Turquoise Energy News #6

Craig Carmichael   August 4th 2008

   I spent most of June working on the Electric Hubcaptm car motor and controller, but it was getting hot and it bounced around a lot. To my surprise making a new "cooler" motor stator and and an improved mounting system took most of July.
   The new motor turns the car wheel, but the motor controller needs some redesign to sync to the rotor position. That's likely to occupy much of August.

   Then I hope to give the new Ni-La (3.25v cells) Turquoise batteriestm more attention. About all I've done since May is acquire a couple of chemicals, added a new "component", and done a couple of intermediate battery chemie things. And I'm waiting for another chemical to arrive.

   It occurs to me, having "wasted" June and July, that if I just change a pulley the wave power unit's generator will probably spin, and that summer is the best time to be at the seashore. But I've dismounted the unit to use the trailer for other things, and for now the car motor and batteries get priority.

   Much of my writing this month is about the motor, with some fairly in depth (ie boring) details.

Topics Below:

Electric Hubcaptm Drive Motor Changes - cooling & mounting mods
Revised Turquoise Motor Controller
Cross Section of the Turquoise Batterytm (as currently envisioned)

The Electric HubcapTM Vehicle Drive Motor

   I found these remarks, in harmony with what I've been saying, on the web site of a company that makes "top end" custom motors:

"New synchronous motor designs using permanent magnets make it possible to have low speed and high torque in one package, eliminating gearboxes and other mechanical components.

"Eliminating the gearbox saves space and installation costs, energy, and maintenance, and provides more flexibility in production line and facility design."

   Amen! Then, changing the common radial flux motor layout to axial flux can provide these advantageous workings in just three inches "length", thus the Electric Hubcap design.

Motor Cooling

   When I put the voltage up to 36 volts (also as much as 36 amps - headed for two horsepower) and tested for a while, I could smell hot polyester. It seemed evident that the motor was going to overheat when used at full power - perhaps 60 to 72 volts - and so I did some redesign. Not wanting the complication of a liquid cooling system, I attacked the problem of better air cooling from every angle and made a new motor stator.

   I discarded the solid cast stator. In its place are individual cast epoxy coils that have no more cast plastic than necessary to "pot" the copper coil with its iron strips and two mounting bolts, into a solid mounting piece. Each coil, exposed to the air on all sides, bolts individually onto the steel backing plate, which is a disk brake rotor with heat dissipation vanes.
   Washers on the mounting bolts hold the coils off the backing plate's surface for all around air exposure, while the bolts and washers carry heat and magnetism from the coil core into the plate.
   The front of a new fairing will have an air scoop so the moving car blows air onto the motor.
   An original cooling system feature remains: the magnets on the rotor (the car wheel) act in principle as a centrifugal fan to blow air across the coils, although the RPM is pretty low except on the highway to move a lot of air. The magnets are mounted on a solid steel disk bolted to the wheel, which may be expected to carry off any heat buildup in the magnets.
   Then, the coil wires have high temperature insulation and the epoxy castings and flat black coating have very high temperature ratings. The stator paint is engine enamel and the rest is copper and steel, so the coils and stator can operate at sizzling temperatures without damage.

   Finally, depending on summer driving tests, if all these "passive" cooling features seem insufficient or marginal, an electric cooling fan - of whatever CFM rating seems necessary - may be mounted. It would start if the motor temperature gets too high, much like a radiator fan, to reinforce and provide airflow even if the car is stopped. The fan's power would be insignificant compared to the power of the large motor it cools.
   It may be that a fan will be found necessary or desirable but runs rarely -- at least in Victoria's mild climate.

   And again, having efficient motors with no gears means less overall power and heat, and having two motors splits the heat there is into two halves. Overall, prospects are excellent that the motors will run a vehicle at trouble free temperatures in most conditions. Fans should guarantee it.

   An interesting feature of the new implementation is the individual rather "generic" bolt-on coils. In theory, they could be plugged into similar motors of various diameters and ratings, from about 8 to 14 inches with perhaps 6, 9, 12 or 15 coils. They can also be replaced individually if necessary.

July 19th: New Stator with better cooling airflow completed.

Motor Mounting

   After doing my best to stiffen the stator mounting bracket, increasing the voltage still resulted in excessive vibration, and finally I realized it would be very difficult to make a sufficiently strong mounting for the stator in the available space, with all the braces having to come around from behind the wheel.

   The answer dawned on me on July 23rd: a thrust plate with ball bearings to tie the rotor and stator together at their centers. That way, the bracket from behind only needs to hold the stator on and keep it from spinning. It doesn't need to be super stiff. The thrust plate keeps everything aligned and sets the stator-rotor air gap.
   That evening and the next day, I puzzled with various bits and pieces, piecing things together and seeing various problems. Finally, some standard steel pipe end plates, a short pipe, and a small "lazy Susan" ball bearing plate, brought it all together.

A running test before supper showed this "platter" had the desired effect. There was much less vibration, even though I only had one of the two "arms" attached and it wasn't even bolted solidly in place.

 "Lazy Susan" thrust plate assembly.
Standard pipe fittings have avoided the need for custom machining.
The square end fits nicely between the four lug bolts on the wheel.

Thrust plate mounted. The length of the post was just short, and a flat 5/16" plate placed on the end provides 3/16" gap between the stator coils and rotor magnets. A large air gap increases efficiency and keeps the magnets from gradually being demagnetized by coil flux and stator heat.

Another Voltage Possibility

   Keeping to very low, and hence relatively safe shock-wise, voltages, is less practical but not impractical. There are nine coils in the motor, three per phase in series. If one wired them separately, 24 volts would provide the same current per coil as 72 volts. They would also draw three times as much total current. (1/3 volts * 3x amps = same power.) One would replace each pair of 100+ volt, 100+ amp rated MOSFETs with three pairs of 50+ volt, 100+ amp MOSFETs, one pair driving each of the nine coils. The total would be 18 MOSFETs instead of 6, and 9 heavy wires from the controller to the motor instead of 3. To prevent the range from being cut by 2/3, one would use 3 sets of 24 volt batteries, either connected in parallel or perhaps one set per each three coils of phase A, B & C.

   However, it would seem easier and more economical to carefully make the 72 volt system as safe as possible. If the power was such that it was difficult to drive with single solid state devices, splitting the coils up with separate drivers might make good sense.

Electric Hubcap Motor Factoids:

* Only the car's wheel turns. The only moving part in the motor is a thrust plate centering the stator on the wheel.
* The virtually frictionless magnetic link to the wheel magnifies useful power by transmitting it all directly to the wheel. There's no losses from a transmission or gears, it requires no gear shifting or other attention by the driver, and it's virtually silent.
* Permanent magnet synchronous motors also have the highest intrinsic efficiency of all electric motor families, further leveraging the efficient power transfer. As a guess, one might perhaps expect up to 50% greater range than other electric motor systems from the same energy, and correspondingly better performance for the same kilowatts of electricity used by the motor.
* Installation requires no connections with or changes to the car's existing mechanical components and systems.
* When not in use, the motor has no more effect on the car than any other 35 pounds of luggage.
* The motor sticks out just 4" from the wheel or a couple of inches past the fender, less protrusion than the outside rear view mirror.
* The RPM with 13 inch wheels is about 10 per one kilometer per hour of speed, that is, 450 RPM at 45 Km/Hour. Most electric motors prefer much higher speeds, but the "Hubcap" has good low RPM torque and power. 120 Km/hour is just 1200 RPM, a stately pace for most electric motors but a good upper range for the "Hubcap".
* The rotor is a 10 inch steel disk brake disk mounted on the wheel lug bolts, 6 poles using 6, 12 or 18, .5" x 1" x 2" NIB supermagnets, glued and-or bolted on.
* The stator is a similar 10 inch brake disk (but with cooling vanes), with 9 epoxy cast coils bolted to it, each of 60 turns of #14 wire, in 3 phase "Y" configuration. Magnetic flux is axial.
* A unique design breakthrough is that the stator coil iron is strips of regular nail gun finishing nails in the coil cores instead of custom die cut iron laminate sheets. With this and no axle or other moving parts, the motor is simple enough to make at home, or the coils could be wound by machine and cast, for super economical mass production. Individual coils can be easily replaced.
* The motors dissipate their waste heat via air cooling, avoiding the complexity of liquid cooling systems. There's maximal coil air exposure and heat sinking with the magnets blowing air in front of them, an air scoop on the front of the fairing and air guide vanes, plus a temperature actuated electric fan in case all else is insufficient at low motor RPMs that don't move much air and high power (eg, climbing hills and mountains).

Turquoise Motor Controller

   I've been pleased that the solid state motor controller with the high power MOSFETs actually drives the motor and nothing blows up until I do something wrong. It's only been tested up to 36 volts so far, but this testing has shown that lower voltages should be used - perhaps 60 to 72 volts instead of 120 - so 36 volts is actually at least half way (and 1/4+ of full power). Though heavier wires are required for the same power, this allows choosing 100 volt, 120 amp MOSFET transistors, more economical than the 200 volt, 65 amp parts. Bonus: the lower voltage is less likely to electrocute anyone, though anything over about 40-50(?) volts or so must still be treated as potentially dangerous.

   However, I've finally realized that the "lumpy" torque that leads to some of the vibration, and the "fragility" of operation - a tendency to suddenly stall when the load gets a bit too much - mean that the drive signals can't simply assume the motor will synchronize to them: the driving power has to synchronze to the actual magnet positions, somewhat like a DC motor with a commutator.
   Now that I'm addressing the issue, I see this can actually simplify the controller circuit as the phases can be generated simply by the optical shaft position sensors. But a slotted disk, optical elements and additional wiring for them have to be added to the motor itself, and the controller will of course need mods, so the changes are likely to take up a good part of August.

Say, a cardboard cutout slotted disk should be much easier to check for fit than a real metal one!

Will it slip on without disassembling the thrust plate assembly?
Best to cut it so it will!
To the left: a convenient old mouse PCB with four pairs of LED + phototransistor. This needs just three sets, one per phase.

   With the revised controller, which will have PWM like most controllers, the gas pedal "must" (simplest) go back from being a "speed select" pedal to a regular "accelerator" pedal. In my view this is largely a negative thing, but it's how we drive now anyway. And it does open the way for a front wheel drive car to become an all wheel drive with electric rear and gas/mechanical front drives, both on at the same time.

 Speaking of which, a differential can let one wheel spin, eg in mud or on ice, while the other is stopped and providing no thrust. Both hubcap motors provide torque at all times: it's a better two wheel drive.

Motor Controller Factoids:

   The controller converts the DC power from the battery to a variable frequency approximation of three phase AC power, on three power wires that go to the motor to create a variable speed rotating magnetic field in the stator, the motionless part of the motor. That field pulls the permanent magnet rotor on the wheel around with it and hence rotates the wheel.
    Optical sensors tell the controller the position of the magnets to time the pulses. The amount of torque is controlled by pulse width modulation proportional to depression of the accelerator pedal. Torque to slow the motor is provided by differently timed pulses proportional to the depression of the brake pedal. A reverse switch switches the signals to reverse the rotation of the magnetic field.

   In accelerating, the motor uses energy from the battery. In decelerating, the motor generates energy, which goes back into the batteries.

Turquoise BatteryTM

   Here's details of the Ni-La bipolar flat plate battery cell as currently envisioned. There are some uncertainties that may necessitate a change or two.

   First, the cell in cross section:

1. Nickel-brass plate cell wall (AKA nickel-silver but actually has no silver so the name is misleading). The current flows through the thickness of this plate across its entire length and width straight into the next cell, so it's good for zillions of amps. The two ends of the battery are extra thick pieces with bolt terminals silver soldered to them, which should handle, say, many hundreds to a couple of thousand amps.

2. A conductive layer of zinc mixed with the positive electrode powder. This zinc is not added but leached from the surface of the nickel brass cell wall sheet by electrophoresis.

3. The positive electrode powder/paste. This is the familiar nickel beta [oxy]hydroxide, with a bit of cobalt trioxide added (becomes cobalt hydroxide on first charging) to improve conductivity. Pretty much the same as in Ni-Fe, Ni-Cd, Ni-MH...

4. The electrode separator sheet.

5. Ferric oxide (or osmium) powder layer. Smeared onto the separator sheet, this acts as a hydride to store protons, hydrogen ions. Os should have better performance, but though only a bit is needed, Os is very expensive. Note that that gives us the basics of a Ni-MH battery cell. The difference is that this is just a thin mid layer. It won't hold many protons, but they just pass through to the negative half of the battery.
   This layer is important because it splits the cell voltage into two parts: the positive side (1.38 volts per Ni-MH usual) and the negative side (-.83v MH - 2.90v La/La(OH)3 = 2.07v). Water based electrolyte starts to break down above around two volts, which is why lead-acid batteries (2.1v) are the highest voltage cells besides lithiums, which use a non-water-based (but very slow) electrolyte. So splitting or "ramping" the voltage permits creating the Ni-La cell, of 3.45v theoretical voltage.

6. Zirconium silicate (zircon) or zirconium oxide (zirconia) layer. This is an electrical insulator that freely passes ions. It insulates the ferric oxide layer from the negative electrode powder while letting the protons and hydroxide ions freely pass.

7. Another electrode separator sheet. I'm not sure I trust the zirconium to have 100% coverage!

8. The negative electrode powder/paste. About 6mm thick. This is a 3 to 1 mix of monel and lanthanum [hydroxide], with a little cobalt chloride thrown in for better conductivity. These ingredients are mixed in a bean paste to add thiamine mononitrate, rolled out thin, dried, and fried until the bean paste catches fire and burns off. (about 90 seconds. This is definitely an outdoor job - the smoke is awful! I turned an electric barbecue into a special oven.) The "sintered" amalgamated powder is scraped off the sheets into containers for later use.

9. An aluminum grid next to the cell wall to improve conductivity.

10. The cell wall of the next cell. There's only one wall between cells.

   There are also the ingredients that aren't bound by the layers. The first of these is Lemon Fresh Sunlight dishsoap (accept no substitutes), which is mixed with the electrode powders and turn into semi solid pastes. The second is the electrolyte, which is poured in as water with potassium chloride (sodium free table salt), a "tincture" of methyl-ethyl-ketone (nasty solvent), and acetyl ester (AKA ethyl acetate).

   The chlorine ions and odd organic ingredients in the electrolyte [should] allow the reversible La <===> La(OH)3 electrochemical reaction for charging and discharging the negative side of the battery. I did see almost three volts in one test, but that was without the rust or zircon layers and the water discharged it quite quickly. (That continuous discharge, increasing rapidly with voltage, is perhaps why it didn't get up to 3.45 V.)

   I'm hoping for many hundreds of amps capacity, even 1000+, in a battery whose maximum actual drain/charge will be under 200 amps with two hubcap motors at full power. (Pushing things to extremes, just think: a 50 amp-hour battery could be charged in just 3 minutes at 1000 amps, a 72 volt battery drawing around 300 amps at 240 VAC. How big is the main breaker on your house?)

Now I've told you all the secrets... so what do you need me for? Well, maybe you should hold off until I've actually got one working.


Victoria BC