 15 KW pancake package 28 inches in diameter and only about five inches thick, with the force elements arrayed only around the outside rim. This configuration gives it nine times the torque of a typical high-torque axial flux supermagnet motor made for vehicles, but with only three times the power. I expect it has the torque and speed range to directly match a vehicle wheel. I intend to connect it directly via a CV drive shaft to a front wheel (or to a differential attached to the driveshafts), though other couplings might also prove practical.

   The design of the Weel calls for the unmodified motive components that three Electric Hubcap motors would use: 24 coils, 32 supermagnets, and 2 motor controllers. Each controller, fed from the same control signals, drives 1/2 of the coils, at 48 volts. The Electric Hubcap motor and matching Turquoise Brushless Motor Controller have undergone three years of development and these components should impart the same 95% efficiency and reliability to the Weel.
   It was originally to have had 27 coils and 36 magnets, the "Triple Electric Hubcap", driven by three 36 volt motor controllers and having 9 times the torque of the Electric Hubcap. But since it's so large, I decided to build the stator as 8 separate octagonal sections that bolt togetther. Since these would have had to have 3.33 coils on each equal section, a 24 coil configuration was decided on, 3 per section with magnets adjusted to suit and (hopefully) using only two motor controllers, at 48 volts.

The steel parts for the prototype Electric Weel motor, cut by abrasive waterjet.
The rotor is made in two parts: a thinner (1/8") plate for lightness with a ring (3/16": total 5/16") for sufficient magnetic
conduction thickness at the rim. The original 36 magnet positions are marked around the rim. (Now there are to be 32 magnets.)

The stator will be composite plastic outside in the vicinity of the magnets. The center plate is 3/16" steel for strength and stiffness.
The eight Polypropylene-epoxy rings will attach to this metal center forming an octagonal body.
The heavy axle and bearings will handle the high torque and loads.

[Thinking back to large sawmill blades of yesteryear, it may be that the rotor would best be "cupped", hammered to a slight bowl shape.
As it spins, it stretches slightly around the outside edge and becomes flat. But blades are cupped to specific RPMs, and for a variable
speed motor it may be necessary use a thicker rotor plate, or to weld ribs onto it. (I wonder if there's anyone still around who knows
how to cup big sawmill blades! Anyway, this rotor is 26", not 48" or larger.) But I digress.]

   Owing to all the other concurrent projects, the first weel motor is still not completed and running, and now the prototype has been sold as a 'kit' to be used as a generator for a hydro power project.


   The motor controller, as mentioned, started out as the wrong type for the job: a variable frequency drive. Starting at low frequency, the motor spins up synchronously with the frequency. For this reason, three phase motors with permanent magnets were called synchronous motors, or "permanent magnet synchronous motors" (PMSM) when I was in school in the 1970s, and I hadn't heard the now usual term "brushless DC" motor (BLDC). This caused most of my trouble in looking for information about them. The main trouble with operating them with a frequency drive is that if the load is too high they get out of sync, lose power, and suddenly stall. Used on 60Hz AC power as three phase motors, they need a starter motor to get them up to speed.
   The integrated circuit I stumbled on for this controller, the IR2130, proved to be a very good 3-phase MOSFET driver chip family for the sort of motor power I'm running, but I didn't recognize it until I had taken some unproductive diversions along the way. One of the 'diversions', however, led to a major advance. But I'm outrunning the story line.

   I created a method with three optical sensors so that the frequency of the 'frequency drive' became synced to the motor rotation instead of the motor to the frequency, and the power and hence speed was regulated by pulse width modulation. This is what moved the car (barely) in 2008.
   Then I found out that my fine sensor system 'invention' is the way everyone does it, except using Hall (magnetic) sensors to pick up the actual magnet positions instead of optics adjusted to align to them. And finally somewhere I picked up the term "BLDC" off a couple of web sites, and - now that I had already done everything wrong and finally gravitated through trial and error to the right techniques - info became much easier to find.

   Another 'diversion' I took was to think in terms of the sort of high voltages everybody seemed to be using for electric vehicles, 144 volts being common for hybrid cars. I started thinking of 120 volts from 10 batteries. The day I carefully wired up 60 volts of batteries in sequence and tested things... and then casually grabbed the 60 volt wire to disconnect the batteries... I started thinking about electrical safety. Nothing happened, but what if that had been 120 volts? What if it had been damp out? That careless grab could have ended my career.
   I decided that after all, power equals volts times amps, and to run at a safe voltage just needed more amps. I took the three coils in series on each phase of the motor, and rewired them in parallel instead. That made the motors 40 volts instead of 120, at three times the current. Each coil has exactly the same voltage and current, so the motor is really exactly the same. Since it's hard to get 40 nominal volts from 12 volt batteries, it became 36 volts.
   I put away the high voltage MOSFETs and found lower voltage ones. To my surprise, they could be had rated for commensurately higher currents, like 120 amps continuous and much higher for brief periods. It seems MOSFETs are also the same power whether configured for high or low voltage.
   The chief differences then between high and low voltages are electrical safety and the thickness of the wires needed. In a building, long runs of very heavy wires would be prohibitively expensive. In a car, the longest run is if the batteries are in the trunk and the motor up front.

   I started to wonder why we didn't hear of more electrocutions of amateurs converting cars to high voltage electric. For all the discussions I had on the subject, no one enlightened me and I didn't find out until I got a converted car myself, that the entire high voltage system is ungrounded, floating. In theory, this means that any point of high voltage wiring touched by a grounded person becomes the ground point... as long as only one person touches only one point at a time, and nothing else grounds the circuit somewhere else. In practice, dust and moisture on the batteries usually cause enough conduction to give the worker a tingle or a shock. But it's enough to prevent a count of bodies of DIY - and factory - electric car workers from piling up.
   Having never considered the floating system idea, conscientiously grounding the negative to the car body, I could easily have been killed by high voltage had I continued in that direction. I still consider lower voltages to be the way to go. If the power needed can be kept down by 'ultra efficient' motors and transmissions, the wiring isn't too heavy. But I also plan to adopt the floating system idea. My larger transport motors such as the Electric Weel will then go up to 48 volts with little worry about electrocution potential.

[Last update: 2014/05/14]