kilowatt-car is about what can be done to make a modern car work on electricity. So far this is just theoretical: I am getting a feel for how fast these technologies are approaching the point where I can take my venerable Jaguar sedan and convert it to electricity:
1. The challenges of an electrification project
2. Dimensioning an electric car - the motor
3. Dimensioning an electric car - the battery
4. Dimensioning an electric car - other factors
5. Battery sizing, perceptions and behavioral changes
6. Electric vehicle architectures
Dimensioning an electric car
When we think of conventional motor cars, those of us with an interest in the subject – even if only as owners and drivers – have an idea of the figures. For a mid-size car, 200 hp accelerates pretty well, a 15-gallon tank will take you 300 or 400 miles and you would look for a 2 – 3 litre, 4 – 6 cylinder engine and 5 forward gears.
But what’s the equivalent for an electric-propulsion train? Are the key parameters even equivalent to the petrol-driven motor car? That’s the purpose of this investigation.
There are different ways to go about it, but this one seems logical to me:
The key questions are how fast and how far the electric car should go. Given reasonable assumptions for these parameters, the remainder should fall into place.
Before we start, note that this project will be much, much easier to do in 5 or 10 years. All the motor and battery equipment is at an early stage of development, and most of it is tied up in deals between suppliers and electric car manufacturers – it is difficult to get good, purpose-built, inexpensive electrification equipment for a car project at the moment. Also, battery technology in particular is advancing at around 8% per year, and this will continue for a few more years, so one can predict that in 5 years’ time batteries will be 40% lighter, 40% smaller and 40% cheaper than today, a huge savings. If these advances come to pass, an electric car will indeed be lighter, accelerate faster and have the motoring range (hopefully) and price point of the petrol equivalent today.
Also note that this is, as yet, a thought experiment. I’m not building one of these vehicles for a couple of years, as it is difficult to get high-specification AC motors and controllers, while advanced batteries, though available, are still prohibitively expensive if one is looking at a large sedan for even a 70+ mile range.
The base car
Like most electrification projects, this concept starts with a conventional car, strips out the petrol engine and all unnecessary equipment then fills the space (and replaces the weight) with an electric motor, controller and battery system. The car I have chosen to dimension is larger and heavier than most, which will make it an expensive project – light weight and low aerodynamic drag are the two best characteristics for an electric car – but you have to start with a car you love, or you will not appreciate the end result.
The car is a 1999 Jaguar XJ8L. I have owned it for 10 years, it’s beautiful, it rides like a dream and it was built long after Ford acquired Jaguar, so the quality is excellent. But to add some variety, here are a couple of comparisons. The Toyota Prius (2004) is already the favorite hybrid; let’s see how it would look in all-electric clothing. And the Mazda Miata is an iconic sports car, and light enough that it is already the subject of a decent EV conversion kit.
The key specifications (the ones we use for these calculations):
(Note that for weight, a critical dimension, I take specified ‘curb weight’ and add 136 kg for driver, etc.)
Now, to dimension the motor.
After all the sound and fury, petrol or other forms of energy put into a car can move you from one place to another, and produce heat. There is drag on the vehicle from friction of movement, which ends up as heat. So does air resistance: it creates turbulence, moves air around and ultimately produces heat. When you use the brakes to slow down, kinetic energy from traveling at speed turns into heat energy from friction between the brake pads and disks. And when you travel up a hill, returning to the original elevation involves working against drag and some braking, for more heat. This section is about how to calculate the drag on the car under different conditions, sizing the motor to overcome that drag and give acceleration with the surplus torque, and adding hill climbing to the equation to ensure the vehicle won’t grind to a halt on steep grades.
At a steady speed, on a level road the motor’s force has to just equal the sum of the different components of drag on the car. The two significant contributors are tire drag and aerodynamic drag.
Tire drag is the force required to roll the tires (and the car) along the road at a given speed. It is mostly due to the energy dissipated in flexing the rubber sidewalls as the tire’s flat footprint on the road moves around. As the sidewalls flex, they become hot, the end result of this energy dissipation.
Tire drag is measured as a coefficient of rolling resistance, Crr. This is a dimensionless constant, which is multiplied by the normal force (or weight) of the car to give a force resisting motion. For our Bridgestone Potenza tires on the Jaguar, we assume a Crr of 0.011, which multiplied by the weight of 1946x9.8 N (Newtons) results in a drag of about 210 N.
There is a secondary effect which causes tire drag to rise with speed, mainly due to the increased aerodynamic downforce on the car body. This adds less than 10% on top of static drag at 30 mph (50 kph) but rises to 50% at 70 mph (110 kph), as it is proportional to the square of the speed.
The second source of drag is air resistance. This is calculated by multiplying Cd, the coefficient of drag, by the frontal area of the car, A. Then multiply by the square of the speed, v2. Cd x A x v2. For frontal area I just take overall height times overall width (less mirrors), which is a little conservative.
The graphs above show the contributions of rolling resistance (RR) and aerodynamic drag on the Jaguar, as a stacked column graph of force in Newtons and the equivalent power in kW. This, the line graph on the right axis, is the important conclusion. It tells us that it takes 2.5 kW to maintain a steady speed of 20 mph, 12 kW for 45 mph and 24 kW for 60 mph. So let us pick a suitable top speed: not snail-like, but neither over-optimistic. 90 mph requires about 71 kW, so I will pick that number and see how the other dimensions come out. Note that depending on conditions, we may never reach that ultimate velocity, unless going downhill, but we should be close. For comparison, here are the constant-speed power curves for all three cars:
At 90 mph, the Jaguar requires 71 kW rather than 52 kW for the Prius and 57 kW for the Miata. The differences are mainly in the Miata’s much lighter weight, and the Prius’s somewhat lighter weight and much better aerodynamic shape. For a check, here’s Tesla’s model for their Roadster (full data here). To maintain higher speeds, the Jag requires twice the power of the Tesla, due to all the factors noted above.
The graphs above show how much power must be delivered to the wheels, just to overcome drag effects and maintain a constant speed. If we want to accelerate, the motor’s power must be greater than this combined drag.
Calculating acceleration requirements
This is getting a little ahead of ourselves, as we can’t really model acceleration until we know about the actual motor characteristics and the gearing; but we can derive an expected or ‘needed’ value of power at the wheel to give us the required performance.
Let's say for argument that we want 0 - 60 mph in better than 9 seconds. This is about average for new US car sales in 2009. In petrol guise, the Jaguar's 0 - 60 time is 6.8 sec. Our goal is to match this performance after the electric conversion.
For a simple power calculation, ignoring drag, we can take the initial energy state (zero), and the final state (E = ½ m x v2) which is the car’s kinetic energy. Plug in the Jaguar’s weight and 60 mph and we get ½ x 1946 x 262 = 658 kJ, and divide by the 9 seconds it takes to get there and we get 72 kW. Now, this is a very coarse estimate because it ignores the drag figures we calculated above, and the fact that the motor doesn’t give constant power over its speed range: early on we will be torque-limited, although an electric motor’s torque is very impressive. Let’s say these effects make us add 50% to the power, then we need 108 kW from the motor/inverter/battery system. This is about 150% of the continuous-power rating of the motor we chose, 72 kW for 90 mph top speed.
Does this mean we need a bigger motor? It doesn’t. The reason is that electric motors are rated at continuous output, but can reach at least double this power for short periods of one or two minutes: plenty of time for normal driving acceleration sequences. So our 72 kW motor will be capable of delivering 140 kW to accelerate: in fact we will do better than the 9 second 0 – 60 mph target.
More detailed calculations… We need to determine the final drive ratio for this motor. While it is possible to use a gearbox in an electric car, it consumes weight, efficiency (more battery drain means lower range!) and it’s more expensive. Most electric cars (the Tesla, for instance, once it got past initial teething troubles) use a fixed gear ratio.
If the car already has a manual gearbox, home-conversions often leave it and the clutch in the car, as it is easy to connect the electric motor to the bell housing via an adaptor plate. And if the power rating of the motor or battery is marginal, it may be useful to shift down to 1st occasionally for steep hills, or up for high speeds. But we are spec’ing out a professional electric car here, so there is no gearbox (for instance, early Tesla’s had a 2-speed transmission but as soon as they could, Tesla got rid of it).
Since electric motors have very good torque over a wide speed range, including full torque from a standing start, it’s quite possible to use a constant-ratio final drive: to connect the motor directly to the final drive differential. One could argue it’s even better to put a motor on each wheel, a concept Michelin has been exploring, but that would need a great deal of engineering. For the Jaguar project, we would mate the motor to the rear-axle differential, probably with a different ratio.
To calculate the ratio, we need to know the highest speed we wish to attain, and the motor’s maximum speed. And for this we need to specify a motor and read its power and torque curves. Unfortunately, AC motors for electric vehicle use, the best technical choice, are not available in the power rating we need for the Jaguar. Sources of AC induction motors include ‘Metric Mind’ Corporation, which retails BRUSA and MES motors. All ‘professional’ electric car developers, such as Tesla, AC Propulsion, Chevrolet Volt, etc, use AC motors. Azure Dynamics sell AC motors up to 40 or 50 kW continuous rating, but not powerful enough for the Jaguar; whereas the major manufacturers, including Siemens, AC Propulsion and UQM, do not sell to the public; all their production is tied up with OEM electric vehicle companies and some academic research labs.
Indeed, Metric Mind used to sell the Siemens 100 kW motor, aimed at the urban hybrid bus market, but Siemens stopped selling their controllers to the public, and since a controller must be matched to its AC motor, this outlet is no longer available. I have not been able to find an AC motor of more than 40 kW that can be purchased today, but undoubtedly they are coming, so I will be taking specs from UQM Technologies, a Colorado company. This is the ‘Power Phase 145’ 75 kW (continuous) AC induction motor. Its torque and power curves are shown below.
The characteristics of electric motors are a good power curve over all the usable rev range, and full torque from rest. The red line above, for torque, shows 400 Nm from rest to 2200 rpm (for reference, the Tesla single-speed gearbox is rated at 380 Nm), when the revs catch up with it and the curve crosses with the peak power curve as it rises to a 140 kW max at 3600 rpm. A reasonable rev limit for this motor might be 5,000 rpm at 90 mph, so we will use those figures to size the differential.
We know that the Jaguar’s tires turn at 785 revs per mile. So at 90 mph, they will be turning at 1177 rpm. If that corresponds to 8000 rpm for the motor, we need a reduction ratio of around 7:1. Now, we know how fast the motor has to spin to move the car at a given speed, since the ratio is constant.
The graph above is derived from the motor torque plot, and the drag calculated earlier (both rolling and aerodynamic resistance). It gives us a 0 – 60 mph time of 7.1 sec for the 7:1 differential ratio, about right. As with a petrol engine, the gear ratio balances top speed against acceleration, improving one at the expense of the other. We could make the car accelerate faster by increasing the ratio, but the motor would be spinning faster, and we would limit the ultimate top speed a bit, as well as increase wear on the motor.
Note that I haven’t accounted for some factors in the calculations above (or later on), such as the efficiency of the diff (should be 96+%, rotational acceleration of the tires and half-shafts, idle braking drag… but the basic drag assumptions are conservative, so they should compensate. But don’t assume these figures are any closer than 10% accurate: hopefully one day we will be able to prove them out with a real car.
Hill climbing ability
All the calculations so far assumed flat roads, but of course cars must be able to climb hills, even accelerate up hills! It’s relatively easy to add hill climbing to the calculations; it works as extra drag on the car.
Here’s the effect of a hill on acceleration, at the kind of grades we usually meet. The steepest streets in San Francisco are about 30%, and at low speeds the Jag will just be able to get up them! At low speeds, with high torque from the motor, acceleration is still very brisk, even with a hill. At highway speeds, it’s a bit more challenging because of the extra aerodynamic drag. The strange plot for a 25% hill says we won’t get above 55 mph; and for a 35% hill, the top speed is 45 mph. Luckily the steepest hills are in cities and provided we can attain 30 mph no one will notice!
Comparison of performance with petrol-engined cars
First of all, for a quick sanity check we compare the electric Jag with its original petrol version. The original’s engine is specified for 216 kW at the flywheel, compared to the UQM motor’s peak power of 140 kW at the final drive. For a fair match at the final drive, subtracting gearbox losses, perhaps it’s 190 kW to 140 kW. The torque of 393 Nm compares to the electric Jag’s 400 Nm. But the original gearing is higher, with the gearbox dropping through ratios of 10.9 for 1st and 6.7 for 2nd gear, by which time it has passed 60 mph. Overall performance shows the electric version is quite close to the original: the 1999 Jaguar XJ8L took 6.8 sec to reach 60 mph while the model predicts 7.1 sec.
For comparison purposes, I compared the electric Jaguar’s acceleration against some 2006-2010 models’ reported performance. The Tesla is, of course, electric, and its specifications are 185 kW peak, 373 Nm with a single-speed gear ratio of about 8:1. So far it’s comparable to our Jag, but the weight, at 1370 kg compared with 1946 kg makes all the difference.
The Jaguar even looks rather anaemic, between the Honda Civic Hybrid and the Accord; that shows how far performance has improved in 10 years. Also, when comparing these electric vehicle specs to conventional cars, consider that no one drives a modern car with their foot on the floor and the engine red-lining. Not for long, anyway, as it will need frequent trips to the mechanic! But there is no reason not to run an electric motor in this way: electric vehicles can accelerate for 5 – 10 seconds at maximum torque or maximum power without any damage or even excessive wear, although there will be stress on the transmission components.
Dimensioning motors for other cars
Just for variety, I wanted to experiment with other car platforms for the model. The idea is to keep the performance close to the original, rather than make them go as fast as possible, because we will see the tradeoff in weight and range later on in the analysis.
Because both the Prius and the Miata are quite a bit lighter than the Jaguar, and the Prius also has low aerodynamic drag, we established earlier that the power required to maintain a constant 85 mph is about 45 kW for each car, so the electric motor chosen for both is the UQM PowerPhase 125, rated at 45 kW continuous and 125 kW peak. Plugging in the equivalent numbers, we get this set of performance curves:
The before-and-after comparison shows the Jaguar is slightly slower, while the Prius and Miata are slightly faster than in petrol guise. The Tesla shows that a light electric car with a high gear ratio has excellent acceleration.
Summary of dimensioning steps
The table below lists the steps in this section.
It doesn’t have to be done exactly this way, but it seems logical to me. The exercise produces a motor/inverter selection and a predicted acceleration curve – assuming of course that the battery can deliver the peak power required by the model. The battery system is the subject of the next section.
And finally, the penalty for getting it wrong
If the motor is too small, the vehicle’s top speed will be limited, and acceleration, both on the flat and uphill will be worse than expected. This could be very disappointing.
But if the motor is too powerful, there is not much to worry about. It will be a bit larger and heavier than necessary, and that will affect range, but only to a small degree. A larger motor does not force a larger battery, for instance. Indeed, the efficiency of AC motors is good across the rpm range. The UQM PowerPhase 145 motor/inverter set, for instance, is rated at 92.5+% efficiency from 3000 to 5500 rpm, and 85+% from 1000 rpm upwards. These efficiency figures will become significant as we dimension the battery for the electric Jaguar.
The battery is the single most expensive component of an electric vehicle, and will be for some time to come. It’s also extremely heavy, and it takes up a good deal of space. Most of the compromises that must be made to convert from petrol to electricity are due to these three parameters of price, weight and size for a given energy storage (or power delivery) level.
As we shall show, the main concern when dimensioning the battery is the car’s required driving distance between charging stops – its range. With an arbitrarily low range, the problem is solved with an arbitrarily small battery pack. If the specified range is too great, the battery will take up so much space there will be little left for people, and its weight will become an anchor on performance, to say nothing of sky-high costs. Therefore much of the skill required to design an electric car at this early stage of the technology goes into reducing aerodynamic and tire drag, specifying the minimum range that will be acceptable to the consumer and coaxing that range out of the smallest possible battery pack.
Desired range: how far should it go?
Another of the articles on this page, ‘Battery sizing, perceptions and behavioral changes’ discusses the various factors influencing an electric car buyer. It concludes that for a suburban passenger car, an achievable 100 mile range is a good minimum target, if the product is to appeal to more than a small niche market. Indeed, the only two mass-produced (450+ copies each) electric cars of the modern era, the Mini-e and Tesla Roadster, have practical horizons of about 100 and 180 miles respectively, so perhaps 100 miles is a little optimistic. Nevertheless, that is the figure we will be using in the calculations below.
How many electrons per mile?
In modeling how far a given battery will take the car, we need to know how much energy will be used per mile – a surprisingly variable figure. The model developed in the ‘motor dimensioning’ article on this page predicts the power level required for steady-speed driving:
By multiplying power by speed, we can derive the energy from the battery in Watt-hours per mile:
We can compare these figures with this excellent graph for the Tesla Roadster taken from the Tesla web site:
As the Tesla engineering blog points out, somewhat tongue in cheek one suspects, Tesla owners have engineering and analytic tendencies, and as they are always measuring and comparing their mileage it is important for Tesla to have good reference models for expected performance. Tesla claims the graph above, and their other figures, are the result of a model they have checked against actual performance, so it has credibility. It broadly follows my calculations, except that at low speeds Tesla factors in a number of transmission-related inefficiencies, so it starts around 350 Wh/mi, decreasing quickly to 150 Wh/mi at 10 mph. After 20 mph, as aerodynamic drag first becomes significant, it begins to rise along the curve I calculated, with aerodynamic and tire effects (related to tire dimensions and construction, and to the car’s weight) providing quickly increasing resistance.
For a couple of check points, here are my figures for the Miata compared with Tesla’s: the model for the Miata predicts 116 Wh/mi at 30 mph, 167 Wh/mi at 40 mph, 230 Wh/mi at 50 mph and 310 Wh/mi at 60 mph.
30 mph: Miata, 116 Wh/mi; Tesla, 150
40 mph: Miata, 167 Wh/mi: Tesla, 175
50 mph: Miata, 230 Wh/mi; Tesla, 210
60 mph: Miata, 310 Wh/mi; Tesla, 250
Compared to the Tesla curve, my Miata model has too little drag at lower speeds (to 40 mph) and too much at higher speeds – although I am loath to admit Tesla is right and I am wrong: of course there may be good reasons for the differences, as they are different cars. But they are not too distant one from the other.
The models above are for constant speeds, so there are no acceleration, braking, cross-wind or other effects. These all affect miles per gallon figures in the real world, and they also affect Wh/mi figures for electric cars. In order to take them into account, a standard driving cycle or course is required. The US Environmental Protection Agency (EPA) has defined a number of driving cycles for just this purpose. Most electric vehicle manufacturers use US06, which is supposed to be a quite rigorous course involving city streets and freeway driving. The graph below is the speed-time profile for US06.
My reading of US06 indicates there is no hill-climbing component: the profile just specifies a forward speed at a given time from the beginning of the test. But the profile above shows some obvious stop-start motoring, some rapid acceleration and braking, and a fairly long 60 – 70 mph component (the freeway part).
Developing a consumption profile for electric vehicles
We already have good estimates for steady-speed energy use in Wh/mi, but we need to develop a model that includes acceleration and braking. Although we developed some figures for acceleration in an earlier note on this page, ‘dimensioning the motor’, life gets a little more complicated when the motor is running at lower utilization levels – an electric motor’s efficiency, while always high, varies with load and acceleration. Generally, the heavier the car, the more energy is required to accelerate it to a higher speed, another reason weight is so important in electric vehicle design.
I’m going to flunk as far as exact calculation goes, it requires quite complicated modeling software, but a few graphs will illustrate the behavior.
Obviously, the higher the speed the higher the kinetic energy of the car: accelerating to a higher speed increases its kinetic energ. Take a figure of 50 mph. The graph above shows that for the Jaguar we require about 470 kJ, or 130 Wh. Add perhaps 20% because the motor, etc are not perfectly efficient, and we have about 160 Wh. This must be added to the energy used to propel the car at constant speeds, integrated over the speeds and distances traveled… which is complicated, and difficult to explain. But consider a piece of road a mile long. Accelerating from rest to 50 mph on this road will take about twice the energy required to just maintain a constant 50 mph, if we were to take a 50 mph rolling start and just maintain the speed. Acceleration eats up battery life, and the heavier the car, the more it takes.
But although acceleration is our enemy in battery terms, deceleration is now our friend. The AC motors used in serious electric cars (the simplest hobbyist’s conversion would be a DC motor, which is more difficult to control in this way) uses regenerative braking. This takes input from either the brake pedal, or even (in the Tesla) when the driver lifts off the accelerator pedal, and sets up the car so the wheels drive the motor, generating electricity that recharges the battery. The harder we brake, the more electricity is returned to the battery. Of course, the cycle is not 100% efficient, but perhaps 60 – 75% of the energy used to accelerate to a given speed can be recovered by regenerative braking, as we bring the car to a stop.
Hybrids such as the Prius and Honda Civic use regenerative braking, although we have not developed the technology to the point where we can dispense with friction brakes – saving the unsprung weight and expense of discs and calipers, hydraulics and boost systems would be nice, and no doubt it will one day become reality. In the Prius there is a transition from regenerative, electric braking at light pedal pressure to friction as the foot gets heavier. Regenerative braking improves an electric car’s city mileage considerably, compared with petrol engines.
Real-world consumption figures for electric cars
The key figure we need to derive is ‘how many Watt-hours per mile’. With that, we can set the desired number of miles and calculate the size of battery required. It’s not easy. As I have noted elsewhere, there are very few reliable records of mass-produced electric cars, tested under repeatable conditions. Here are a few data points from the Web (DC figures are battery-to-wheel, AC are wall-to-wheel, so they should be reduced by perhaps 10-20% to allow for charging inefficiencies):
So it is reasonable that a small, light electric car with a battery that doesn’t add too much weight, and using regenerative braking, can achieve around 4 miles to the kWh in ‘normal’ driving. Our Jaguar XJ8 is considerably heavier than the Tesla or Mini-e, so I will bump it up by 25% to 310 Wh/mi. This is the figure I will use to dimension the battery.
Required range and battery capacity
This is a very simple calculation: if we require the car to go 100 miles, and it takes 310 Wh/mi, we need a 31 kWh battery pack.
But anyone who owns a cellphone or laptop computer knows that the batteries don’t take kindly to complete discharge – running flat shortens the battery’s life and limits its charge capacity – and the batteries used in electric cars are no different. So if we assume that we run down to 25% of charge at a minimum, we will need a 42 kWh battery in the Jaguar (rounding up). This is just a bit larger than the Mini-e’s battery pack, which holds 35 kWh nominal and 30 kWh usable.
Battery chemistry and physical characteristics
Most hobbyists today would use the large Lithium-ion batteries made in China under the brand ‘Thundersky’ or ‘Sky Energy’. Some choose batteries from ‘A123’, but these are difficult for the home hobbyist to purchase. Thundersky’s batteries come in large, rectangular cases in a variety of capacities, where each battery is a cell with a nominal usable voltage of 3.2 Volts. Therefore a 100 Amp-hour battery would hold 320 Watt-hours, depending on the particular battery type. In this note, I have brushed over the many variations on battery chemistry: Lithium-Iron-Phosphate, Lithium-Polymer, etc, as battery science and chemistry, while fascinating, is a distraction from the task of designing a successful electric vehicle. I take the Thundersky Lithium-ion specifications as the standard, for these purposes.
To design the battery pack we need to connect a number of cells, and we can choose series or parallel circuitry for this. The easiest way to use the large-format Thundersky cells is to pick the desired pack voltage, find how many cells are necessary and then calculate the Ah capacity of each cell. For example, if the motor controller is designed for 320V, it will take 100 cells (at 3.2V each) in series to achieve that voltage. And if the required capacity is 42 kWh, each cell must hold 131Ah (42,000/320). The closest Thundersky battery would be the 160Ah model, TS-LFP160AHA. Since it’s a little larger than we need, we could adjust and use just 80 of them. (In practice this balance between voltage and capacity would be given a lot more consideration, as it’s usually most efficient to use the highest possible voltage the controller will accept.)
Now we have a battery pack design of 80 Thundersky 160Ah cells connected in series, providing a nominal capacity of 41 kWh. The capacity figure in kWh is directly related to the vehicle’s range, but we also need to check that performance will be adequate for the acceleration and top speed we require. This is related to the maximum discharge rate, in Amps, the battery can drive (assuming it’s not limited in the controller or motor – either way we need to check the figure).
The Thundersky data sheet tells us the maximum constant discharge current is 3C, and the maximum pulse discharge current is 20C (the ‘C’ or ‘1C’ figure is the nominal Amp-hour figure of the battery). In this case, the 3C rate is 3x160 = 480Amps, giving a maximum continuous power figure of 256V x 480A = 122 kW. This is well in excess of the maximum we need to sustain a steady 90mph in the Jaguar, which from the graph earlier in this article is 70 kW, and it will not impact the performance (acceleration) curves calculated earlier. In general, we would expect this result: if a Lithium-ion battery pack is designed for 100+ miles range, with a nominal voltage of 200-400V, it should not be limited by peak power. Indeed, if we were to use the ‘pulse’ figure of 20C, we would be capable of 800+ kW – a figure we would never reach because the motor controller is programmed with limits.
The diagram above shows the discharge curve of the Thundersky TS-LFP160AHA. It demonstrates that we would like to keep the maximum current draw to a minimum because, all things being equal, the higher the current (5C is high, 0.5C is low) the less energy we can get out of the battery: capacity is not a static figure, but dependent on this and other parameters such as temperature. The Thundersky battery loses 10-15% of its nominal capacity at a temperature of -25C. As we often find in life, the nominal figures are closer to best- than worst-case.
Physical characteristics of the battery pack
We have established that we require 80 of Thundersky’s 160Ah cells to make the battery pack. Each cell is nicely cube-oid:
Dimensions for the cell are shown above, and we can calculate that the overall pack will take up 288,419 cm3, 0.288 m3, or close to 300 litres. This could be packaged as 1m x 1m x 0.3m, or 3x3x1 foot: it’s quite large. In the Mini-e the bulk of the battery is where the rear seat should be, in general it would be good to keep it central in the car, and as under the floor if possible to keep the centre of gravity low. Of course the battery pack will include more than just the cells: wiring, battery management PCB units and cooling will have to be accommodated.
Since each cell weighs 5.7 kg, the whole pack will be 456 kg, or 1,000 lb. For reference, the Mini-e’s battery pack is specified at 260 kg for a nominal 35 kWh, so the Thundersky cells appear to be a lot less energy-dense than the A123 cells used in the Mini. It is clear that with this degree of extra weight, it will be very difficult to keep the converted Jaguar’s weight close to the petrol version. Despite removing the entire engine and transmission, petrol tank and radiator, the addition of the electric motor and its controller as well as the battery pack will more than compensate. The Mini-e ends up at 1465 kg (3230 lb) compared to the stock Mini Cooper at 1206 kg (2660 lb), so we should probably expect a 20% increase in weight compared to the petrol Jaguar. And if this is true, my performance calculations earlier are all optimistic, as they took the car’s current weight… but I am indeed optimistic that battery technology will improve quite rapidly in this regard, perhaps reducing weight by as much as 20% in three or four years.
Using current specifications, the currently-available Thundersky battery specifications and some simple theory we established that a 256V, 160Ah battery pack would be sufficient to drive the electric Jaguar XJ8 for an effective range of 100 miles, leaving a 20% safety margin in the battery. The battery pack will be bulky and, most importantly, very heavy, but we should remember that prior to Lithium-ion technology, lead-acid batteries and latterly Nickel-Metal-Hydride, were even worse: a lead-acid battery would be at least twice the weight of the Thundersky equivalent. It is difficult to predict how reliable and long-lived the battery would be: Thundersky quickly established a very poor reputation for quality, but comments since the end of 2008 have been more positive in terms of meeting their specifications, at least when delivered. Thundersky’s spec sheet claims at least 3000 charge cycles (to 80% discharge), enough for 8 years at one charge per day, but I am not aware of any independent long-term test results. Tesla now has a 7-year life estimate on its battery, increased from an original 5 year figure.
The battery pack designed here would cost $15,500 from EV-Components near Seattle, the price should also be coming down over the next few years, but is certainly significant today.
The pace of technology is impressive: before 2009, very little of the information I have drawn on and referenced was available, particularly from the Tesla Roadster and Mini-e fleets, which must have tripled the number of electric vehicles on the road in daily use. These two vehicles are the first real large-scale, mass-produced electric cars, but hopefully we will see many more in 2010 and subsequent years. If we are lucky, the next few years will see new advances in battery chemistry offering breakthrough improvements – research laboratories report such possibilities frequently, but as yet the Lithium-ion and Lithium-Poly batteries are the best to find their way to commercial service.
In this article I have ignored a number of practical areas that a serious conversion would require – including battery box construction, cooling, battery management systems, cooling design and cabling considerations, and the whole field of charging and fast charging – but the discussion above should serve to introduce the considerations and mathematics required to dimension a battery pack. As time permits, I would like to investigate these topics in other notes on the site.
Once the motor and the battery are dimensioned, there are some secondary factors to consider, such as weight, cooling and auxiliary systems.
In the technology industry, most new products, and even product ideas are routinely portrayed as ‘revolutionary’ or ‘game-changing’, to the extent that the terms are cheapened by over-use and over-stretch. This is a shame, as the technology observer becomes inured to unreasonable claims. But if we see 100 potential revolutions for every one that succeeds, we also recognize a corollary: a truly revolutionary product will initially appeal to only a minority of the market, and behavioral changes may be required to take full advantage of its benefits and avoid its drawbacks – and we also recognize that behavior and customs are difficult to change. Both Geoffrey Moore of ‘Crossing the Chasm’ fame and Clayton Christensen who wrote ‘The Innovator’s Dilemma’ postulate, interpreted loosely, that the truly innovative breakthroughs don’t perform existing tasks with new technology, but completely change the way we work, opening up better methods of achieving our desired objectives.
During the early-adopter phase of an innovation, pioneering customers are drawn by one or two key attributes of the new technology, allowing them to achieve something they value so much that they are willing to put up with the many rough edges elsewhere. These customers become the core cheerleading constituency for the innovator, providing enough of a market that he can sell his product, even while improving it to the level where its price and performance become acceptable to a broad audience. This is also the phase where mainstream attitudes change, from initial resistance to change to eventual support of the new technology, and where we can start to see whether people are prepared to change their behavior to work around those areas where the new product does not measure up to the prior solution.
We face a number of barriers to acceptance when promoting electric car technology. In this note we will explore two major objections, ‘it only goes 60 miles between charges’ and ‘it takes 6 hours to recharge’, to see whether they can be overcome by technology, or if we must find a behavioral solution. Almost every evangelizer of electric vehicles has encountered these arguments, and we must recognize that they are rooted in fact, and overcoming these ‘inconvenient truths’ is indeed central to our ability to build acceptable vehicles for everyday consumers.
Comparing batteries with petrol tanks
The root of the problem, indeed the key to delivering a successful electric car, is battery technology. Conventional cars incorporate a fuel tank that has little intrinsic weight and is relatively small – it doesn’t cost much in design compromises to double the size of a fuel tank – so the range can be doubled at the cost of the weight and volume of the extra liquid, and petrol is a remarkably energy-dense fuel. We have developed, over the evolution of the conventional car, a design formula where 10 – 20 gallons of on-board fuel is sufficient to drive 300 – 400 miles, and we have grown comfortable with the consequence of needing to visit a service station every few days, or periodically on longer trips. We also get nervous when we know we have less than perhaps 30 miles range left.
The engineering and economic tradeoffs of electric cars lead us to a different design solution. Since the battery is now the heaviest and most expensive component in the car, and an empty battery is ostensibly the same as a charged one, increasing the range drives an unavoidable increase in the weight and space of the battery pack, and significantly affects the cost of the car. Over the next few years, electric cars will be designed with the minimum ‘adequate’ battery size, so it is most important to identify what that minimum limit for acceptance may be, and also to understand and test the public’s perception of adequate range, and behavior when driving a vehicle with much-reduced range compared to conventional cars.
To take some figures, most electric cars can travel around 4 miles per kWh of electrical energy from the battery. This means that for a range of 60 miles we require a battery pack with a usable 15 kWh (nominal 20kWh, as batteries cannot be fully-discharged). With today’s Lithium-ion technology a nominal 20 kWh battery pack weighs 280 kg, takes up 18 litres of volume and costs $7,500 (retail hobbyist’s price).
To build an electric car with a 300 mile range would require 5 times these dimensions, or 1400 kg, 90 litres and $37,500. And because the car would be carrying the extra weight, its range would actually be less than the expected 300 miles: an electric car’s range is sensitive to weight.
The Tesla Roadster, possibly the longest-range production electric car today, uses a very light, aerodynamic frame and carries a very high price tag. Tesla has chosen this market niche because it allows them to equip the Roadster with a large, expensive 53 kWh (nominal) battery pack which gives a nominal 250 mile range. They could have chosen a much smaller battery: a half-size pack would allow Tesla to reduce the cost by perhaps $30,000, but might have reduced the vehicle’s attractiveness to buyers. So the design formula appears to be right on the money, but a car that sells for $100,000+, of which perhaps 35% of the cost is in the batteries, is not a mass-market car. The Tesla is a remarkable success story, but if there’s a conclusion to draw from the battery size, it is that the public won’t accept a nominal range less than 250 miles. If electric cars are not viable with shorter range than 250 miles, it will take many years before they approach the cost of their conventional equivalents.
One way to solve this cost barrier is, of course, technology. Battery experts have suggested an improvement curve where Lithium-based battery weight and cost is reduced by 5 – 8% per year for a few years, but 250 miles on a charge will be out of reach of all but the most expensive cars for at least a decade at that rate. Alternatively we can wait for a better type of battery, and research labs report many candidates in the early stages of development; but naked optimism is a poor policy approach.
This explains why the consumer’s perception and daily habits must change if electric cars are to be successful. A number of driving phenomena are different in the electric world, and we can perhaps take advantage of them to realize a viable market for a car with a range of less than 100 miles. Some of the key questions are the viability of a car with 100 mile range (followed by a multi-hour charging period), how to charge cars that are parked outside overnight, and how day-long or multi-day trips can be handled.
In our daily driving, our electric charging habits will be different than for petrol cars. For those who garage their cars at night, it will become second-nature to plug in when returning home. This means that instead of refueling at the petrol station every few days, we start each day fully-charged. Even better, if employers, shopping centres and other destinations can be persuaded to add recharging stations, the longest stretch between charges will become home-to-work rather than the daily round-trip. In practical terms, the electric car owner should start to think of the longest daily trip, or even half that trip, as the benchmark for the car’s range between charges. With this behavioral change, electric cars become will viable for everyday use, even in the suburbs.
This is not a new argument of course, but it is one where electric car proponents and the rest of the public talk past each other. The obvious question is ‘what’s the maximum distance you drive in a day?’ But as consumers hear this, there is an instinctive push-back, at least for any range below 100 miles.
Range anxiety in the suburbs
Perhaps as a lesson from conventional cars, consumers develop range anxiety at a relatively high threshold. Conventional cars have conditioned us to fill up when there is anything from half a tank to a gallon of petrol remaining – say 30 miles for most people, before they start to worry about finding a filing station soon. Two factors that contribute to this are the uncertainty of how much petrol remains – car fuel gauges are analog and not so accurate around the one-gallon mark – and also the variation in miles per gallon due to acceleration, hill climbing or high-speed driving. So it is this 30 mile distance that tends to stick in the mind, and this is a much more significant fraction of a 60-mile electric car’s range than for a 300-mile conventional car. One study in Japan indicates that even when drivers know that a car is good for another 40 -50 miles, they become anxious and seek out a charging station: in other words, they won’t use the car to its potential. This reinforces the 100-mile target as a ‘safe’ range for a ‘usable’ electric vehicle in the public’s mind although, of course, it can also be attacked by the way the range is displayed.
Technology may be able to help. If the figure for remaining charge is accurate, and it is possible to predict the future usage patterns (through the use of GPS and route history, perhaps), an accurate and steady figure for remaining range can be displayed. If the consumer develops confidence in this estimate, this should lead to better use of available charge. Alternatively, some sleight of hand in the presentation of reserves may help. If it is safe to discharge the battery down to 20% of capacity, for instance, would people be happier if 30% charge is shown as 30% rather than 10%? Somehow we have to make the public more comfortable driving with 20-miles of charge remaining, than they are in a petrol car with one gallon in the tank, or there will be a lot of unused battery capacity on the roads, and electric cars will be over-specified and more expensive than necessary.
Curb your velocity
Another aspect of electric cars is the very steep increase in power consumption with speed. Wind drag is an exponential resistance factor, hence Tesla estimates their Roadster can travel 240 miles on a full charge at a steady 55 mph, but only 200 miles at 65 mph and 170 miles at 75 mph. This will constrain long trips even more, as drivers seek to minimize the number and duration of recharging stops. Long trips at 80+ mph will be difficult to sustain, unless aerodynamic design can advance. Without new technology, the more aggressive among us may find this a source of significant frustration.
A lifestyle solution for long trips
Long-distance travel introduces two further effects, the need for more frequent recharging stops and a much longer wait during recharging. We can gain insight into long-distance electric travel from Tesla owners, the early-adopters of ‘new’ electric car technology. A splendid blog from the summer of 2009 recounts a Seattle resident’s trip home from San Francisco with his new Tesla. He completed the 850-odd miles in 3 days, including around 17 hours’ driving, while recharging overnight and mid-day at RV campgrounds with 240 Volt, 50 Amp charging stations. The approach was to recharge completely overnight, a 6 hour exercise at 40 Amps, to start out each day with a nominal 250 mile range. After driving for the morning, 80-160 miles in perhaps three hours, he would stop at lunch time for a few hours, partially or fully recharging to cover another 80-160 miles in the afternoon. On one very long day he recharged several times through a very long day, covering 440 miles but taking 20 hours to do so, including stops.
At 60 mph, a Tesla discharges the battery at about 15 kW, and when charging on a 240V, 50A (40A practical) circuit, it would replace energy at about 10 kW. Overall a Tesla driver would spend 50% more time charging than driving. This works well when the charging is overnight, but the 200 mile range (practical figure) would not allow a good daily range without that long lunch, and even then the 400 miles is less than a conventional car by perhaps 30%.
We conclude that, even with a large battery, long trips in an electric vehicle require the consumer to accept constraints, and suffer inconvenience. Recharging stops are more frequent and last considerably longer than for petrol cars.
The recharging stop
One possible technology solution would be a viable fast charging technology. If the four or five hours to fully recharge a Tesla could be reduced to 10 minutes, it would become comparable to the petrol alternative. The three techniques postulated for fast charging are battery pack replacement, using ultra-capacitors rather than batteries, and improved battery chemistry.
Replaceable batteries may become successful, but they require both engineering changes – the battery pack dimensions must be standardized, at a time when fitting the bulky items into the car is already difficult enough – and introduces other issues, as packs will still be very expensive, and have different value depending on their age and usage history. How will the driver know that the replacement battery he is given will not be on its last legs? Won’t this be the ultimate ‘market for lemons’? The approach from ‘Better Place’ is that the company will own all battery packs, charging the consumer per-mile for use, but this is a new and untested business model. The adoption of ultracapacitors is likely to depend on their cost and effectiveness under driving conditions, but if a combination of battery-ultracapacitor power has good performance characteristics, it would help with fast charging. And breakthroughs in battery technology for fast charging are announced every month, but none are yet commercially available, and again, cost and weight will be the primary drivers for battery adoption, rather than charging characteristics. Also, to deliver anything beyond 15 kW continuously, a charging station would have to include more than a conventional 240 Volt outlet.
Without the huge battery, impractical for cost and weight reasons, and without fast charging technology, we must recognize that the practical electric car will be able to travel only perhaps 200 – 400 miles in a day, and we will have to adapt to this constraint. A number of solutions have been proposed.
The easiest case: city cars
One is the focus on urban cars. Electric vehicles are much more energy-efficient than petrol at low speeds and when idling: the Tesla Roadster uses around 15 kW for a steady 60 mph, but only 5 kW for 30 mph , due to lower aerodynamic drag. Low speeds and limited range requirements about-town mean the battery can be smaller, and the car lighter and cheaper, and there is already a demographic trend to ‘town cars’ such as the Smart from Mercedes Benz, so one can argue that we are already on the way to changing consumer behavior and acceptance in this setting. Perhaps we only need a captivating design to jump-start this market, as the technology is already attainable. Not surprisingly, most of the major car manufacturers have shown concept cars in this category: during 2010 and 2011 we should see the first mass-produced offerings for urban drivers, and it will be interesting to see how far these cars can travel on a charge, as the designers will be keenly aware of the cost - range tradeoff. It is quite possible some of them will offer a 60 mile car for urban use, but doing so risks the consumer comparison-shopping and showing a preference for a 100 mile alternative from a competitor, should this exist.
(This trend is taken to extremes with the ‘neighborhood vehicle’, a legal category including golf carts. NVs are governed to a limited top speed, but are exempt from safety and other regulation. Other early approaches to small electric vehicles use loopholes to be licensed as motorcycles, avoiding much of the safety and other testing required for conventional cars. We can assume that most of these approaches are attractive to small entrants in the market with limited capital, but will disappear as the major manufacturers become involved.)
As one moves from the city to the suburbs, a habitual commute in the order of 10 – 30 miles each way or longer can be accommodated by the overnight-charge regime, with optional charging at work. But the conceptual challenges are greater than for urban cars. Because distances are greater, batteries must be larger and more expensive. But more than that, suburban driving does not follow the same route every day. Detours for shopping, ferrying children to school and sports events, doctors and other appointments, all extend the daily distance. However infrequent, it is the maximum daily drive that sticks in drivers’ minds, especially when making the purchase decision. A car will be of no use to them (they think) if it can’t stretch to their longest errand-running day. This means that rather than sizing for a 60 mile range, the ‘average’ daily suburban round-trip, it is probable that manufacturers will have to launch electric cars in the 100 mile category for suburban consumers, increasing their cost and weight.
It is widely accepted now that charging an electric car from a household electrical circuit is much cheaper than buying the petrol to power it. By way of example, a car that attains 20 mpg at $2.00/gallon costs $0.10/mile, while at 330 Wh/mile from the wall and $0.15/kWh it the electrical equivalent would cost half of that, or $0.05/mile. And of course the cost of maintenance and replacement parts for the electric car should be lower: no engine, gearbox (multi-gear anyway), exhaust system, catalytic converter, etc, and industrial electric motor technology is already extremely reliable and low-maintenance. Electric car proponents quote a cost per mile of 25% to 50% of a petrol car.
But there is one critical difference: the battery. Current electric car companies (Tesla is the only one in mass production where the customer buys the car and the battery) quote a 10 year or 100,000 mile life while retaining about 80% of capacity, but whether it’s 7 years or 12, in the mind of the consumer there will be a question of how soon and how expensive the replacement exercise will occur. Accounts of the Chevrolet Volt, a range-extended hybrid, suggest the designers devote significant testing cycles to verifying battery life over a range of conditions, so we should see credible figures from a variety of sources over the next few years, but it will take a while before the public becomes attuned to the different cost-of-ownership characteristics of the electric car. As with so many aspects of green energy (e.g. CFL or LED bulbs substituting for incandescents), a high initial cost is combined with low ongoing running costs for a low lifetime cost, but the high initial costs, and the likelihood that early failures will be heavily publicized, mean that the risk of premature battery failure and the high replacement cost will be an obstacle to purchase in the mind of the consumer.
Technology, perceptions and behavior
A few conclusions can be drawn from the discussion above. First, we can’t have electric cars that perform in every way as conventional cars do. The electric car owner will not be able to drive as far on a charge, or as long in a day as with a petrol car. So we must focus our design efforts on cars for those people who will value the distinct advantages offered by electric cars, while dealing with these drawbacks. This points the way to urban cars with light weight, good acceleration performance and low emissions – a seemingly winning combination, especially if the cost can come into line, or undercut conventional high-performance small cars.
For suburban drivers, the proposition is more nuanced and we need to segment carefully and consider where we can nudge behavior changes to cover the gaps. Families with multiple cars may be able to keep a long-distance hybrid, and add a daily electric commuter as their second car. Those who garage overnight are obvious targets, while cars parked on the street will be more difficult to recharge easily, and we may need to postpone our efforts to reach this audience. We have argued here that a 100 mile range is probably the minimum required to reach a significant segment of the mass suburban market, due to the worst-case-trip and limited-range anxiety effects.
And long-distance driving is the most difficult to accommodate. For those with time on their hands, a ‘drive for three hours, recharge for four’ regime can work well with extended lunch breaks, but type-A personalities will fret while waiting at the charging station. Battery-pack swapping is the only credible, albeit imperfect, technology for now. Alternative solutions include a load-when-required petrol-generator for a temporary hybrid car, or even temporarily renting an add-on battery pack for long trips, although this is but a partial solution given the size and weight of today’s batteries.
Even as we evangelize electric vehicles with habitual Silicon Valley optimism, these issues are real and will not disappear without a combination of new technology and new consumer habits. Of the two, habits are more difficult to change and must not be underestimated!
A few weeks after writing this, I came across a very comprehensive column in the Economist. It covers much of the same ground, but elegantly and succinctly: professional writers (and thinkers) are so impressive! The reader is invited to check it out.
This note covers some of the different options for architecting an electric car. At the ‘simple’ end of the scale are small-car conversion projects for the hobbyist, moving through concepts for larger and more sophisticated cars, until at the end of the note we discuss what can be done if a car is designed from scratch to utilize the advantages of electrical propulsion.
Simple home conversions
Converting a petrol-engined car to electric power is conceptually quite easy. Hobbyists report home conversion projects taking anything from a few weeks to a year or more, often including some custom fabrication – although some companies sell kits for particular makes and models, there doesn’t seem to be enough volume to standardize on either a unique year/model of donor car, or a standard conversion specification. Of course, even for the mechanically-adept these are not necessarily simple projects: the amount of ingenuity and perseverance required can be considerable.
The easiest conversion looks like the diagram above. Start with a light, aerodynamic car with a standard (manual) transmission. Unbolt the engine from the bellhousing and dispose of it. Similarly with all the associated equipment: petrol tank, radiator, exhaust system, etc.
Now fabricate a conversion plate to fit over the bell-housing and allow the new electric motor face to bolt to it, with the original flywheel or a planed-down replica on its shaft, connecting to the gearbox via the existing clutch assembly. The original gearbox will do the job it did before, even putting the car into reverse, so the electric motor only needs to run forwards. Since it has a much wider torque curve than the petrol engine, the gearbox is used much less – many owners say they only use second gear except for special situations like very steep hills.
The second major component is an electronics package to control the motor. With a DC motor up to perhaps 30 kW continuous rating, sufficient to power a small sports car, this is quite a small box and requires only fan cooling. The accelerator pedal is connected to a potentiometer modulating the controller, so all controls: gear stick, clutch, brake and accelerator are familiar.
The largest, heaviest and most expensive new component is the battery pack. Most home conversions now use Lithium-ion batteries of some type. Since these are bulky and weighty, even for a 50 – 100 mile range in a light sports car, it’s often necessary to use the trunk or rear seat, as well as engine-compartment space to house batteries. Sometimes the batteries are so heavy that springs and suspension must be stiffened to keep the car at the same ride-height as before. Over time, batteries will improve and these constraints will lessen, but the tradeoff between range on the one hand and the cost and weight of the battery pack will be significant for many years.
If you picked an old British sports car without power-assisted braking or steering, and without air conditioning, there’s not much more to be done; but if it’s a more modern car, like a Miata, these components will need to be powered, either by attaching a pump to the tail-shaft of the electric motor, or with separate electric motor assemblies. Similarly, cabin heating must now be electric, perhaps using a ceramic heating element.
Most conversions keep the original 12V battery to power ancillaries such as these, as well as lighting, horn, wipers, etc. This battery can be charged on a separate circuit, but most projects include a small DC-DC converter to take power from the main battery pack as required.
One item that can be difficult to work around is the ECU, the electronic brain box of modern cars. This has a large number of sensors, and when the petrol engine and other components are removed, it may be necessary to replace some sensors with components that fool the ECU into thinking it has a functioning engine still attached. The older the car, the less likely such a problem will be presented.
A charging unit for the main battery is usually kept on-board, although if the car were to be ‘sleeping’ at home every night that would not be necessary. But most owners like to plug in the big orange cable under the filler cap, and that normally involves 120V or 240V AC with an on-board charger.
Fixed-gear drive trains
The plan discussed above is very simple, but it leaves a few things in the car that are not strictly necessary. The diagram below shows how the clutch and gearbox (and prop shaft) can be removed, saving space and weight, and the electric motor mounted directly to the differential. Automatic transmissions are usually best removed in this way, because they are very power-hungry.
A few considerations make this a more complex project. Firstly, while it is quite possible to use a single gear ratio (the Tesla Roadster does just this), the gearing will need to be much higher than for a conventional differential: closer to 7:1 rather than 2:1 or 3:1. And since an electric motor extends high torque from a standing start, torque rating may also be a consideration. Indeed, the Tesla was delayed for several months by gearbox vendors’ delays, so this is not a trivial component. For a home conversion project, the car’s original differential will probably need replacing, so a search for alternative gearing will be necessary.
Also, most projects moving to this configuration would use an AC motor rather than DC. While DC motors are cheap and simple, there are many advantages to AC, mostly in control, reversing (again, without contactors) and thinner, more flexible wiring. But they are more expensive and, especially if larger than 30 kW, much harder for the hobbyist to purchase.
Perhaps the best aspect of AC motors is their ability to perform regenerative braking, where the motor is used to slow the car, producing electricity that is used to recharge the battery. Regenerative braking is well-established in mass-produced hybrids, but for the hobbyist there may be some difficulties in blending electric with friction braking, as the original brakes cannot be removed, and there should be a smooth transition between modes as the pedal is depressed.
There may also be a requirement for temperature control. All the power components – motor, controller and battery pack – are very efficient, but when dealing with power levels of 40kW or more, the level required to maintain a constant 70 mph in a Jaguar XJ8, even a 90% efficient power train will need to dissipate 4 kW continuously, quite a large amount.
Since there should be good air-flow at high speeds, the worst-case might not be high-speed travel, but with several passengers, accelerating up a steep hill into a head-wind on a hot day. Either way, there may be a need for cooling beyond an electric fan on the controller and battery pack. Some projects have used water cooling, plumbing from the original radiator, although the requirement is much less severe than for a petrol engine. It is to be hoped that ever-increasing electrical efficiency, wide-temperature-range batteries and good air-flow design will allow such complex cooling designs to be avoided.
The next stage takes us beyond the scope of a home conversion, although in the next few years we may see kits produced for this type of project. Wheel-mounted motors offer a number of advantages over a central motor on a differential.
One advantage should be simplicity. With no drive shafts and differential, it should be possible to reduce weight, and to avoid increasing unsprung weight. But the bigger improvement will be in control. Now each motor can run independently, allowing precise acceleration, braking, individual torque control on each wheel like a combination limited-slip differential and anti-skid brake system; all controlled electronically, by software. Even steering could be enhanced by driving the wheel on the outside of the curve slightly faster than the inside one.
Further, if the design is good, it will avoid the need for friction brakes on the driven wheels, as the motors will be able to do all the braking work. Motors for light vehicles already incorporate a parking brake (handbrake) function. More weight avoidance, more simplicity.
These are not entirely imaginary capabilities. Michelin has had its ‘Active Wheel’ in concept cars for a while, and promises production in 2010. Active Wheel includes a prime motion electric motor, regenerative braking and electronic active suspension in a single package, as a wheel hub.
Four-wheel motor drive
The final step of evolution is perhaps obvious. As George Orwell could have written, ‘Two motors good, four motors better.’
By applying four motors, one per wheel, we make each one smaller and lighter, and allow software to control many more functions. For one, we can avoid adding a friction braking altogether. But differential torque control allows all kinds of sophisticated acceleration and cornering enhancement, with symmetric advantages under braking. In effect, we have a drive-by-wire car for control purposes.
With this design and even the previous step, many conventional constraints on car design are lifted. Each wheel needs a motor unit behind it, but cross-body axles and prop shafts disappear. Batteries will still be bulky and heavy, but can be packaged to fit in different parts of the car; perhaps under the seats, or under the floor to keep the center of gravity low. The passenger compartment can be expanded, or alternatively the whole car can be shrunk, within an aerodynamic envelope of course.
This freedom to re-architect the passenger car that is one of the most exciting aspects of future electric vehicles.
Hybrids, fuel cells, battery-electric... so many choices! I wish them all luck, and the field is wide open. We need to discover what 'works' technically, what satisfies the consumer, and which of the consumer's habits can be changed. My heart is in the battery-powered all-electric vehicle, but even in that category there are many design choices to make.
And my motivation is 'energy efficiency' rather than overtly to save the planet, because it's the kind of thing that interests engineers.