Ann Johnson is Assistant Professor of History at the University of South Carolina.

Hitting the BRAKES

Engineering Design and the Production of KnowledgeBy Ann Johnson

Duke University Press

Copyright © 2009 Duke University Press
All right reserved.
ISBN: 978-0-8223-4526-8

Contents

Preface…………………………………………………………………………………………………xvAcknowledgments………………………………………………………………………………………….1one Design and the Knowledge Community…………………………………………………………………….23two A Genealogy of Knowledge Communities and Their Artifacts…………………………………………………37three The British Road Research Laboratory: Constructing the Questions………………………………………..63four The Track and the Lab: Brake Testing from Dynamometers to Simulations…………………………………….85five From Things Back to Ideas: Constructing Theories of Vehicle Dynamics……………………………………..103six Learning from Failure: Antilock Systems Emerge in the United States……………………………………….117seven Eines ist sicher! Successful Antilock Systems in West Germany…………………………………………..137eight Public Proprietary Knowledge? Knowledge Communities between the Private and Public Sectors…………………157epilogue ABS and Risk Compensation………………………………………………………………………..167Notes…………………………………………………………………………………………………..187Bibliography…………………………………………………………………………………………….201

Chapter One

Design and the Knowledge Community

In 13 August 2000 an anonymous author with the screen name “Starrion” posted a review of the 2000 Mercury Sable on the website Epinions.com. Starrion told the following story:

I was on Rt 24 headed for Walnut Creek, California. And I’m approaching my exit. It’s a beautiful day in the San Francisco area, I’ve got the sunroof open, going 75, the local radio station playing, and all is well. I know I need to drop a lane, and when an opening appears behind a smart-looking silver Grand Am, I take it. I have only a few feet to spare. Just as I enter the lane, a cloud comes up from the road, the Grand Am’s brake lights come on, and she panic-stops. Just as I put my foot to the floor, I see a tire twenty feet in the air-complete with aluminum hub-come bounding over the Pontiac headed right for me. I brake ever harder and the Sable stops flat. No control loss, pointed straight and I out-stopped the Grand Am. The tire lands about a foot from my passenger side door and rebounds into panicking traffic behind me. It was so close that chips of rubber landed in my car through the sunroof. As I started off, I noticed a Jeep Cherokee sans rear-mounted spare about a half a mile up. As I made my exit, I was very glad I had clean underwear in my suitcase.

How was this ordinary family sedan able to avoid this hazard and come to an uneventful stop? Starrion’s answer came in the next line of the post: “The Sable LS has great brakes. Better than any other rental car I’ve driven. This car had ABS or else I wouldn’t be writing this. I highly recommend it. Extra highly.” While there are many reasons for the differences in the ways various cars handle, in this case, the car’s behavior is typical of cars with antilock braking systems, or ABS. When Starrion slammed on the brakes, the Mercury Sable went into an antilock mode. Its computer compared the angular speeds (i.e., the speed of rotation) of all four wheels. The computer also compared the angular velocity of the wheel to the linear velocity of the car. If the computer determined that one or more wheels were decelerating more quickly than the vehicle, the computer could send an electrical signal to the hydraulic system controlling the brakes. That signal would tell the brake caliper (at the wheel that was decelerating too quickly) to pulse, reducing the braking pressure on that wheel. This computerized system is able to apply and release the brakes dozens of times per second, much faster than adrenaline-infused Starrion could pump the brakes. Once the rate of deceleration of that wheel was back in line with the vehicle speed, the brake would revert back to normal operation, even though the wheel sensors would continue to monitor the rate of deceleration for further problems. The result of this active monitoring and modulation of the automobile’s brakes meant that the car never started to skid, and as a result Starrion avoided the crash.

Skidding is dangerous for two reasons: a driver cannot steer a skidding car effectively, and a skidding car can take longer to stop. Although it may seem counterintuitive that a car with locked wheels stops less quickly, under most conditions sliding wheels encounter less resistance than rolling wheels. This is also the reason a skidding car often spins; the wheels with the lower sliding resistance move faster than the wheels with the higher rolling resistance. As a result, if the rear wheels start to skid, as they commonly do on a pickup, which usually carries less weight on the rear axle, the vehicle will spin 180 degrees to face the oncoming traffic. No amount of steering correction or countersteering can overcome the friction differential because the direction in which the wheels are pointing does not matter when wheels are sliding, only when they are rolling. Because of the car’s ABS, Starrion avoided skidding and therefore maintained directional control, making possible the emergency lane change and quick stop. In this scenario, the proper use of antilock brakes allowed the driver to avoid losing control, to change lanes under emergency braking conditions, and to stop as quickly as possible. The antilock braking system is a well-designed invention for several reasons. Once invented, it appears obvious. What it does and how it does it are easy to understand, at least superficially. Its function is socially desirable; in principle all drivers prefer not to lose control of their vehicles and to be able to stop quickly. An antilock braking system is fundamentally a technology of control, often represented as a triumph of engineering over dangerous road conditions and unpredictable drivers. In early advertisements, ABS was often presented as a great idea whose time had finally come. Skidding had been a problem on cars for decades. The fact that antilock systems only begin to proliferate in the 1980s points to a difference between ABS as an idea and ABS as a functional, real-world product. From an engineering perspective, ABS is a complex, electronically controlled, mechanical system. At its heart, it is a measuring apparatus, but not one which exists in a clean, temperature-controlled laboratory. Instead, ABS has to send electrical signals along the bottom of automobiles, which vibrate, get wet, dirty, and salty, operate in temperature extremes, and operate at speeds from 0 to over 100 miles per hour. An antilock system has to interface with but not interfere with several other electrical systems. If ABS fails, the braking system itself must continue to work normally. Making ABS meet these real-world conditions proved to be far more difficult than simply dreaming up the idea. A functional ABS required newly invented components, from sensors to computers, circuits to valves, and new understandings of how drivers behave and how other parts of the automobile function. An antilock braking system may be easy to imagine, but it was difficult to make.

The difficulty of the design problem is best shown by the timeline of ABS’S development. Patents for antilock systems appeared as early as the 1930s. These early patents were never realized commercially, but they do indicate that some engineers were thinking about inventing devices to prevent skidding nearly a half century before other engineers succeeded in doing so. This lag between the realization of the problem and the appearance of the first technological fix in the mid-1960s is indicative of the complexity of designing and producing antilock systems. Broad awareness of ABS by consumers and market saturation in the United States and Europe did not occur until the late 1980s and early 1990s. In short, developing ABS took most of the twentieth century and involved dozens of different companies and government agencies and hundreds of engineers in Europe, the United States, Canada, and Japan.

THE DESIGN QUESTION: PUTTING IDEAS AND THINGS INTO PLAY

This is not principally a book about automobiles, but about engineering design. I focus on the design of ABS as a case study to illuminate the complicated process of actually designing something. If one draws from recent approaches in science and technology studies and focuses on the practices of engineers-that is, what they do-then the central problem of design is not where ideas come from but how ideas become things. It is worth noting at the outset that ideas that do not work are without value in the marketplace. In the world of engineering, artifacts clearly trump ideas and propositions. This is not to underestimate the role of knowledge in engineering. Quite the contrary, to transform an idea into a functional thing requires knowledge. Furthermore, the transformation from idea into thing or one thing into another usually requires new knowledge, not simply repackaged or applied scientific knowledge. In fact, making substantially new things requires substantially new knowledge; ideas, therefore, do matter.

Knowledge is produced in and by communities. In this book I use a framework I call knowledge communities. Through the case study of antilock braking systems I show that knowledge communities are the basic locus of knowledge production in design engineering and much science (and certainly they appear and function in many other human endeavors). A knowledge community is a socio-epistemological structure, and I do not privilege either the social or the epistemological dimension. As a result, I do not want to describe a body of knowledge independently of a community, implying that free-floating ideas and artifacts came together and attracted a community of practitioners. In this scenario a basic causal problem arises: What agents determined the central body of knowledge in the first place? Conversely the community cannot be prioritized either, since it has to be clear what attractor (e.g., a problem or a set of ideas) brought together a multidisciplinary group of practitioners in the first place. Consequently, an analysis of knowledge communities requires a real commitment to symmetry between knowledge and social organization; it may even require the collapse of both into a single entity.

In terms of methodology, knowledge communities lend themselves to historical analysis, or at least historical sociology, because describing the development of a particular knowledge community keeps social and epistemological factors properly integrated. One can track the changing definition of the community’s central problem, knowledge, and artifacts while simultaneously watching participants come into and out of the community. In other words, the development of the knowledge community must be described in order to determine whether a knowledge community exists at all. In addition, a developmental narrative shows the dynamics of the knowledge community. Dealing separately with knowledge and then social arrangements creates the impression that both are static, as the epistemological narrative must be held up as the social narrative catches up, or vice versa. The shifts in what knowledge was valued by the knowledge community and the concurrent changes in community membership are very telling in the case of ABS. In fact, these shifts show the contingent process of the social construction of knowledge; ABS had many different possible paths of development, but certain paths were more highly valued than others. There were also important physical and economic constraints on the system’s development. To explain the convergence of the ABS design around an electronically controlled hydraulic system produced by Robert Bosch, GmbH, a West German company, in the late 1970s, I must also explain what kinds of knowledge and practitioners were valued. The knowledge community dictated these preferences and values.

There are several other salient features of the knowledge community. Knowledge communities in engineering initially form around a communally defined problem, which I call the attractor. But the problem that sits in the center of the community and that attracts the often diverse practitioners is never static; the community is constantly redefining it, ruling certain solutions and definitions in and out. This is an informal process; in fact, formalizing the process often marks the transformation of the knowledge community into something else, perhaps a professional community. Because the focus of a knowledge community is so narrow (far narrower than the focus of a discipline or a professional society), the group is quite small, varying significantly over time but always remaining intimate. Intimacy in a knowledge community means that the practitioners all know each other, or each other’s work. In the development of ABS no more than two degrees of separation existed between members of the knowledge community. As a result of that requirement, knowledge communities rarely exceed more than a few hundred people. Those that grow larger begin to splinter as the practitioners develop discrete problems on which to focus. Furthermore, knowledge communities are nonexclusive, so engineers may work within several different knowledge communities. Because a knowledge community centers on a problem, one can be engaged in several different problems simultaneously. New ideas and tools move into the community in part because participants move between communities, and ideas require human vectors.

Thus I can trace knowledge and community evolution through a series of how questions that united the effort to design antilock systems. The nascent community initially asked, How can we reduce skidding accidents in passenger cars? This question splintered into questions about driver education and psychology, tire design, road surface design, and finally brake design. The group that defined skidding as a braking problem continually refined their central question. They first asked, How can we design a braking system to prevent skidding in panic-stop situations without requiring the driver to change his or her habits? Then they asked, How can we determine that a car is about to skid so that a system can prevent skidding? How should we design an electronic system to measure imminent skidding, then a hydraulic system to react to that electrical signal? And finally they asked, How can we make a mass-producible electronic control system that will integrate with the braking system and prevent the car from skidding without the driver even knowing it is there? Whether the systems completely accommodate these requirements is still unclear; the question of assessment is the subject of this book’s epilogue. But I want to emphasize the changing questions the community asked and show how these changes were interwoven with the changing membership of the knowledge community.

This project of connecting engineering design to the production of engineering knowledge is hardly new or novel; any appearance of novelty comes from the fact that, compared to scientific knowledge (or science more generally), engineering knowledge (or engineering more generally) is significantly understudied. In the small but growing field of engineering studies, one significant focus has been on design and on the nature of engineering knowledge. In What Engineers Know and How They Know It, Walter Vincenti equates design and knowledge by explaining that design and problem solving are synonymous at the working level. He writes, “Day-to-day design practice not only uses engineering knowledge, it also contributes to it.” Designing is problem solving, and solutions and designs both constitute new knowledge. But the voluminous literature on scientific knowledge can actually interfere with an understanding of engineering knowledge. Philosophy of science, along with epistemology, the philosophical study of knowledge, makes certain claims about the nature of knowledge, and these claims tend to be incommensurable with notions of engineering knowledge that focus on the process of design. Epistemologists have long focused on propositional knowledge and argued that knowledge is “justified true belief.” As a result of these two deeply held principles, philosophers of science have tended to focus on theories as exemplars of knowledge. These analyses of knowledge as universal and propositional do not fit well into technological, engineering, or design knowledge; one needs a broader, less formal, more inclusive concept of knowledge. Engineering knowledge rarely takes the form of formal scientific theories; more often engineers’ knowledge resides in the artifacts they design or in the processes of designing those artifacts.

The philosopher Davis Baird has tackled this problem of broadening a definition of knowledge in his book Thing Knowledge: A Philosophy of Scientific Instruments, published in 2004. Baird shows that in addition to propositional, theoretical knowledge, science and technology also both need and produce knowledge in physical, artifactual forms. He breaks this concept of the physical instantiation of knowledge into three kinds:

1. Working knowledge: knowledge communicated by the operation of a device or instrument; knowledge by action.

2. Model knowledge: knowledge by representation, using physical models.

3. Encapsulating knowledge: knowledge which both represents and acts; incorporates both of the previous categories and in most cases provides measurements. The best examples are scientific instruments.

(Continues…)

Excerpted from Hitting the BRAKESby Ann Johnson Copyright © 2009 by Duke University Press. Excerpted by permission.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.