|The Real World: Human Interface in Optimizing Racing Suspension Dynamics
Note: This paper was originally prepared for the SAE Motorsports Engineering Conference, November 1999 Our purpose in this paper is to explore a question faced daily in "the real world" by mechanical engineers who work with race teams. How does the suspension engineer engage a crew chief and driver in a technical discussion regarding the setup of the racecar? As the technology of racing continues to evolve more quickly than ever before, it is ever more critical that the racing engineer, the driver and the crew chief have a more precise understanding of their dynamic systems and have the facility to explain what is happening, real time, on the track.
Two years ago, a 2-pound change in air pressure was considered reasonable. Today, to quote a Winston Cup driver, "a quarter-pound change in the left rear transformed the car." How does the engineer optimize performance when changes to suit the driver's "feel" are moving away from the optimal set up? Worse: how does he explain what's needed when the 'human interface' doesn't understand what he's saying? There is a set of numbers which in physics and engineering terms describes the limits of a racecar's performance. The basic theory of suspension is to minimize the effects of the tradeoffs resulting from disproportionate loading. A common tool to graph these tradeoffs is the friction circle. The object of racers is simple: to try to keep the driver's input "line" as close to the outside of that circle as possible. Theoretically the performance should always equal the outside limit; but driver reaction time, lap geometry, mistakes, changes in conditions, etc. all move the actual performance away from the ideal. The ideal "grip circle" for set-up takes into consideration conditions where traction is not the limiter, engine size and output are. To work with the concept of the "grip circle" both the engineer and the crewchief must be able to communicate how the suspension system works under varying loads and conditions, and how it interacts with the other vehicle components. They say an aircraft is "a series of compromises, running in close formation." The same can be said of a car. In many cases dynamic instability is the way to reach the ideal numbers. Virtually everything you do that departs from the ideal -- to give the crew chief and driver a car that "feels right"-- slows it down. The more the crew chief and driver understand the system's dynamics, the more accurate are their descriptions of what's happening on the track, and the better able the driver is to adjust his driving so that what "feels good" is closer to the performance limits. The suspension engineer whose crew and driver understand the physics and work in sympathy with what the car says is better able to approach those ideal numbers. Just as in combat aircraft, what's comfortable or even physically possible for the human driver in a racecar can be diametrically opposed to what's needed for performance. There is a constant struggle between optimizing those ideal numbers and optimizing the human element's contribution. As Alan Jenkins, Technical Director of Stewart Grand Prix observed, "What Jan [Magnussen's] engineer felt he was able to try with Jan and what he thought Jan might appreciate changed quite a bit over the year." There are compromises to be made in what's ideal for one set of conditions and what's demanded by others within the same performance arena. Peter Wright described the issue quite elegantly in an article in Racecar Engineering, while exploring a comparison between the on-board automatic control systems being tested by a road course driver in California and a rally driver in Wales. He commented that top-level circuit drivers operate at the limit, virtually open-loop, or without modification: "If [a circuit racer] driver knows the circuit well enough and has good confidence in that knowledge he will input brake, steering and throttle with almost no corrections to take the car to, but not over, the limit of grip
The [rally] driver receives signals giving him the essential information he needs about position, speed and available grip that enable him to adjust, closed-loop, the brakes, throttle and steering to gain the response from the car that he wants. How much feedback he needs to know how hard he has to work, adjusting the response to the car, will depend on how accurate the initial, open-loop brake, steer and throttle inputs were." The control will also depend on how specifically he and his crew were able to quantify the set-up. It is then a logical extrapolation to argue that the more "open-loop" driving (i.e. repetitive, as on a Winston Cup oval) is required to optimize performance, the more important it is that the driver understands the feedback he's getting from the car and is able to communicate that to his crew chief. As Joe Amato describes it, top fuel dragsters have evolved to such a degree that basically the only job the driver has is "to mash down on the right pedal and steer." The car does the rest. When top speeds were 0 to 260 in a quarter-mile, the best drivers had the reflexes to make the gear changes and keep the wheels straight; but when speeds are 0-to-275 mph in 660 feet, only a Lenco-style transmission and lock-up clutch can make the changes accurately and quickly enough. The engineering compromises are pretty much all car. For another example, take caster set-up. Uncompensated zero-caster reacts to every input from the driver and the road, making it impossible for the driver to keep up with the adjustments. As happened in the F-22 test when the computer shut down during the landing sequence, 100 feet off the runway, and the aircraft was bobbing up and down in perfect asynchronous wave-forms, sometimes the driver simply can't keep ahead of the corrections. As Peter Wright concludes: "It is a fundamental of both aircraft and racecar dynamics that an unstable vehicle will be the quickest to respond to inputs" When the reaction time of a system is in thousandths of a second, the poor old human - with several times slower that response rate on his best day - is going to lose. The nexus of the problem is that when the driver/crew chief and the race engineer can't discuss the nuances that lead to this sort of field result, the damage can be greater than not optimizing performance. The two basic approaches to racecar engineering are those who base their set-ups on experience, and those who do the math and apply the numbers. GIGO applies in each case: there are those who, without any theoretical grounding, don't question their assumptions (either the assumptions which have been "learned" in OJT training or those they draw, uncritically, from what they "see.") There are equal numbers of button sorters who don't question the machine either. Engineering relies on the scientific method, which is not intuitive. The scientific method is difficult to teach yourself, as has been demonstrated by many of the top teams, but is possible to learn outside of the strict academic environment. The major stumbling block is that both ends of the spectrum tend to put an emotional investment in their positions; and, with those vested interests, are not willing to make the effort to simply see what works. We are not putting forward a case to discount the abilities of stock car drivers and their crew chiefs, or to suggest that they don't know what they're talking about in the real world/ "hands-on" terms they have devoted decades to acquiring. To the contrary, the focus of this paper is to examine the issues, and hopefully, to provide a model which helps both engineers and non-engineers - both of whom "know what they're talking about" -- speak the same language. The research Leading Edge Motorsport has conducted among Winston Cup drivers and crewchiefs is intended to set a framework to which virtually every engineer who works with a NASCAR team can relate. If, at the top end of the sport, the driver and crew chief have "x" level of engineering background; the teams further down the chain have proportionately less - as readers will attest. Leading Edge Motorsport is addressing the communications gap with a specific approach. To establish the rationale for that approach it is necessary to define the working environment. An illustration of the environment to which we're referring: the use of the term "push" instead of "understeer." The driver says the car is "pushing." The engineer has to ask, "How does it feel? Is there not enough spring on the right front corner? Too much toe-in in the rear? Not enough caster? Too much? Not enough air pressure? Too much? What about roll rate? Is the problem in the front of the car or the back? Or both? Is it the way the car feels when you turn into the corner? Or is it all of the above?" Without understanding what and how each of the elements affects the dynamic system, the driver - and the crew chef - can't begin to provide the information the engineer needs to help them localize and describe how to solve the problem. There is frequent attention given to the problem of providing real world experience to lab engineers. John Valentine, Chief Engineer for Ford Motorsports notes that "Many engineers come out of school with very good computer and analytical skills, but without much hands-on experience. We find working with motorsports teams really lets them get their hands dirty, so to speak. They work with the actual hardware and try to apply analytical skills to real-world problems, while seeing the results very quickly." Lou Patane reported similar results from the Chrysler Technology Center where "racing is an integral part of the product development process and a training ground which uses the production platform envelope of motorsports to push the technological environment." "Working with the racing team generally helps our Ford engineering teams do work they couldn't otherwise experience," Valentine continues. "While they can't drive the car, they have to listen to what the driver is saying. With our engineers working along and listening to what the crewchief says, they can offer suggestions that utilize different types of tests." When one side of the conversation is handled by someone who thinks in qualitative stages, and the other side is covered by one who thinks holistically, and in generalities, the communication is nil; yet like the proverbial married couple, each party thinks they know what the other is talking about. OJT provides a clear demonstration of what happens in any given system - but not the why. But one of the things that are frequently overlooked in advanced education is the concept of cause and effect: they understand the theory, but they don't have to produce a result. The numbers say the system "works" but the real world engineer experiences the daily frustration of working everything "by the numbers" only to find that it still doesn't work. OJT produces the experience typical of the way Ronnie Hopkins Jr. describes chassis and suspension-building legend, Ronnie Hopkins Sr.: "I know that I've got the best education ever in racing. It was like one brilliant college professor dedicated to only one student for more than 20 years. And it was a hands-on thing-not just teaching and showing, but doing it." Ron Hopkins Sr., was, in fact, the man credited with teaching Ray Evernham - whom he described as "a quick study" - about set-up and suspension. The Hopkins-Evernham team is proof of exactly the thesis being put forward: that it is possible to produce top engineering results without a formal education if the language and perceptions are precise and shared. The human interface between the set-up of the car and the dynamics of the system on the track is critical in achieving optimal performance from the system. The goal is to produce a dialogue which allows the non-engineer to tell the engineer "We did this, this and this; and it didn't work" in such a way that the engineer can look at the actual numbers describing what the team did in order to find out why "it doesn't work." It is an essential part of the racing engineer's job to recognize the limits of knowledge and experience which are at work in any given situation or team of people - not theirs to understand what the engineer, whom they're paying, is talking about. To achieve the performance goal, one needs to look at the components as a system, as a process, rather than from the typical vantage points -- such as the slices of time in the endlessly repetitive loops of a practice session. One also needs to look at the psychology of comprehension. Taking the process management approach, one can focus on the issue at the core of the communications gap, analyze the complexities contributing to the lack of clarity, determine the most effective countermeasures and than execute those solutions. Ø Focus Ø Analyze Ø Develop Ø Execute Having focused on the link where performance optimization often breaks down - those situations where the human interface is a variable, one must then examine the inputs, both engineering and psychographic which are influencing factors. Many committed individuals muddy the concept of "Focus" by concentrating so narrowly they lose all situational awareness. Over-concentration gives one nothing to analyze. Nor is it the goal of the enterprise to answer individual engineering challenges monolithically - the job is to make the car win. There are several obstacles to the goal of bringing the ideal engineering set-up and "what's comfortable" for the real-world driver together. The two major categories of factors influencing the input-side of the set-up process are 1) mechanics and 2) the ability to communicate. The mechanics are the set of numbers already mentioned which equate to the perfect set-up. While being finite limits, these parameters are by no means simple to determine. As Alan Jenkins comments, "The trick is determining the whole interaction between the various elements
to build the databases of information that let you run combinations that bring you back to similar wheel rates or ride height conditions." The problem of repeatability is exacerbated when for example one finds that, outside of Winston Cup, race teams simply don't maintain solid race books. By insisting that they focus on the numbers they are producing, one can begin to direct their thinking into the Focus-Analyze-Develop-Execute paradigm. The objective of this paper is to explore how the engineer and the team can discuss these factors in a common, accurate language. When one analyzes the source of the communications, one finds not only the usual invisible barriers between areas of specialty, but also an education gap - and one which is adversely affected by several inhibitors inherent in the psychology of learning. What one could characterize as "the psychology of learning," as it relates to both engineering and human interface dynamics, is affected by three major factors, based on field observations in a variety of different racing arenas, as well as in other situations. A good example of the impact of this consideration was observed in NASA's Apollo program. One would hardly characterize American astronauts as uneducated; however, NASA recognized that the traditional classroom models for teaching geology to those crews (whose assignment was to explore the lunar landscape and collect samples which geologists hoped would provide insight into the origin of the Moon) were not achieving the anticipated objective. The Agency hired Lee Silver, Ph.D. Professor of Geology at Cal Tech to accomplish the task. An inspiration to anyone in this difficult role, Professor Silver taught the Apollo 15, 16 and 17 crews to speak the language of geology, so that those who were responsible for examining and interpreting the hard data could communicate and optimize the results of the lunar explorations. Thus, one must examine the reasons for non-communication before recommending solutions to the problem. The chief categories are: · Educational limits · Preconceptions and "Rules of Thumb" · Advice from Associates/ What others are doing Assessing the Education Factor The first, limited formal education, provides the platform on which the other factors stand and which could otherwise be overcome by simple demonstration. It is unfortunately the case that those without the educational background to argue on an informed basis are even more resistant when talking to the formally educated. Witness the now legendary "two-tire change" controversy. At issue is the difficulty the engineer has in explaining "how things do work" vs. "how they seem to work" in a manner which will directly contribute to improved performance. There is an irreducible minimum comprehension level. It is necessary to understand the concepts as well as recognize practical examples in order to make improvements in suspension dynamics. When it is a fact of life that the average American is not prepared to extrapolate concepts from perceived data, the engineer has to limit his discussions to the level of comprehension. Without belaboring the point, it is useful to quantify the education factor, especially as it fosters the climate where "rule of thumb" and "what the other guy is doing" dominate. According to the 1992 National Adult Literacy Survey, more than fifty percent of the adult American population "were apt to experience considerable difficulty in performing tasks that required them to integrate or synthesize information from complex or lengthy texts or to perform quantitative tasks that involved two or more sequential operations." There is a correlation between number of years of formal education and the prose, document and quantitative literacy levels achieved. Only 3 to 4 percent sampling of the population demonstrated the skills, which enabled the reader to "compare and contrast complex information or generate new information by combining the information with common knowledge [italics added]
when high-level, text-based inferences are needed or when specialized background knowledge is required." Only this small percentage could "to integrate multiple pieces of information from one or more documents or generate new information by entering required information in the proper place [note: such as a set-up sheet] when information displays are complex
or when specialized knowledge is required." And just 4 percent displayed the Level 5 ability to "infer the necessary arithmetic operation or perform multiple arithmetic operations sequentially
when background knowledge is required to determine the quantities or operations needed." When manufacturers must rewrite their procedure manuals to lower levels so their employees can understand them; when USA Today's stylebook requires copy written to a 6th-grade reading level; when a crew chief and his driver don't know that a vertical line on a "force-over-time" shock dynamometer graph shows the opening of a valve; where are we to start in explaining how shock absorbers work? In Formula One, the Stewart Grand Prix team (a relatively small team by Formula One standards) employs more than a dozen engineers, from Technical Director Alan Jenkins to specialists in shocks, composites, stress, aerodynamics, gearbox, hydraulics, electronics and systems & detail design. They interface with Cosworth Engineering and Advanced Vehicle Technology division at Ford. McLaren has dozens of engineers on its staff of 160. No CART team is without several engineers on staff and each driver has his own personal race engineer. The sportscar crowd covers the range between the two extremes, with engineering expertise present in direct relationship to the team budget. . Engineering input in Winston Cup is a function of team budget, ranging from a 22-year-old to the full resources of Ford SVO. Based on a recent survey by Leading Edge Motorsport, the average education completed by a Winston Cup crewchief is high school. The technical training consists, in the majority of cases, of 20 years on the job. For the teams with whom many SAE engineers work, the drivers and crewchiefs have proportionately less engineering, or even scientific approach, training than Winston Cup.
As motorsports automotive engineers know from daily experience, most of the stockcar people they encounter are working primarily with On the Job Training. It is not argued that there aren't many extremely valuable efforts underway to improve the education - particularly the engineering education - of those who are involved in the industry of motorsports. The University of Central Florida through Professor Robert Hueskstra is involved in an engine research project. The new Chrysler development center in Auburn Hills, Michigan was custom-built to provide an academic and research atmosphere for recruiting the best and brightest students interested in performance vehicle dynamics. Delta Community College in Saginaw set up an automotive maintenance training program for General Motors a decade ago. And there are dozens of engineering schools and programs available at the post-secondary level. Like Ford Motorsports, other manufacturers are realizing the crossover benefits of OJT engineering training. Norton Manufacturing, maker of crankshafts is one, working with Terra (OH) Community College. And fortunately, at a time when vocational education funding has been reduced, partnerships like the Street Stock program created by Mesa Marin Raceway and the Kern High School District in Bakersfield CA (implemented in large part for "a segment of our school population that does not get connected in a traditional sense
students who would not normally be involved in school") are being championed. While not providing an engineering background to these future racers, the alliance between formal education and backlot garages is a hopeful sign. But the typical experience is that just ten years ago, Dave Charpentier (now crew chief for Rick Mast), whose US Navy training provided him with post-military employment as a supervisor/engineer at a nuclear power plant, used his background in electronics and instrumentation to design and build monitors for engine dynos. In just three years the business he founded was serving the needs of over 80 shops in various race divisions all over the country. In 1996 - seven years later -- he was hired by SABCO as team engineer. Robin Pemberton says the most difficult job he has to fill is that of shock engineer. Certainly there are more engineers in NASCAR now than there ever were; just as there is a higher proportion of high school graduates who go on to college than there was after the Second World War - when the legendary names of NASCAR began to build their dynasties. More prospective drivers and more prospective crew chiefs are getting at least some of the basics. To the purpose of this examination, however, it is fact that - like astronauts who see rocks instead of bits of the primordial lunar crust -- basic ignorance of engineering continues to detract from the optimization of performance. OJT is the reality in most applications; and the problem with OJT is the lack of education in theory. Ad hoc training may provide the "what" and the "how;" but it rarely considers the "why." As a result, when process is being analyzed and no theory is brought to the analysis, a modification may look successful; but often the change did not affect the outcome, and the observer doesn't notice because he doesn't know what to look for. In this area of dynamics, errors are compounded because the circumstances of the test result are similar, but are, in fact, not the same and the conclusions drawn are flat wrong. When one recalls from the Literacy Survey that nearly half of all American adults are not proficient enough to perform sequential arithmetical operations, it becomes clear why improving the communications process in a racing dynamics situation can be so impactful. Preconceptions/Rules of Thumb The second psychological factor influencing the human interface dynamic is Preconceptions/"Rules of Thumb." It is not surprising that an environment characterized by limited formal training is an environment where "conventional wisdom" holds sway. In motorsports environments where engineering expertise is considered elemental - such as Formula One and CART - teams are less likely to look at "rules of thumb" and more likely to demand proofs in real numbers in real world conditions. The engine and chassis parameters that are measured and recorded on the Stewart Grand Prix vehicles number in the hundreds. The problem in the situation under examination is that outdated thinking -and outdated engineering - becomes codified. As a result, conventional wisdom is given a foothold as all but a handful of crew chiefs try to cobble solutions to ever-more-advanced problems out of limited options. A clear example is the growing predilection of Winston Cup teams to go with qualifying-session setups which progressively jack the car down in the rear. This tendency is so extreme that one can easily find cars back in the garage area, which are still rebounding from the shock compression, long after the driver has exited the car. These jack-down set-ups work according to the rule of thumb that if one can get the rear spoiler out of the air flow by dialing in excessive rebound over bump, drag will be significantly reduced. While in the case of NASCAR rear spoilers, this approach does work, the problem - as a number of drivers at Talladega were quick to point out - is that the car becomes virtually uncontrollable. The trade-offs become quickly disadvantageous: as the compression increases in the rear, the ability of the driver to compensate - in effect, to drive the car in the corners - decreases. As was stated at the start, a certain set of numbers defines the "stick" of the car - the capability the machine has in terms of velocity. In all attitudes there are compromises, such as asking the car to turn while braking or accelerate while turning, or take a disproportionate load (i.e. anything that causes any one corner to do more than a 25% share and reduces the capability of the system to perform to 100%.) When you have all those "compromises flying in close formation," you must make choices: setting the car up for cornering and braking is not what you do to get maximum acceleration. You, as engineer, have to agree along with the crewchief and driver how to spend your "grip budget." To paraphrase Ben Franklin: one toenail beyond the limit equals disaster; one toenail inside the limit is bliss. The "rule of thumb" example of jacking the car down for qualifying may be bad engineering; but it does suggest the objective considered in this paper can be realized: drivers and their crew chiefs are willing to drive set-ups which are supposed to improve performance, even if that makes the set-up less comfortable for the driver. Michael Schumacher is known to prefer a hyper-reactive car because he is willing to drive to the limits to achieve performance. He is constantly adapting his driving style to the set of numbers which define what is possible within the dynamic parameters of the system. As a result, he is closer to the limits of the grip circle than competitors farther back in the pack. Advice from Friends & Associates Process input factor number three is "advice from friends and associates" or "what the other guy is doing." Again engineering solutions militate against "follow the pack thinking" while OJT environments allow random casting for answers to fill the void created by the lack of hard information and theoretical knowledge. Roger Penske never cared a whit about what other teams were doing; he has always insisted on knowing what his own numbers were telling him. His dominating championship year with Mercedes was the result of that insistence on measurable, repeatable performance. Frank Williams and Ron Dennis have been successful because they did what the numbers said. Determining solutions to the problem As Paul Van Valkenburgh has observed of the comprehension gap, especially in the context of short track technology, "In this game, the open questions vastly outnumber the known answers. None of [the racing books that focus on short track technology] really give any scientific set-up secrets. They mention all the key set-up considerations and then just seem to conclude, 'Cut and try.'" The solution which has proven effective in the author's field tests is to break the dynamic set-up problem into its components, applying the same focus-analyze-develop-execute process management disciplines applied in this paper. Good "rule of thumb" has trained even the Saturday night racer to record tire pressures and temperatures on a set-up sheet. When more data is collected - within a context that the both the basic and the sophisticated team can understand - the affective factors emerge and can be defined and explained by the engineer. The author uses the appended set-up sheet when working with race teams. It is more detailed than most commercially-available forms and was specifically designed to map the correlation among tire pressure and temperature; shock, wheel and spring rate, and driver effect. As each input is recorded in a format which shows the interrelationship of even the alien concepts, the big picture - quite literally - begins to emerge. It is far easier to examine the suspension set-up when the crew chief understands that it is wheel rate that is important, not spring rate, to offer the most basic - and most frequently overlooked example. The recording of data is equally effective in opening communication with the driver. Too much reliance is put on driver input; in fact hard information frequently goes unheeded if it conflicts with the driver input. Once the evidence is presented in a format for review and analysis, one can begin to work with the team through set-up and driving style alternatives - using the scientific method - to advance toward optimized performance. When these sub-process loops are taken into consideration they not only allow, but often force, a real understanding of what's going on in the real world. It is then our job as racing engineers to explain, to define, to teach. If the goal is a winning team, we can't be satisfied that the crew chief and driver might be wil
Back to LEMD University