Figure 2 - Arrays Charging at Sunrayce 99
In the design of a solar car, as in any project, the first step is to identify the car's purpose. For an American college, this will typically be to race in the biennial American Solar Challenge. This specific goal will shape the design priorities of the car and will impose certain constraints in the form of the race regulations. The regulations provide maximum dimensions, safety requirements, and performance requirements. These requirements may vary from year to year and from race to race. For example starting in 1996, the World Solar Challenge imposed a rule that all cars must be able to maintain a minimum speed of 38 km/h (23.6 mi/h), and Sunrayce 99 disallowed three-wheeled cars, which had previously been quite common.
In addition to these general requirements, most races have different classes in which cars can be entered. In the World Solar Challenge, although class definitions change from year to year, university teams, private teams, and corporate teams generally race in separate classes so cars in each class face similar competition. In the American Solar Challenge, where the entrants are primarily university teams, the distinction is between stock and open classes. In the open class, any battery and solar cell type can be used, while stock class cars are limited to commercially produced batteries and cells that are available to all teams. Additionally, in the stock class, a dollar/watt price cap is put on solar cells, and until 2005, only inexpensive lead acid batteries were permitted. As extremely efficient solar arrays built with space-grade cells can cost hundreds of thousands of dollars, the stock class allows teams with a limited budget to build and race a car without needing an extremely expensive solar array to be competitive.
Once the constraints of race regulations have been set, the type of race must be considered and an overall design philosophy for the car established. Aiming to build the car with the fastest 0-60 mi/h (97 km/h) time or best cornering ability might be appropriate for a track based competition like the Formula Sun Grand Prix, but is unlikely to be a successful strategy for a multi-day road race. Appropriate priorities and goals must be established, ranging from the car being under a certain weight, to being able to maintain a certain speed on solar power, to simply completing the race without breaking down. Many teams building their first solar car strive to make it the most reliable car possible. Robustness is emphasized over weight savings and critical components are over-built. This approach has served many teams quite well and the relatively brief history of solar racing shows numerous examples of extremely well performing cars having to stop to perform repairs, and ultimately loosing to less efficient, but more reliable cars. An alternate approach is to build a car that is as light as possible. Each component is pared to its minimum weight and every addition is carefully scrutinized. As the car's weight decreases, this process builds on itself, as a lighter car needs a correspondingly lighter suspension, further reducing weight. Extremely ultra-light cars can be very energy efficient but require careful engineering and extensive testing to ensure that they are durable enough to complete the race.
Ultimately, solar car performance comes down to efficiency. Only the energy stored in the battery pack at the start of the race and any additional power generated by the solar array is available to propel the car. While more expensive solar arrays will produce more power, total available power remains the limiting factor in how far and how fast a car can go. Solar cars are often described as operating merely on the power used by a household hair drier. Power output from a solar car's array in bright sun is usually between 0.75 and 2 kW and most races limit energy stored in the battery pack to 5 kWh.[11] Compared to the 32 kWh available from a single gallon of gasoline,[12] solar racing clearly operates on an entirely different level of power consumption and efficiency management from conventional automobiles. To make the best use of the limited energy that is available, aerodynamic drag and mechanical rolling resistanceare the two largest forces that must be minimized. There are also power losses in the motor, drive train, and electrical system.
At highway speeds, 50 - 75% of the total resistance experienced by a car is from aerodynamic drag (Tamai 1999); clearly then, the shape of the body is a primary factor in car performance, and it is also the area where improvements are most effective. Calculations given in the technical report from the 1996 World Solar Challenge show that reducing drag area by 10% results in a 3.1% increase in average speed, while reducing rolling resistance by 10% results in only a 1.4% increase in average speed (Roche and others 1997). As with all things, there are compromises, and an airfoil with minimum drag may need to be modified to take into account wheel locations, road visibility for the driver, and exposure of the solar cells. While the overall shape of the body is typically determined relatively early in the design process, numerous smaller elements - surface finish, wheel fairings, wheel well sealing, and protrusions such as mirrors - also contribute to aerodynamic efficiency.
The second largest power drain is rolling resistance: the result of friction between the tire and the road surface. This is minimized with special solar racing tires, now used by most teams, but suspension design and alignment also play a significant role in minimizing energy usage. The remaining power drains are fairly constant, or parasitic, inefficiencies. The in-hub motor, solar array power-point trackers, DC-DC converters, and even the batteries themselves are not 100% efficient. Every time electricity passes through one of these components, some power is dissipated as heat. For most electrical components, this unavoidable reality is dealt with by selecting components with the highest efficiency possible, but there are trade-offs. For example, in the wire connecting all of the electrical components, larger wire dissipates correspondingly less power as heat, but it is also heavier. As additional weight increases rolling resistance and requires more power to accelerate the car, increased weight must be balanced against decreased resistance to find the wire size with minimum power loss for the individual car.
As components are designed and selected, a power use equation for the car can be developed.[13] Using a computer model, the power consumed by each component of the car can be characterized. This allows analysis of the various components and investigation into which areas of the car are most likely open to improvement. Such a model is also extremely useful during racing to optimize performance and power consumption.
Solar car electrical systems are generally subdivided into two categories: high and low voltages. The high voltage components, those above 36 V, are the motor, batteries, solar array, and their corresponding accessories: controllers, relays, switches, fuses, and cables. All other components are low voltage and include both basic car functions - accelerator pedal, fans, lights, and horn - as well as data acquisition, monitoring, and control systems.
The power to propel a solar car is drawn from either the batteries or the solar array, which are both connected in parallel with the motor. If the solar array is producing more power than is currently required to propel the car, the excess goes to charging the batteries, while if there is insufficient sunlight to meet the motor's power requirements, the batteries will supply the difference. The motor is paired with a motor controller, which receives instructions from the low voltage control systems - forward/reverse, what speed to go, how much regenerative braking to apply - and then supplies the appropriate power to the motor. One of the unique features of the electric drive systems used in solar cars is their regenerative braking ability. A motor also functions as a generator and when desired, such as coasting down a hill or coming to a stop, the generated power can be fed back to the batteries while the resulting force slows the car.
Early solar cars made use of a variety of small high efficiency motors, typically DC brushless motors, coupled to a drive wheel by a belt or chain. As there are inefficiencies in such a transmission and maintaining belt or chain tension requires careful attention, few teams now use a power transmission unless the budget is extremely tight. At the 1993 World Solar Challenge, three teams independently developed an entirely new style of motor:[14] the in-hub, or wheel, motor. By taking advantage of axial magnetic flux, instead of the radial flux of conventional motors, the magnets can be shaped as thin discs. This allows the motor to attach directly to the wheel, often nesting entirely within it. Drive-train inefficiency is eliminated, but there are mechanical difficulties: the suspension must cope with the additional wheel weight; the motor must turn at the same speed as the wheel, which is often less efficient; and with some designs the entire motor must be removed to allow for tire changes. Most solar car teams now use in-hub motors, with the most common model being made by New Generation Motors (NGM).[15]
While there are only a few motors used by solar car teams, there are many more choices for batteries. In some of the first solar car races, there were few regulations governing batteries and teams were free to select both the type of battery and the quantity of them in the car. Balancing the weight of the batteries vs. the anticipated need to store energy led teams to use battery packs from 2 to 11 kWh. Experience showed that the additional flexibility of greater battery capacity was almost always advantageous, so, starting with the 1990 WSC, a limit of 5 kWh in batteries was established (Roche and others 1997). This limit has been used for most major solar car events since then, although given the difficulty in accurately measuring battery capacity, race regulations typically specify a maximum weight for each type of battery chemistry, determined by the average energy density of that chemistry.
Accurate specifications for batteries are difficult to find. The physical dimensions, terminal type, and weight are relatively straight forward, but the electrical characteristics are not. Manufacturers will typically state a nominal voltage and capacity, but to properly quantify a battery, several charge/discharge cycles need to be conducted and the variables plotted. The output voltage of a battery varies with state of charge and load and energy capacity is dependent upon the rate of charge. These factors are all temperature dependent to varying degrees, and there are additional characteristics such as terminal resistance and charge efficiency- the ratio of how much of the energy put in to a battery when charging can be recovered while discharging.
Numerous varieties of batteries are used in solar cars, with the most common being lead-acid batteries, the same type of battery found in nearly every commercial automobile. The only type of battery allowed in Sunrayce 93 through 97 and in the stock class of the American Solar Challenge through 2003, lead-acid batteries are inexpensive, reliable, and relatively easy to manage. As nearly all solar car teams choose sealed lead-acid batteries (SLA), either absorbent glass matt (AGM) or gel cell, the disadvantages of automotive flooded batteries - electrolyte spillage and hydrogen out gassing - are avoided, but lead acid batteries still have a very low energy density, typically around 30 Wh/kg. Until recently, teams seeking higher energy densities typically chose silver-zinc or nickel-zinc batteries, which have energy densities of around 125 or 60 Wh/kg respectively. While the energy density of silver-zinc batteries is very appealing, a battery pack can cost tens of thousands of dollars and lasts for only a few cycles.
More recently, most teams looking for mid-range batteries have chosen Nickel-Metal Hydride (NiMH)[16] which typically have an energy density of 70 Wh/kg. Although NiMH batteries have considerably higher energy density than lead acid batteries, they have a lower charge efficiency and some feel that this is significant enough to negate their advantage (Carroll 2003). The battery chemistry of choice for most current teams is lithium ion.[17] Driven by the rapid expansion of battery demand for portable electronics, the price of lithium ion batteries has fallen drastically, and some have energy densities as high as 180 Wh/kg.
The solar array is a network of small photovoltaic cells each converting the solar energy that it is exposed to into a small amount of electrical energy. The ratio of this solar energy to the produced electrical energy is the cell's efficiency. Solar radiation at sea level at midday is approximately 1000 W/m2, so one square meter of 15% efficient solar cells would produce 150 W of electrical power. An average solar car has 8 m2 of solar cells and might have cells from 10 to 25% efficient, giving an 800 to 2000 W solar array under optimal conditions. Cell efficiency usually correlates quite closely with cost, with the most efficient cells commanding a significant premium. Although there have been a few cases of teams associated with research universities constructing their own solar cells, nearly all teams purchase commercially available cells. Stock class cars are limited to cells valued at less than $10 per Watt and produced in sufficient quantities to be available to all teams.[18] These stock class cells are generally considered terrestrial-grade and are usually silicon cells mass-produced for commercial power applications. In the open class, without any price limits, teams generally try to obtain space-gradesolar cells that are designed for satellites and strive for maximum efficiency regardless of cost.[19] Currently most of these cells are based on gallium arsenide (GaAs) semiconductors.
Solar cells are most commonly manufactured from semiconductor wafers that are sliced from cylindrical blanks. To maximize the size of the cell that can be made from each wafer, cells are typically not true squares, but rather have clipped corners. When these cells are laid out in panels, there are blank spaces where these corners meet. For most commercial applications, this wasted space is insignificant, but with the limited area on a solar car, additional power can be gained by cutting off the sides of each cell, turning them into rectangles. Generally, teams choose not to cut their cells either because there is not enough time before the race to have them cut, or because of the additional cost. In addition to the cost of the cutting itself, considerably more cells need to be purchased. As each cell is now smaller, there are then more cells on the car, and there can be large numbers of cells broken in the cutting process. In the past, solar cells were principally laser cut, but the heat of the laser has been shown to slightly reduce cell efficiency, so now many teams have their cells saw cut (Rogow 2002).
Each silicon cell only produces around 0.7 V, or around 2 V to 3 V for multi-junction GaAs, so many cells are connected in series by soldered tabs to reach a suitable working voltage. These assemblies are referred to as strings or chains. Silicon cells are typically around 125 mmsquare and so chains of around 10 cells are usually a good working length and several of these chains will also be connected in series. Smaller GaAs cells are often formed into much longer chains. Most solar cells have their positive contact on one side and the negative on the other, so metal tabs are woven in between the cells to interconnect them. Some cells have been developed with both contacts on the rear of the cell, which adds a small amount of effective solar array area, since there are no metal contacts obstructing the top of the cell (Verlinden Sinton Wickham Crane and Swanson 1997). Sometimes a technique referred to as shinglingis used to achieve a similar effect. Shingled cells are cut down so that they contain only a single contact strip, which is then placed under the edge of the next cell. This is a very delicate process and the final assembly can be fragile, but is used by teams aiming for maximum possible power from their array.
Once the cells are connected in chains, those chains are then either directly mounted onto the car, or several chains are grouped together and encapsulated into modules. If individual chains are being mounted, they are usually secured with double-sided tape or a silicone elastomer, and covered with a conformal coating. Modules can be attached in a similar fashion, but already possess their own protective coating. Preassembling cells into modules better protects the cells and can provide for a smoother surface, as well as allowing the majority of array assembly to take place before the body is ready. Using modules also lends itself towards outsourcing, and many teams are now choosing to have one of several companies[20] catering to the solar industry perform the time-consuming soldering and encapsulation process and then constructing their array out of these panels rather than starting from bare cells. Constructing the array out of discrete modules also allows the possibility of later removing the modules and using them for a future car.
Once the chains are on the car, matching numbers of cells are then connected in parallel to a power-point tracker. A power-point tracker, generally referred to as a PPT or MPPT (maximum power-point tracker), is a highly-efficient intelligent DC-DC voltage converter.[21] Since the current-voltage curve of a solar cell is non-linear, with a sharp drop-off in current beyond the operating voltage, at a particular voltage and current, output power is maximized. The PPT loads the cells at this maximum (or peak) power-point and converts this voltage to a preprogrammed output voltage that matches the voltage at which the batteries and motor operate. PPTs can be either buck type, where the output voltage is lower than the input voltage or boost type, where the output voltage is higher than the input voltage. Boost PPTs are generally more economical then buck PPTs but are also generally less efficient. Additionally, buck/boost PPTs are also available that provide maximum flexibility in array configuration. PPTs are available commercially, but they are also a component that many teams choose to design and build themselves.
Chains of connected solar cells typically perform only as well as the weakest connected cell. Therefore, care must be given to how the cells are laid out. If one cell in a chain is shaded, the whole chain functions as if it is the shade. For this reason the array is usually divided into sections, each connected to its own power-point tracker, so that the shadowing of one section will not affect the others. The number of sections is usually somewhere between three and ten and is limited by the cost and extra weight of additional power-point trackers, the additional wire needed, and the need to have sufficient cells in series to reach a voltage in the PPT's operating range. Flat cars generally need few PPTs as there is less likelihood to be significant difference in the sunlight reaching different portions of the array, while steeply sloped cars with a center spine may have half of the array receiving direct sunlight while the other half receives virtually none. The outputs of all of the PPTs are connected in parallel and then connected to the battery and the motor.
In addition to configuring the cells into sections for each PPT, diodes are often added to minimize the effects of shading. In series connected chains, the current flowing through each cell is equal. Therefore, when one cell is shaded so that its ability to source current decreases below the level of the current forced through it by the other cells in the chain, that cell must then operate in a region of reverse voltage, bringing down the overall voltage level of the chain and overheating the shaded cell. To avoid permanently damaging cells from overheating and the temporary power reduction of that chain, bypass diodesare installed in parallel with the cells. When a cell is shaded, the diode is forward biased and conducts, bypassing the shaded cell so the series chain suffers only the 0.3 - 0.7 V drop of the diode. Bypass diodes can be installed on every cell, and this is often done on GaAs cells using small surface mount diodes that are installed inside the encapsulated modules,[22] but more often bypass diodes are installed in parallel with an entire chain. (Watkinson 2003b) The number of cells to bypass with each diode is a compromise that must take into account the additional time needed to install the diodes and the weight and complexity added by their wiring.
While bypass diodes are installed in parallel with series connected chains, blocking or isolating diodes are connected in series with chains that are connected in parallel. Unlike bypass diodes, blocking diodes are always in the circuit, so Schottky type diodes are generally used to minimize the voltage drop across them. In commercial installations, blocking diodes are used to prevent current from flowing out of the batteries and back into the solar cells over night when the cells are a lower potential than the batteries. This is not generally a concern in solar cars where the batteries are disconnected at night, and most PPTs also provide this isolation, but a similar situation can occur when chains are connected in parallel. If the voltage drops in one chain from a cell being shaded, this will also cause the voltage to drop in the parallel chains, reducing output power. If there is a blocking diode in series with the shaded chain, that diode will then be reversed biased and isolate the chain from the remaining chains.
The final element in the solar array is the array stand, also known as a charging stand. While driving, the solar array is attached to the car and its orientation cannot be changed.[23] However, most races also allow charging before and after driving hours, at which time the array no longer needs to be attached to the car and can instead be put on a stand and aligned with the sun. The power produced by a solar cell varies by approximately the cosine of the angle between the cell and the sun (θ). Additionally, due to the thickness of the atmosphere, the angle of the sun incident to the ground (φ) is also a factor. 
Particularly in the early morning and late afternoon, aligning the array with the sun can increase power output by as much as 30% (Carroll 2003).
While the high voltage systems of all solar cars tend to be rather similar, generally differing only in what type and brand of components are chosen, the low voltage systems allow far more flexibility. The low voltage elements of a car can generally be grouped into two systems: car operation and telemetry. The car operation components - accelerator control, lights, fans, horn - are essential to the car, while telemetry is composed of supplemental components to allow the driver and the rest of the team, via a radio link, to monitor the status of the car's performance.
Car operation components are required by ASC regulations to be powered from the car's main battery pack and not from an auxiliary power source, therefore a DC-DC converter is required. This device acts as a bridge between the high and low voltage systems, taking the 50-100 V of the primary batteries and converting it to low voltage, typically 12 V or 24 V. Many solar car teams use a rugged and efficient DC-DC converter manufactured by Vicor, Andover, MA. The output of the DC-DC converter is fused and split to several locations. It is directly connected to cooling fans for the batteries, which ASC regulations require to be operating whenever the car is on. Fans for power-point trackers and other electronics may also be necessary, as the interior of a solar car is generally extremely hot, which shortens the life of electrical components and decreases PPT efficiency. 12 V power is also supplied to brake lights and turn signals. A switch is mounted to the brake pedal so that power is applied to the brake lights whenever the mechanical brake is activated and a driver accessible switch is located to activate the turn signals. Simply having the driver toggle this switch may provide for the blinking of the turn signals or a small circuit may be fitted to blink the lights while the switch is depressed. Additionally, if an electric, as opposed to an air, horn is used, it will also be fed from the DC-DC converter via a driver accessible switch.
Various motor control functions are also part of the car operation low voltage systems. For simple analog control, two potentiometers and two switches need to be connected to the motor controller. One switch enables the motor and the other indicates the rotation direction, i.e. forward or reverse. Solar cars aiming for maximum simplicity have been known to eliminate these switches - wiring the motor to be always enabled[24] and going forward.[25]
The two control potentiometers are for acceleration and regenerative braking (regen) and they can be integrated into pedals or hand controls. Either linear or single-turn rotary potentiometers are generally used and the acceleration potentiometer is required to have a spring return to the off state. New motor controllers also offer more sophisticated digital control where the motor is controlled via digital codes instead of analog voltages. Some form of microprocessor communicates with the motor controller and provides it with the necessary instructions. The microprocessor may have generated these instructions from analog controls connected to it or a program may be run as cruise control providing optimum acceleration and regen values. If cruise control is used, the switch connected to the brake pedal is required to be connected to the microprocessor and disable cruise control whenever the mechanical brakes are activated. The microprocessor can also receive data back from the motor controller via the digital connection about the controller's status, such as speed, voltage, and current draw. Microprocessor based car control systems are often tightly integrated with telemetry systems and while more complicated and thus more difficult to implement reliably than analog systems, allow much greater fine control and awareness of the car's performance.
Telemetry, or data acquisition, systems provide data about the car's status. In a traditional automobile, the various gauges on the dashboard - speedometer, fuel gauge, warning lights - are a form of telemetry. Much of the telemetry data in a car will be displayed in real time to the driver in the form of LCD displays panels, LED readouts, or analog gauges. This allows the driver to optimize his or her behavior, but telemetry is also useful when looked at over time. Accurate data allows modifications to the car to be tested and problems to be anticipated. To accommodate this, some form of data-logger is typically installed in the car. In addition, the telemetry data is usually transmitted to a computer in the chase vehicle via a wireless radio link. This allows the computer to store the data and for team members in the chase vehicle to calculate strategy and advise the driver
In a solar car, the most important information is the voltage of the batteries. Voltage provides a rough measurement of the capacity remaining in the batteries, although battery discharge curves are non-linear and more sophisticated means are generally used to measure energy. However, voltage measurement is even more important from the standpoint of safety. Being discharged below a minimum safe point can damage all battery types. Further, many battery chemistries have the possibility of exploding if they are overcharged. Although many teams choose to use lithium-ion and lithium-polymer cells for their high energy density, they are quite dangerous if overcharged or overheated. Lithium batteries are required to have battery protection circuitry,[26] often supplied by the battery manufacturer, which monitors the voltage of each battery or group of batteries in parallel and the temperature of their enclosure. If safe limits are exceeded, the protection circuitry will open the relay that connects the batteries to the rest of the electrical system so that they cannot be further overcharged. As the protection circuitry measures battery voltage and temperature, it is usually connected to the telemetry system to make use of that data. However, if there is no protection circuitry, small wire leads will be connected to each battery allowing its voltage to be measured, whether by connecting these leads to analog inputs on a microprocessor or simply having test points where a hand-held voltmeter can be connected.
Current consumption is the second value monitored. Ideally, the current supplied from the batteries and the solar array as well as that being consumed by the motor and the DC-DC converter will each be individually measured. The DC-DC converter current is generally small and can be discounted; one of the three primary currents (batteries, array, motor) can then be inferred from the other two. The simplest method of measuring current is to place a shunt in-line with the connection being measured. A shunt is a piece of metal with a small resistance, typically only a few hundred micro-ohms, which has been measured with high precision. The voltage across the shunt is then measured and the current is determined by dividing that voltage by the known resistance of the shunt. Shunts are simple and accurate, but they are often rather heavy and some power is dissipated as heat by their resistance. In addition, each shunt adds two electrical connections where the cable needs to be terminated, inspected, and periodically tightened.[27] Alternatively, non-contact current measurement can be used. The most common of these are Hall-effect sensors.[28] Hall-effect sensors are lightweight and avoid having to splice a sensor into the high voltage electrical system, but they can be subject to electromagnetic interference from the motor and also need to be calibrated for the high temperatures in the car. Some recent motor controllers and power-point trackers monitor voltage and current internally and output this data as a digital value, eliminating the need for external sensors.
The current measurements collected by the above sensors need to be displayed to the driver so that he or she can try to minimize power drain through careful acceleration. In addition, by monitoring the current going into and out of the battery pack, an accurate measure of the pack's state-of-charge can be determined. Either analog circuitry or a microprocessor can integrate the measured currents to provide a functional 'fuel gauge.'
A third datum of interest is the solar car's current speed. This is sometimes available as a digital signal directly from the motor controller or it can be measured with a wheel mounted sensor. A small magnet or a reflective patch placed on one of the wheel rims and an appropriate sensor attached to the suspension will register each time the wheel rotates. Sensors of this type can be fabricated for use with a custom data acquisition system or purchased from bicycle shops as complete systems with a small battery powered LCD display. It is also possible to include many other sensors into a telemetry system, for example an accelerometer to measure the force the car is under or a slopemeter to measure the car's angle. The additional data from these sensors can be used to optimize racing strategy.
The mechanical systems of a solar car include everything that is not the electrical system, other than the outer aerodynamic shell of the car (the body), which is typically, but not always, treated as a separate system. The largest and most central of the mechanical components is the frame, which is the structure that holds and protects the driver and to which all other components attach. The remaining mechanical components are moving parts: wheels, suspension, steering, and braking.
There are two primary frame styles: space frame and monocoque. A space frame is constructed out of tubes and is complete in and of itself. Wheels, motor, batteries, and other electronics can be mounted on it and it can be driven as a car, essentially an oversized go-cart, without attaching the body. A monocoque, French for single shell, instead has the body and frame as a single integrated piece, so a single body/frame structure supports all of the loads in the car. Also possible is the semi-monocoque car, where the frame is a separate structure, but relies on its attachment to the body to provide additional strength. This term is applied both to cars with a metal frame reinforced by its attachment to a body intended as a structural member and to those cars where the 'frame' is composed of composite panel bulkheads.
Space frames are generally assembled from welded metal tubes. Each element of the frame is selected to balance the necessary strength and weight, by choosing the material, shape, and size of the tube. Many different materials are used in space frame construction, although perhaps the most common is SAE J404 chromium-molybdenum alloy steel, typically referred to as 4130 steel or just chromoly. This steel alloy is often used in bicycles and go-carts and can be cut, machined, and welded with relatively little experience or specialized equipment. Teams seeking a higher strength to weight ratio have also used common aluminums such as 6061 and 7075, ceramic-matrix reinforced aluminum, titanium, and carbon fiber tubes, like those used to make high-end golf clubs. In order to be used safely, these materials require a person with considerable experience using that material and a specialized shop to work with them. Choice in material also has an impact on field repairs during the race, as chromoly can be brazed with a torch, while aluminum and titanium require bulky gas welders.
The tube shapes used in most solar cars is divided between rectangular and round cross section tubes. When constructing the frame, rectangular or square tubes are considerably easier to cut and weld at angles and provide ease in attaching brackets and flanges for other parts to attach. Round tubes however, are stronger by unit weight than rectangular tubes, so the completed frame can be lighter. This advantage of round tubes over rectangular ones is offset by considerable drawbacks in the construction process. Every joint between two round tubes, even one as simple as two tubes meeting at a right angle, needs to be cut with a hole saw of the same outer diameter of the tube to form a curve in one tube, allowing it to sit flush against the other tube. Nodes where several tubes meet, often required for strength, are even more difficult to construct.
The size of a tube is specified by its outer diameter (O.D.) and its wall thickness, which is the difference between the outer diameter and the inner diameter (I.D.). A tube's strength is primarily proportional to its outer diameter, but larger tubes are also correspondingly bulkier, heavier, harder to cut, and more expensive. One way to reduce the weight of large tubes is by decreases the tube's wall thickness. Shrinking the wall thickness does little to effect the primary strength of the tube in compression, tension, or bending, but does greatly decrease the tube's resistance to buckling - the tube collapsing in on to itself from a point load. With proper design, buckling forces should not occur in responses to the normal stresses of solar car use, such as tight turns or pot holes, but in an accident they could be a problem, requiring caution when reducing wall thickness.
Instead of the metal tubes of a space frame, monocoques are primarily constructed from the same composite materials as the body of the car, generally in sheet form, and require very different design and construction techniques then working with tubes. The simplest monocoque design is the tub, which greatly resembles its namesake, the household bathtub. By forming composite material into an open-topped, round-cornered box, the material supports itself and creates a strong, but lightweight central compartment for a car. The bottom of this box also serves as the bottom of the car and a primary structural member. Suspension and other components are typically attached to the composite structure by bonding metal connection plates to the car. Composites are also often supplemented with metal tubing to create the roll cage.
One of the largest drawbacks of composite construction however, is the difficulty in properly analyzing it. Metal tube construction has been used for centuries and numerous techniques have been developed to predict its performance. Calculating forces on beams is a relatively straightforward physics problem and software packages, including several open-source options to analyze complex structures are readily available. Composites however, are a relatively recent innovation and depend upon large surfaces bonded together, which require more complicated calculations to analyze. The strength of composites is also rather variable, depending upon the conditions of the materials, the resin used and how well it is distributed, as well as under what conditions - temperature and pressure - the part is cured. These same factors also affect the weight of the finished part, and attempts to save weight by using the minimum amount of resin may impair the part's strength.
Perhaps the most important mechanical decision is the layout of the wheels. While some solar cars follow the traditional four-wheeled automobile plan, many teams pursue a three-wheeled design. Eliminating a wheel removes one-quarter of the cars un-sprung weight and reduces rolling resistance. However, having only three wheels can lead to stability problems, particularly during an accident or tire-blowout. A traditional tricycle has a single front wheel for steering and two rear wheels. The most common three-wheeled design in solar cars however, is the inverted tricycle or tail-dragger. This has two widely spaced front wheels, which are used for steering, and one centered rear wheel, which is typically the drive wheel. A variant on this has four wheels, but the two rear wheels are very closely spaced. By sharing some of the rear suspension, two closely spaced rear wheels eliminates some of the weight of a true four wheel design while still meeting potential race requirements for four wheels.[29] Such a design is still heavier than a normal three-wheeled tail dragger and also lacks some of the advantages of normal four wheeled vehicles - such as having the rear wheels inline with the front wheels, reducing drag.
As the majority of aerodynamic losses occur at the front, or leading, edge of the car, there is a potentially significant reduction in drag by having only a single wheel obstructing airflow at the front of the car instead of a pair. Team Lux's computer analysis of the John Lee showed that a tricycle configuration had 15-20% less drag than a tail-dragger (Johnson 2002b). With only a single front wheel for steering, there are also potential weight reductions. Despite this, the tricycle design has rarely been implemented in a solar car due to definite stability problems in single wheel steering at high speed and braking concerns as the majority of braking is done with the front wheels. Having only a single front wheel cuts front braking performance in half, and if the front wheel also contains an in-hub motor with regenerative braking, mounting an additional mechanical brake can be quite difficult. These problems have been overcome, such as in the cars of the Aurora Vehicle Association, which had been quite successful in several World Solar Challenges,[30] but a successful tricycle design is a significant engineering challenge.
Regardless of the number of wheels, they are the interface between the car and the road and their selection is quite important. A wheel is composed of the rubber tire, possibly with a pressurized inner tube, and a rim that supports the tire. This combination must be lightweight, able to withstand the stresses of highway driving, and should have the lowest possible rolling resistance. While many early solar cars used bicycle components to meet these needs, they were generally not durable enough for extended high-speed use. Tires designed for the front wheels of dragsters have also been used, but these are rather heavy and have a higher rolling resistance than is desirable. A few tire manufacturers introduced products specific to solar racing, and now most teams use either Bridgestone's Ecopia tire[31] or Michelin's solar racing tire.[32] These are both thin, high-pressure[33] tubeless design, resulting in extremely low rolling resistance.
These specialized solar racing tires require corresponding specialized rims to mount them. While the Ecopia tires were distributed through NGM, who manufactured a matching machined aluminum rim designed to interface with their in-hub motor, mounting the Michelin tires is not so simple. Commercially available dragster and go-cart rims and hubs exist, but these are usually quite heavy. Many teams either purchase expensive carbon fiber rims from GH Craft in Japan or have their own rims custom machined, often from magnesium.
Once the wheel configuration is determined, suspension needs to be designed to hold the wheels in the correct position. Suspension is perhaps the most complicated of the mechanical systems, as it must allow the wheels to spin freely while transferring force from the wheels to the frame, absorbing shocks from bumps in the road, and providing for mounting braking and steering components. In addition, the suspension should be adjustable, so that the wheels can be properly aligned with the road and the body's angle of attack can be tuned for best aerodynamic performance. While accommodating all of these constraints, the suspension must also allow a blown tire to be quickly and easily changed, so little time is lost during the race.
Among the variables determined by the suspension is the car's travel. Travel is the distance above and below the car's resting position that the car can move in response to bumps and holes in the road. This is determined by how far the spring or shock can extend. How much force it takes to achieve this motion is the suspension's stiffness. A great deal of suspension travel, several inches, gives the car a very smooth ride, but energy is expended by the car moving up and down and oscillating about its rest position. Conversely, on a car with very little travel, or very stiff suspension, the entire car moves in response to a bump. While this prevents energy from being dissipated in the suspension, it also subjects the entire car to increased shock loads and raises the possibility of damage, as well as the moving body presenting a larger area for aerodynamic drag.
In aligning the suspension, the three primary factors are caster, camber, and toe angles. The caster angle is the angle from vertical of the wheel's support. For example, the front wheels of a shopping cart have a very high caster angle, with the wheel essentially dragged behind the attachment point at the front of the cart. This leads to the wheels being pulled in to a straight line and reduces any tendency to curve, but the wobbling of the wheels as they are dragged into the straight path is wasteful of energy. A wheel's camber is its offset from vertical when viewed head-on. Finally, the toe angle is any deviation from straight ahead when viewed from above. Any toe angle is inefficient as the tires cannot roll straight ahead, but to some degree must be dragged across the difference in angle. Despite this, a small toe angle is sometimes necessary to achieve sufficient cornering performance. Additionally, as the suspension is not a static item, all of the angles must be evaluated throughout the range of the suspension's travel.
Many different suspension designs have been used on solar cars, often adapted from regular automobiles, motorcycles, and even bicycles. The most common configuration however, has been parallel a-arms, also known as double wishbone suspension, on the front wheels and either additional a-arms or trailing arms for the rear wheel(s). A-arms are actually composed of a pair of "Y" or "V" shaped pieces, placed parallel with the road and stacked vertically. They each connect to the frame at both of the legs at the top of the 'letter' and the single point at the base of each letter is joined with a vertical piece, called the upright or kingpin. The upright typically continues past the lower a-arm to a point where the wheel attaches. A shock absorber attaches to the base of the lower a-arm back to a point on the frame in-between the arms of the upper a-arm. All of the joints are designed to allow the connected structure to pivot up and down, ideally with the road contact patch at the base of the tire moving vertically with the suspension travel. Any side-to-side movement of the tire, or scrub, wastes energy; this is minimized by carefully adjusting the length of the a-arms and the angle of the upright, typically using computer modeling.
For the rear wheels, without the complications of steering, simpler trailing armsare often used. These are large tubes or trusses of smaller tubes that extend back from the frame to a wheel and pivot about their connecting point to the frame. Typically, there may also be a diagonal support element to help control lateral loads that occur in turns. As with a-arms, a shock absorber, connected to the arm near the wheel, connects to the frame to absorb energy and limit travel.
Steering on a solar car is typically straightforward, with the two front wheels connected by steering arms to a rack and pinion gear that is turned by a steering wheel. There are many variants to this however, with teams often substituting cables for the rack and pinion in order to save weight. With the cramped conditions in the cockpit of a solar car, some teams also omit the steering wheel, which is often in the way of a quick emergency exit from the car. Push-pull steering, similar to that on a bicycle, can be used with handles to either side of the driver directly connected to a single front wheel or as part of a cable steer system.
Teams will typically try to minimize the range of wheel motion required for steering, as a greater range of wheel movement requires larger fairings to enclose the wheels, which can significantly increase drag. This minimization is limited by race regulations specifying minimum turning requirements, typically in the form of a slalom or figure eight course that the car must navigate. Some teams have also tried four wheel steering, which requires less movement of each wheel, although this prevented the University of Michigan from qualifying for ASC 2003, as their remote control system for the rear wheels was not responsive enough to navigate the slalom course at the minimum speed.
The last major mechanical component is braking. The motor's regenerative braking is used to slow the car, but is generally insufficient to stop the car, so mechanical brakes are required. Current ASC regulations require a redundant mechanical braking system, either separate independently operated systems for the front and rear wheels, or dual systems in either the front or rear. The brake handle or pedal and the brake disks can be shared between the two systems, but separate calipers are required and hydraulic systems must have separate master cylinders (ASC 2003). Hydraulic disk brakes are the most common type used in solar cars. Generally go-cart or motorcycle sized brakes are chosen, but for very light cars high end racing bicycle components are sometimes used. Most commercially available brakes do not fully retract the brake calipers, but leave them lightly touching the brake disk for a quick response. This wastes energy however, so calipers need to be chosen that can be adjusted to fully retract the brake pads.
Beyond the simple fact of having solar cells, what separates solar racers from other electrical vehicles is their bodies. The body of a solar car serves two purposes: aerodynamic streamlining and holding the solar array. To maximize power from the sun, the body is made as large as possible, up to 2 m wide and 6 m long, larger than most pickup trucks. Even this much area provides very little power, so extreme aerodynamic optimization is necessary. To create such a body, advanced composite construction techniques are required.
Although aerodynamics is arguably the most important aspect of solar car design, it can also be the most difficult. Most team members will not have studied the fluid dynamics and advanced mathematics necessary to solve aerodynamic problems and much of aerodynamics can be counter-intuitive. Perhaps because of this, new solar car bodies are usually iterative rather than revolutionary, copying a great deal from previous successful designs.
Some of the earliest solar cars were essentially go-carts with a large flat solar array suspended over them. As designs evolved, and efficiency goals increased, solar arrays were integrated into the surface of the car body and aerodynamics became king. Many unusual shapes have been tried, but two shapes have become increasingly common. The first, the bubble canopy style car, has a thin body and a large flat top for the solar array, with room for the driver being provided in a clear bubble that gives this style its name. An alternate body shape, the central spine - sometimes referred to as a Manta type car, after MIT's Sunrayce 95 winning solar car - features a thicker body with a sloped front, allowing the driver to fit entirely inside the body, followed by a central spine and sloped sides running the remaining length of the car. While there is some debate about the respective aerodynamic benefits of each design, computer analysis conducted in 2002 by Team Lux to evaluate body shapes for the John Lee showed that traveling at 65 mi/h (105 km/h), the central spine design was subject to 30% less drag then a comparable bubble canopy car. Although crosswinds exerted 27% more net force on the spine than the bubble canopy, the center of pressure of the force was aligned with the car's center of mass, resulting in 84% less torque (Johnson 2002a). The high torque exerted by crosswinds could be potentially dangerous to a bubble canopy car. Additionally the bubble canopy was subject to significant lift, while the spine was almost perfectly neutral. Despite these deficiencies, a bubble canopy design allows simpler array construction and layout, more uniform exposure of solar cells to sunlight, and better driver visibility.
One of the principal ways to reduce drag is to add wheel fairings, also known as spats. Essentially an airfoil constructed around each wheel, fairings reduce turbulence and can significantly improve the performance of a car. They require careful design however to ensure that they do not interfere with the motion of the wheels and that they are actually functioning as expected. Several strategies have been used to fully enclose the wheels over their full range of motion without having overly massive fairings. The simplest is to use half fairings, or fairinglets, which are placed only in front of and behind the wheel, with the sides open. This greatly simplifies construction while still providing most of the aerodynamic advantage of full fairings. Another option is to construct fairing doors, which swing open when the wheel turns, but provide a full enclosure when car is traveling forward. The most complicated solution, but which can provide the optimum performance, is fairings that rotate with the wheel. By securing the fairings to the suspension rather than the body, the fairings are always oriented with the direction of travel and provide the maximum aerodynamic benefits. A variation on this is to also have the fairing move up and down with suspension travel, which allows the fairing to extend almost entirely to the road surface, but creating a smooth seam between the moving fairing and the body is quite difficult.
Aerodynamic optimization is not the sole constraint that defines the body shape, driver visibility and egress, suspension travel, access to the interior for maintenance, and the solar array all also need to be accounted for. Accommodating the solar cells typically leads to the largest tradeoffs with aerodynamics. The flowing curved surfaces of an airfoil conflict with the flat outline of a solar cell that is most efficient when oriented perpendicular to the sun. Most cells will only bend slightly, if at all, and are in danger of breaking if over bent, so large relatively low curvature areas are required for the solar cells. Another subject of much debate is where to place the seams in the body. Many teams simply place the seam, where it results from construction of the body in two halves - typically at the widest point on the car. This requires no additional work, and as there is a stagnation point right at the nose of the car, this arguably places the seam where its roughness will have the least affect on aerodynamics. Other teams try to avoid having a seam at the nose at all, either by having the section of the top containing the array be removable, or by moving the seam to the underside of the car.
Solar car bodies are most often constructed from composite panels, which are a multi-layer sandwich laminated together with epoxy. The inner and outer skins are sheets of synthetic cloth, usually Kevlar®, an aramid fiber developed by DuPont; carbon fiber, also known as graphite; or fiberglass. The filler material between these skins, known as core, can be virtually any material, from foam insulation or plywood to solid aluminum sheets. Many solar car teams use a honeycomb structure made out of Nomex® paper, another DuPont aramid, which has been dipped in resin. By themselves, the materials of a composite panel are usually flexible, so they can be conformed to curves, but once they are laminated together, a very stiff, lightweight panel is formed.
There are two basic processes for making solar car bodies: assembling them out of prefabricated flat panels or laying up custom bodies in a mold. Flat paneled bodies are cheaper and faster than mold based bodies, but have many aerodynamic drawbacks and are typically used only for a team's first solar car. Most composite bodies are made in a mold through a process called lay-up. There are also two primary methods for this: wet lay-upor dry lay-up, also known as prepreg. In wet lay-up, the skin material is laid into a mold and the liquid resin is worked into it before the core is placed on top of it. In the prepreg process, the skin material has been pre-impregnated with resin at the factory. It is laid into the mold dry with the core and the resin is later released by heat. Prepreg materials are easier to work with and are significantly less messy. They also allow a longer working time to correctly position all of the materials. Liquid resin will start to harden in as little as twenty minutes, so performing wet lay-up on an entire solar car body is usually quite a race. However, prepreg materials are more expensive and need to be kept frozen when not in use, raising transportation and storage costs. Surplus 'expired' prepreg material can often be obtained inexpensively, since prepregs are widely used in the aerospace industry, which has strict regulations on the length of time that the material can be stored. As solar car bodies do not require the same strengths and safety factors of airplanes, the use of expired prepregs is not usually a problem, but the obstacle that keeps most teams from using prepregs is the need to bake the entire part (the sandwich of skin and core) in order to activate the resin. Access to an oven or autoclave large enough to hold an entire solar car can be prohibitively expensive if one is available at all, although some teams have managed to construct their own ovens.
Regardless of whether a wet or dry process is used, the basic steps are the same. The mold is prepared with wax or a special release agent to keep the part from adhering to it and then an appropriate skin is carefully laid into it. This skin will form the outer surface of the car. During wet lay-up, resin is worked into the skin. Pieces of core which have been cut to shape are laid on top of the skin, followed by breather cloth, which will absorb the excess resin released while the part cures, and then vacuum bagging, essentially a thin plastic sheet. The vacuum bag is sealed around the edge of the mold with a double-stick tape known as press-tite and inlet ports are installed in the bag. These are then connected to a vacuum pump and the part is then left to cure under vacuum for an extended period. If prepregs are used the mold is also baked at this stage. In an autoclave, this might be for only a few hours, but for wet lay-up at room temperature, the part may need to cure for 24 hours or longer. After the part has cured, the vacuum bag and breather cloth are removed and discarded and the part is checked to see if it cleanly released from the mold and if the outer skin fully laminated to the core. The part is then returned to the mold and the lay-up process is repeated for the inner skin. Once the inner skin has cured, the outer shell of the body is essentially complete. Mold imperfections may result in some defects in the body that need to be filled and sanded, but with a well prepared mold, the part should be ready for painting once it is released.
The main task in body construction is the creation of the mold, also known as a tool. The mold is a full-scale inverse image of the solar car body, and is often referred to as 'female,' as opposed to the 'male' part that fits within it. Typically, a solar car body is made from two molds, one for the top half of the car and one for the bottom. The created parts may then be permanently joined together and access panels cut into the body or the mold line may be used as the opening point of the body and the two parts are secured with temporary fasteners. Additional molds are sometimes used for faring and canopy parts if their curvature would make the main molds difficult to construct or release.
Commercial molds for mass-produced parts are often made of metal, typically steel or aluminum depending upon the required durability, and are directly cut on CNC machining centers.[34] This is extremely expensive and solar car molds generally do not need the durability of metal as they are used only once or twice. Instead, the molds are often made out of fiberglass, which has been sprayed over a male plug. The plug is a full-scale replica of the desired final part, and when the fiberglass is removed, the desired inverse mold is created.
These plugs are typically made of foam and the optimal way to create them is to have them cut by a CNC machine. They are cut from a blank, which is block of high-density foam large enough to encompass the whole plug - either obtained as a single piece or assembled by gluing together smaller pieces - which is then formed into the appropriate shape by a high speed cutter guided by a tool path file that has been made from the team's CAD design for their body. This process requires minimal labor and produces an extremely accurate plug, but can be expensive if a sponsor is not found who will not donate the machine time. If the CNC machine is not local, which is often the case, shipping the plugs, which can weigh thousands of pounds, will typically require a tractor-trailer, adding time and expense to the construction process. If a machine large enough to handle the entire plug cannot be located, it is also possible to assemble the plug from halves or quarters that are cut on a smaller machine.
It is also possible for teams to manufacture their plugs by themselves. One method is to assemble the plug from a stack of thin sections. Numerous sections are laid out on sheets of foam, each the full width and height of the car, but perhaps only an inch of its length. These are then cut out and glued together, creating the profile of the car. The more sections that are used the more accurate the plug, but the more time consuming the assembly. To create an accurate profile, the sections must then be sanded smooth, which is a process that can introduce considerable inaccuracy. The layout and cutting of the sections by hand is also a source of inaccuracy. However, as these are considerably smaller parts than an entire solar car, it may be easier to find a CNC machine capable of cutting these sections, than one large enough for the entire car. As an alternative to cutting all of the sections of the car, it is possible to cut only selected sections out of a stiffer material than foam, such as plywood, omitting the sections where no change of curvature occurs. Thin sheet stock can then be applied over these sections creating the surface of the plug. While this eliminates the sanding of foam sections and results in fewer sections to cut, it is difficult to accurately position the covering and the use of a solid surface covering limits the choices of curvature that can be used.
Whether machine or hand made, the most important step of plug construction is preparing its surface for the mold. Any imperfections in its surface will be transferred to the mold and then to the solar car body. Smoothness is important not only to the final aerodynamics of the final body, but it is essential for its creation. Imperfections create places where the fiberglass covering the plug will adhere, preventing the mold from releasing from the plug. The plug generally receives as many as a dozen coats of wax before fiberglass is sprayed over it to ensure that the fiberglass mold will release from it. In addition, depending upon the plug's material, the plug may need to be protected from the fiberglass spraying process. Fiberglass spraying is an aggressive process that can eat away at a soft mold and a part to be sprayed is often cleaned first with acetone, which reacts with many types of foam. One method Team Lux used to protect a foam plug was to cover it with a thin layer of epoxy resin; however, the resin itself would have damaged the foam, so in addition a thin layer of sheet rock mud was added over the foam before the resin was applied. The sheet-rock mud had the added benefit of being easily sanded to form a smooth surface.
The primary factors driving material choice for solar car teams are typically cost and availability. For the skin, there are three main choices: fiberglass is the least expensive, followed by carbon fiber, and then Kevlar. In addition to its low cost, fiberglass is also widely available from marine suppliers, but it has the lowest strength/weight ratio and is therefore least desirable. Carbon fiber parts are extremely strong and stiff and have very good working qualities as they drill well and can be sanded, however carbon fiber is conductive which can cause shorts and even fires when used near solar cells. Kevlar is the strongest of the materials, but it is not as rigid as carbon fiber and finished parts do not sand or drill well, however it is non-conductive making it the material of choice for array panels.
Core selection is generally simpler, as a paper honeycomb is almost universally used in solar cars, although the honeycomb can also be constructed from plastic and, in extremely high-strength applications, aluminum or even carbon fiber honeycomb is sometimes used. Some teams have bodies very well supported by the frame and choose a foam core which, while producing a heavier and less stiff composite than honeycomb core, is considerably cheaper and requires less skill to ensure a good lamination. Beyond material, the two primary variables in core are the thickness, which largely determines the strength of the final body, and the density and shape of the honeycomb pattern, which determines the flexibility of the core. The largest supplier of core, as well as skin material and fully assembled flat panels, is Hexcel Corporation, Stamford, CT, which has donated or discounted body materials to numerous solar car teams. Hexcel's HRH-10 Nomex core, used by many teams, is available with standard hexagonal cells, OX-Core® (over-expanded) cells which bend easily in one direction, or Flex-Core® which supports compound curvature (Hexcel 2005).
The ability to store energy in a set of batteries is what makes strategy necessary in solar racing. Without batteries, each car would simply use all of the power available from their solar array to travel as fast as possible and the race results would be determined by the design of the cars and the vagaries of the weather. By adding energy storage, the team must decide whether to store energy or to use energy reserves for speed. Strategy during a race can be as simple as driving the speed limit until the batteries are drained and then stopping to recharge with the solar array, but typically a more thorough and complex approach is used.
The optimum speed for the car to travel must be calculated based upon its power use characteristics. Given prevailing conditions - solar flux, wind speed, altitude, road grade, and road surface - the car's power use equation will calculate how much power is needed to achieve a given velocity. If the car's telemetry is in working order, the state of charge of the batteries will be known, and predictions can be made about how far and how fast the car can travel. Many teams develop computer-modeling programs that use data on the race route from GPS surveys and input from the car's telemetry to make recommendations on optimum speed. Some of these models are detailed to the point of timings of traffic lights and locations of rail crossings and cattle grates.
Anticipating how much sunlight will be available for charging before, during, and after racing hours guides how much of the energy stored in the batteries can be safely used for a day of racing. This makes weather forecasting an extremely important part of race time strategy, and many teams will have historical weather data for the route on hand, coupled with instrumentation in the chase vehicle and a satellite uplink to professional weather services. Knowing whether the sun that is currently out needs to be taken advantage of by stopping and charging, or whether to push on and hope for sun later, can be the difference between winning and loosing. Accurate predictions can be crucial, as was shown in the first World Solar Challenge, where GM raced all out on the first day, draining their battery pack. The lead they gained however, was sufficient to move them ahead of a storm system, allowing them to proceed for the rest of the race in sunlight, while teams further back on the race route were plagued with thunderstorms and hail (Tuckey 1989).
Race performance depends upon considerably more than just the car. Solar racing is a team sport and participating in a race requires people from drivers to tire changers. A car does not win a race, but rather a team does, which requires having both a high-performance car and highly skilled personnel. Before Hans Tholstrup founded the World Solar Challenge, he was involved in fuel economy trials and showed that driver skill could result in as much as a 30% reduction in fuel consumption (WSC 1998). Beyond this direct efficiency gain, the success of a team relies on the smooth interaction of its members. They must be thoroughly trained in the normal operation of the car, prepared for unexpected contingencies, and sufficiently supplied to maintain the car and support themselves for the duration of the race.
It has been shown time and again that the most successful teams drive hundreds of test miles before a race (King 1993). While this certainly helps uncover technical problems with the car, it also prepares the team itself. Solar racing is a foreign experience to nearly all college students, with very few having experience in convoy driving, energy optimized driving, radio communications, or Formula-1 style pit stops. Obviously, the driver of the solar car needs time to become acclimated to an unusual vehicle, but the importance of training and experience for all of the other members of the race team is often overlooked.
Dealing with confusion of intersections, heavy traffic, and aggressiveness is instinctive to most drivers, but to a solar car convoy, these elements provide unique safety and energy management challenges. The first responsibility of a team on the road is safety. Although the race regulations set out minimum safety requirements, such as a fully encompassing roll cage, crush zones, helmet,[35] and a five point harness, a conscientious designer goes beyond these minimums. However, it is unlikely that a solar car will ever be as safe as a normal car with its much greater mass and additional safety features such as airbags. It is therefore best to avoid an accident in the first place, which is one of the prime roles of the lead and chase vehicles.[36]
As solar cars are quite low to the ground, it is difficult for other drivers to see them. Additionally, their appearance is quite different from ordinary cars, so when they are seen, other drivers often behave in unpredictable ways. To mitigate this, the lead and chase vehicles, often vans, are equipped with flashing orange lights and serve both to alert other drivers on the road to the presence of the solar car, as well as to keep them away from the solar car. Maintaining an evenly spaced convoy is a tricky skill requiring practice and coordination. It is quite frustrating, and wasteful of energy, to have the lead vehicle go through a yellow light and then the solar car and chase vehicle forced to slam on their brakes to avoid oncoming traffic when the light changes to red. The lead vehicle, although responsible for the navigation of the convoy, must allow itself to be coordinated by the chase vehicle so that the solar car is not wasting energy trying to catch up with the lead vehicle. In some situations, such as unprotected left turns, the chase vehicle might stop in the intersection to provide a barrier while the solar car passes and then overtake it to resume normal operation.
To ensure smooth operations, it is usual for there to be radio communication between the solar car and team members in the support vehicles. Many teams set up a dedicated radio link between the solar car driver and one handler, so that the solar car driver is not overwhelmed with extraneous communication. A separate radio channel is then used for coordination among the rest of the support vehicles. ASC regulations also require each vehicle to have a CB radio for inter-team communication, to facilitate such matters as passing.
The only way to perfect racing procedures is through a great deal of practice and preplanning for possible emergencies. It is equally as important to be prepared for routine situations, as any unnecessary delays cost race position. The most common delay is in tire changing, which should only take minutes with a well-practiced team, but often takes a half hour or even longer if replacement tires are not ready to be attached or access to the suspension is obstructed. Many teams largely avoid this problem by installing a fresh set of tires for every day of the race, but this can be an expensive strategy. Nevertheless, it is thoroughness to this degree combined with proper planning, such as pre- and post-race day checklists, that leads to an effective race team.
Solar cars are expensive. Some have budgets in the millions of dollars, and while a few are built on shoestring budgets for only a few thousand dollars,[37] most will cost at least $150,000. While some cars are built by large corporations (most notably General Motors and Honda), and others are built by teams of private individuals, in the United States nearly all cars are sponsored by academic institutions. The sponsoring institution is typically a source of significant support to a solar car team, in the forms of technical advice from faculty, workspace, and direct financial support. For Sunrayce 90 through 95, all teams had received some funding from the race sponsors, but fundraising has always been a necessary component of any solar car program.
Seeking sponsorships and donations is a skill that is foreign to many engineering students, but it must either be developed or outside assistance brought in. Most universities have a development office, but other students can also be of assistance. The University of Michigan team, for example, is noted for recruiting business majors to handle the financial side of the project, which has helped it to become one of the best-funded and most successful programs in the country. Whether it is handled by the same people designing the car or a separate business department, corporations, individuals, and foundations need to be approached and sold on the idea of supporting a solar car. Family ties and alumni loyalty can be called upon, but sometimes sponsors are simply intrigued by the idea of solar racing. Having their name or logo on a solar car is a rather unique opportunity.
While direct cash donations are often most useful, they are the hardest type of support to obtain. Requests for 'in-kind' donations of materials can be more successful, as many manufacturers are intrigued by the novel uses solar car teams find their products. Requests for donations must be carefully coordinated with the engineering aspects to avoid receiving unsuitable products. Another common problem is the car's design team being overly specific in their requests. With the numerous disciplines that solar racing draws upon, it is difficult to be familiar with standard practices in all of them, and the design team is often unsure of what basic assumptions their supplier might be making.[38] Of course, not just parts for the solar car need to be solicited, but also supplies for every aspect of the project: tools, office supplies, maintenance for support vehicles, and food and housing for the race itself. Sponsors' employees also often offer valuable advice and experience in their respective fields. More substantial partnerships, such as that between McGill University's team iSun and Bell Helicopter or between Yale's Team Lux and Sikorsky Aircraft, are invaluable to the design and construction of technically advanced cars.
Given resources that the team has been able to, or hopes to, secure, a budget for the car must be established. Some car components, such as a specialized solar car motor, are unlikely to be donated, so cash must be available for those purchases. For other items, the team must evaluate its budget and its priorities. One type of battery might be available for donation, but not be as desirable as another variety that needs to be purchased. These real world choices greatly affect the design of the car, from what class the car will enter in a race to whether suspension components need to be made by the team or outsourced to a commercial machine shop. Of course, limited budgets also provide an opportunity for engineering innovation, where a solution must be devised that does not rely on throwing more money at the problem. Additionally, while the top solar cars have generally been among the most expensive, not much further down in the rankings there is not a strong correlation between price and performance (King 1993; King and others 1996).
Beyond the cost of producing the car, money has to be put aside for the race itself. Insufficient funds often prevent teams from competing internationally, but even without high shipping costs, race time expenses have kept teams from entering races (King and others 1996). Simply purchasing sufficient fuel for the (non-solar) support vehicles can cost several thousand dollars.[39] Everyone on the team, usually 10 to 15 people, need foods and lodging for the duration of the race plus the travel time to and from the race. Additionally during the race, the observer must also be provided for. Given the difficulty in finding sponsorships for such items, some teams ask each member to help contribute to these expenses.
For Sunrayce, with all the teams stopping each night in the same location, the race organizers typically arranged for sleeping accommodations in a gymnasium or community center, and an economical meal plan was available. In the WSC, and for the ASC, teams are entirely on their own and must make camp wherever they stop for the night. For the first World Solar Challenge, GM hired a touring company and each night a virtual tent city was constructed, complete with catered meals, hot showers, and a satellite uplink for the journalists with the caravan (Tuckey 1989). Most teams are not this extravagant, but some rent RVs or budget for the use of motels when available. Many teams, however, simply sleep in their support vehicles or in tents.
[11] A capacity of 5 kWh nominally represents a 5 kW draw for 1 hour or a 1 kW draw for 5 hours; the actual value must be calculated through integration and is dependent upon the rate of discharge and the design of the battery. Capacities are often quoted at a 'C' rate, where a C/20 rate would mean the capacity of the battery when it was fully discharged over a 20-hour period.
[12] Gasoline has an energy density of 31.1 MJ/l which is 8.6 kWh/l (Thomas 2000).
[13] Carroll (2003) provides a model for developing such an equation.
[14] Wheel motors were used in the 1993 WSC by Honda, Ingenieurschule Biel, and Northern Territory University (NTU). The success of these new designs led to them being used extensively in the 1996 WSC, with five more teams developing their own motors and five using motors by New Generation Motors (NGM) (Roche and others 1997).
[15] NGM (Ashburn, VA) was founded by alumni of the George Washington University solar car team (Schneider 2004) who had licensed NTU's motor design for their Sunrayce 95 entry. This motor/controller combination had an estimated 93% efficiency, while the most common motor at Sunrayce 95, the Unique Mobility DR086s, was only 85% efficient (King and others 1996). With the success of this motor, NGM was founded and they began commercial production. The primary drawback of this first generation motor was that the motor needed to be removed from the car and disassembled in order to adjust the air gap. Similar to switching gears in a conventional automobile, changing the size of the gap between the rotor and stator alters the motor's torque constant, allowing the balance between maximum speed and torque to be optimized. The next generation NGM motor, introduced for Sunrayce 99, had an external gear to adjust the air gap on the fly (NGM 2002), and although there have been several upgrades to the NGM motor controller, this motor remains the one used by most American teams.
[16] In the US, these batteries were first allowed in Sunrayce 99, with several teams using packs produced by GM Ovonic.
[17] Starting with NASC 2005, lithium ion batteries are even available to stock class teams (ASC 2003).
[18] At this price, typically terrestrial-grade monocrystaline silicon solar cells are available, although the efficiency of those cells has increased with each race. In 1997, the best cells available were only 14% efficient. This had increased to 16.5% efficient in 2001, while for NASC 2005, many stock class teams used the 20% efficient SunPower A-300, available for only $7.50 per Watt (Cole 2004).
[19] Retail prices on these cells can be as high as $500 per Watt, in part because their demanding specifications lead to low production yields. This benefits solar car teams however, as there are then many cells that are slightly out of tolerance for use on a satellite, but are perfectly suitable for a solar car. The surplus cells are often available for donation or extremely reduced prices, although their limited quantities exclude them from the stock class.
[20] Among the first integrators to work closely with solar car teams were Alain Chuzel of SunCat Solar, Phoenix, AZ, and Hans Gochermann, now of Gochermann Solar Technology, Holm, Germany.
[21] Stuart Watkinson, managing director of one of the leading PPT manufacturers - Australian Energy Research Labs (AERL), provides a good explanation of how these high efficiencies (98-99%) are achieved (2001a, 2003a).
[22] GaAs cells are more susceptible to thermal damage than silicon cells and, as each GaAs cell produces two to three times the voltage of a silicon cell, fewer of them will be connected in series, so it is common to protect them individually.
[23] There have been some asymmetric car designs that allow the array to be sloped one direction in the morning and the opposite direction in the afternoon (Kyle 1991).
[24] ASC regulations require the motor to have a dedicated high voltage power switch, so a low voltage enable line can be considered redundant (ASC 2003).
[25] Given the relatively fragile nature of solar cars, particularly their extended tails, combined with their notoriously poor driver visibility, many feel it is unwise for them to back up under driver control, choosing instead to have the car pushed by other team members when necessary.
[26] Sometimes referred to as the battery protection module or battery safety module.
[27] Connections in the high voltage electrical system may need to carry as much as 50 A at 100 V or more. Correspondingly, they have to be fairly substantial, and therefore heavy, to avoid resistive losses. Crimped or soldered lugs secured with bolts are the most common connector, although the easily disconnectable Anderson Powerpole® is also popular.
[28] These sensors take advantage of the phenomena first measured by Edwin Hall (1855-1938), where a magnetic field perpendicular to a flat metal conductor forces the electrons to move to one side of the strip, creating a measurable voltage drop between the edges of the strip (Serway 1997). By measuring this voltage and knowing the charge density of the strip, the magnitude and direction of the magnetic field can be determined. Since a current flowing in a wire produces a magnetic field around the wire proportional to the current, by passing the wire through a round strip configured as a hall-effect sensor, the current can be measured without a direct electrical connection to the wire.
[29] This was the case for Sunrayce 99, though after much debate, the rule was dropped for ASC 2001. (ASC 2000; Sunrayce 1999)
[30] Aurora raced tricycle-designed cars to win WSC 99 and place second in the following three World Solar Challenges (WSC 1999; 2001; 2003; 2005).
[31] These tires were supplied by Bridgestone free of charge to all teams participating in Sunrayce 97 through ASC 2001 (Hayslett 2002; Sunrayce 1996b).
[32] The Michelin tire was used by 66% of the entries in WSC 99, including 8 of the top 10 teams (Selwood 2002).
[33] Solar car tires are typically inflated to 100 PSI (690 kPa) or higher, compared to the approximately 30 PSI (210 kPa) of a conventional automobile. Experiments conducted by the GM Sunraycer team showed that increasing tire inflation pressure can significantly reduce rolling resistance (Carroll 2003).
[34] CNC stands for Computer Numeric Controlled and refers to any machine tool where the actions of a human operator have been replaced by motors guided by a computer program. In a process known as CAM (Computer Aided Machining), a drawing of an object generated with a CAD (Computer Aided Design) program is turned into a tool path file that guides the CNC machine. A machining center is an advanced CNC milling machine that may have up to five axes of movement and automatic bit changing capability.
[35] Sunrayce required only bicycle helmets (Snell B90/ASTM F1447), while for the ASC motorcycle helmets (Snell M95/DOT) were required (ASC 2000; Sunrayce 1999).
[36] Although there have been numerous accidents involving solar cars, for the most part safety precautions have been sufficient and drivers have suffered at most minor injuries. This safety record was called into question in the summer of 2004 when a University of Toronto student was killed in a head on collision while driving their solar car in a Canadian showcase event (Stirling 2004).
[37] The winners of the Sunrayce 95 Cost Effectiveness Award, Northern Essex Community College, built their car for only $20,000 (King and others 1996) while Honda reported their budget for WSC 96 as $1.8 million (Roche and others 1997).
[38] Two examples of difficulties with donations that Team Lux experienced were in wire donated for Lux Perpetua and the aluminum tubing for Lux Aeterna. The donated wire was rated for 1000V instead of the more common 300V, so had much thicker, and heavier, insulation than required. In response to Team Lux's specifications, the manufacturer produced a custom extruded ceramic matrix aluminum tubing for the team, where one of their stock sizes would have likely served with only a few modifications to the car's design.
[39] The fuel consumption of the solar car's support vehicles is one of the great ironies of solar racing. In an effort to correct this, 'Green Caravan Awards' were introduced for NASC 2005 to promote fuel efficiency and the use of alternative fuels (RRVCCC 2005).
[40] Generally, Yale's four solar cars are considered to be Lux Aeterna, Lux Perpetua, Lux Millennia, and the John Lee. Team Lux's first attempt at solar racing, to build a car named Handsome Dan, never came together into a complete car and the John Lee 1.5 and John Lee 2 designations represent a series of electrical and mechanical upgrades to the original John Lee. Complete specifications for each of these cars can be found in Appendix A.
