Chapter 18

Doing History by Reverse Engineering Electronic Devices

Yana Boeva, Devon Elliott, Edward Jones-Imhotep, Shezan Muhammedi, and William J. Turkel

In this chapter we describe three experiences of collaboratively reverse engineering historical electronic artifacts at the University of Western Ontario’s Lab for Humanistic Fabrication, a setting that supports hands-on fabrication and experimentation, programming and computer-aided design, and traditional historical research. Our first case study comes from Elliott’s work on the re-creation of wireless effects designed by early-twentieth-century magicians to simulate mind reading. These effects depended on electromagnetic induction, a technique that has recently come to prominence for its applications in radio-frequency ID tagging, wireless charging, and near-field and secure wireless communications, largely driven by interest in the Internet of Things. The second case explores Muhammedi’s software re-creation of an analog electronic musical instrument built by the Canadian engineering pioneer Hugh Le Caine. We used a visual programming language to create a low-cost, widely accessible, and suggestive model of Le Caine’s device, one that simultaneously helps us to interrogate the digital-analog divide. A final case study focuses on the analog electronic computer — part of a long-dismissed tradition of analog computing that was overshadowed by digital developments. Here, Jones-Imhotep and Turkel make small analog computers to further explore the relation between the digital and the analog and to gain insights about the crucial role of analogs in the history of scientific instrumentation. In all of these cases, we are interested in exploring how collaborative acts of making and experimentation can deepen scholarly and public understanding of history.

We refer to our approach as “humanistic fabrication” and believe that it provides a performative and ontological complement to the epistemological or representational work that comprises most day-to-day practice in historical disciplines. Rather than attempting to reproduce (knowledge of) any particular technological past, our goal is to stage situations that allow us to experience a specific process, moment, or flow.[1] Humanistic fabrication is thus a kind of “ontological theatre,” to use the terminology of Andrew Pickering (2010).[2] Our concern is to engage with the world rather than to create representations of it. Sometimes we do learn things that we can put into words, but this “entail[s] a detour, away from performance and through the space of representation, which has the effect of veiling the world of performance from us” (Pickering 20, emphasis in original). So rather than writing about humanistic fabrication, we typically share this research practice with colleagues, students, and the general public by offering dozens of collaborative, hands-on workshops each year: in classrooms and hackerspaces; at libraries, museums, and conferences; and during intensive venues such as the Digital Humanities Summer Institute (DHSI). We also document our performances in online forums, on our own websites, and in GitHub, so that others may be encouraged to stage comparable performances of their own. Take this chapter as an offering in a similar spirit (Turkel and Elliott 2014).[3]

In addition to terms like “re-making” or “re-creating,” we use the expression “reverse engineering” to focus on the exploratory, experiential, and performative aspect of our work. What is important to us is immersing ourselves in a setting where we are collaboratively and actively engaged with design, intention, material constraint, and the “adjacent possible”[4] — with process rather than product. Within the history of technology, scholars have long privileged large-scale industrial technologies as sites to understand social transformation (Hughes 1983). But electrical or transportation infrastructures lend themselves poorly to experimental, hands-on exploration. By shifting the scale to smaller, often more intimate objects such as household technologies and personal computing devices, re-making or reverse engineering can provide unusual insights about the micro-level contours that shape our broader historical understandings. And in the absence of historical evidence about the design, use, and impact of material devices, those practices can help us to reenact and reimagine portions of the historical worlds they occupied and the meanings they held (Turkel 2011).

Case 1: Mind Reading with Magnetic Induction

In a 1909 book debunking spiritualistic phenomena, an amateur magician named David Abbott (1977) wrote of experiments using induction to transmit voices. Abbott’s brother-in-law, an engineer at Western Electric, had recently seen a wireless telephony device demonstrated at White City.[5] Abbott adapted this device to the purpose of illusion, concealing a coil and telephone receiver in the body of a papier-mâché kettle. During the performance of one of Abbott’s magical acts, voices were transmitted via a hidden Dictaphone to an accomplice in another room. The accomplice, in turn, could speak into a telephone transmitter that was attached to a large coil hidden under the performance room. Through induction, the coil transmitted the accomplice’s voice into the body of the kettle. Held anywhere in the performance space, the kettle appeared to speak to Abbott and his audience. He soon modified the method to suit stage magic. Instead of direct contact, a coil on one’s person was all that was needed to get voices out of a receiver. Moving about freely, the performer could surreptitiously receive information at any time, while the audience remained entirely unaware. The induction method soon spread beyond Abbott’s own performances. By the 1930s, he had produced talking tea kettles for magicians around the world. They were joined by talking vases, lamps, Buddha statues, and mind reading sets available by mail order from suppliers such as Thayer Magic in California.

Magicians put induction to work for a variety of purposes, yet they abandoned the technique fairly quickly, despite much practice and refinement in the early twentieth century. With the current renewal of interest in near-field effects and communications, we were motivated to revisit these experiences and techniques, particularly to explore questions regarding their aural history. The attempt to recapture the historical experience of sound has received considerable scholarly attention in the last two decades (Corbin 1998; Johnson 1996; Thompson 2004). But these valuable studies have privileged scholarly imagination over the historically minded re-creation of objects and effects. What might these induction devices and effects have sounded like? If a kettle or vase was supposedly a spirit voice, then how could a spirit plausibly sound? For the performers who were literally wrapping themselves in the technology, what was it like to receive information? And, with contested claims over the effectiveness of magnetic induction, can we determine which methods might have worked best and under what circumstances?

In the case of stage magic, the original device or exact replicas are often not available. They have been lost or destroyed, are held in private collections, or are otherwise beyond our resources to acquire or study. Nevertheless, it is possible to experiment with these technologies in a way that is sensitive to existing sources. Elliott’s prior work on the technology and culture of stage magic has shown that there is a great deal to be learned by attempting to re-create these devices in some form (Elliott, MacDougall, and Turkel 2012). One major insight involves intentional lacunae in historical records. Some explanations for magical illusions are incomplete or incorrect — descriptions and drawings might be missing crucial elements, components might not function as depicted, and generalizations were used to omit details while still revealing most elements of a secret. Stage magicians balanced the need to keep their own methods secret with a desire to gain reputation or notoriety amongst their peers, and they delighted in revealing the secrets of competitors and publicly debunking mediums. When a magical apparatus is built according to the directions given in historical sources, it often does not work exactly as claimed, but one does not learn that until actually trying to rebuild the apparatus.[6]

Even when the historical sources were not intentionally designed to mislead, they assume a fair degree of tacit knowledge, and experimentation with small models has been useful for bringing some of this experience into sharp relief. For example, the original descriptions of how to accomplish levitation effects often do not include information about sets, accompanying props, backdrops, lighting, and other details that were essential for proper performance. Attempting to restage small-scale forms of the performance quickly brings those issues to the forefront, as elements that are supposed to be hidden are blatantly obvious, lighting casts unforgiving shadows that reveal how an effect is accomplished, or backdrops signal actions that were supposed to be concealed. Model-scale experiments can also be recorded and examined after the event and compared with recordings where conditions were controlled or altered.

Elliott’s models of induction illusions use wire coils, circuits, audio pickups, and telephone transmitters that allow him to record and measure the results of transmitting at various powers and volumes, with stationary and moving receivers, and with varying orientations between the receiver and the transmitter coil. These models are by no means replicas of the originals, but they give him a sense of how the originals may have worked, what they may have sounded like, and how to provide a more nuanced interpretation of extant representational sources of past magical performances. And they allow him to share these more embodied and tacit understandings with other people in a workshop setting. At the end of the day, Elliott has not provided definitive answers to any of the questions that motivated his exploration, but we would not expect him to, any more than we would expect the players or audience of Hamlet to come to a satisfactory conclusion about the meaning of life. What we do learn from performative engagements like these is that they are, for us at least, very much worth doing. They give us a nuanced feeling for past technologies and an imaginative sense of historical milieu that we have not found any other way to access. As Michael Polanyi wrote in The Tacit Dimension (2009), “We can know more than we can tell” (4, emphasis in original). Humanistic fabrication allows us to share some of what we cannot put into words.

Case 2: Translating the Sonde into Code

Reverse engineering and simulating an analog hardware device in the form of digital software provides a distinct set of opportunities for hands-on learning. In this section we describe Muhammedi’s re-creation of the Sonde, one of Hugh Le Caine’s lesser-known devices for composing and producing electronic music. Born in 1914, Hugh Le Caine was a Canadian pioneer in the field of electronic music (Young 1989). He studied applied science at Queen’s University and spent many years working for the National Research Council as a physicist. Throughout that time he was an avid creator of musical instruments. He is most famous for inventing the Sackbut, one of the world’s first synthesizers, in 1945. When he switched to being a full-time musician in 1954, Le Caine was able to create a host of new electronic instruments, including the Touch Sensitive Organ (1955), the Multi-track Tape Recorder (1955), the Oscillator Bank (1957), and the Serial Sound Structure Generator (1965).

The Sonde was one of Le Caine’s final projects before he retired, adapted from a previous instrument of his, the Tone Mixture Generator. The Sonde’s key feature was its ability to produce 200 sine waves simultaneously, ranging from five to 1,000 Hz in five-Hz increments (Le Caine and Ciamaga 1970). Using matrix generation, 10 convertor oscillators spaced by 100 Hz were connected with 20 fixed oscillators ranging from zero to 95 Hz. By subtracting the frequencies of the combined sine waves, multiple sine waves could be heard at once. The amplitudes of each of the 200 generated tones could be controlled independently using a vertical slider in addition to an overall volume control. This interface afforded the creation of complex sine-tone mixtures, eliminating the need to continuously record and combine individual sine waves by dubbing. Ultimately, the Sonde reduced the cumbersome tape hiss generated when recording multiple sounds over the same reel.

For this project, Muhammedi used a visual programming language called Max 6, in which code consists of symbolic messages and audio data communicated in real-time between a collection of objects using virtual “patch cords.” Max is a particularly appropriate choice for this project, since the language was originally developed by electronic musicians using the metaphor of a network of analog devices connected by physical patch cords. Many software objects in Max directly correspond to the electronic equipment of analog studios: sine wave oscillators, switches, filters, oscilloscopes, and so on. The Sonde itself was one such analog electronic device.

Working in software has many well-known advantages. The first is drastically reduced costs. Adding another sine wave oscillator to a software synthesizer is basically free, whereas building another electronic module is definitely not. Although there is a licensing fee for programmers who want to code in the Max language, a runtime module can be used to freely distribute a Max program to as many people as possible.[7] The second advantage is accessibility: the software re-creation allows anyone to utilize and experiment with the digital version. Individuals are free to generate their own sounds, increasing the opportunities for public interaction and engagement. The final benefit is that software has very different affordances than hardware. As we explored different ways to represent the components of the Sonde, we continually confronted these differences. In his interface, for example, Le Caine used a multi-pole, multi-position crossbar switch once commonly used for telephony. How closely should we attempt to model this piece of now unfamiliar technology? Real analog equipment has a warm and somewhat unpredictable sound that is now highly valued among electronic musicians. Should we add additional code to simulate that?

Le Caine’s original instrument used 30 physical oscillators wired to telephone switches, run through a final filter, and then amplified to create the desired sound. In the early 1970s, the University of Toronto Electronic Music Studio improved upon the Sonde, replacing the sliders of the original prototype with a grid of printed circuit keys (Le Caine and Ciamaga 1970). Since we were prototyping in software, implementing versions of both interfaces was relatively easy, and it gave us a glimpse of some changes occurring in electronic music composition during the late 1960s and early 1970s. At the same time, we recognize that our simulated instruments have none of the tactility of the originals.

Complex sine tone mixtures are readily generated in software, and their sound is familiar to anyone who listens to contemporary ambient electronic music. Although Le Caine anticipated the use of computers for music creation, he never embraced the world of computer-generated music. Gayle Young argues that Le Caine’s understanding of music was rooted in analog terms: “His dislike of stepped systems and disinterest in computers was linked with his inclination toward analogue systems, an inclination that went deeper than a simple use of available (analogue) techniques to what could be called analogue imagination” (166). Le Caine was certainly aware of how computer scientists might improve upon the Sonde but felt that such improvement would alter his own appreciation for analogue musical instruments. In our case, working with digital simulations of Le Caine’s creation continually drew our attention to material aspects — especially the cadence of adjusting controls to produce desired effects — that we had not yet considered and also led us to experiment with analog synthesis and live electronics (Collins 2010). We cannot be sure what tacit features Le Caine most appreciated when composing with the Sonde, but we do have a new appreciation for some of the possibilities the device may have afforded, and we hope that others who play with it have a similar experience.

Case 3: Thinking with Analogs

There was a time when electronics meant something other than digital. For decades spanning the middle of the twentieth century, making electronics meant building analog machines. Like their namesakes in literature and biology, electronic analogs sought to establish correspondences and harmonies. In place of tropes and organisms, however, they took phenomena like time and motion and rendered them as physical quantities — voltages, positions, and currents. Analogs in the humanities and the natural sciences served the interests of explanation, classification, and clarification. Electronic analogs, by contrast, served the interests of calculation, modeling, and intervention. They found their way into everything from music synthesizers to automata, and ultimately computers. Through that remarkably simple premise — that one object might stand for another — analog computers sought to represent the flux of the world, as well as the possibilities for controlling it, through the continuous variation of electrical values.

What happened to analog computers? One account, often repeated in the history of computing, suggests that they died out because they were not fast enough and did not lend themselves to the kinds of pressing applications that digital machines handled easily. That view turns out to be inaccurate in important ways. As historian David Mindell (2002) explains: “Analog and digital arose together, as distinct but related approaches to representing the world in machines. In general, historians of computing have neglected analog computing, viewing it primarily as an obsolete predecessor to digital. . . . On the contrary, we have not yet begun to understand the history and significance of analog computing, especially the relationship between analog and digital machines” (10). Up to the 1960s, analog computing was faster, more advanced, and more obviously applicable to problems of machine control than were digital computers (Edwards 1997). Analog computers were also easier to integrate with the analog world of things. While digital computing handled discrete entities well, analog computing was the better match for continuous and fluid systems. Analog devices were well suited for measurement and instrumentation, simulation of movement, visualization via pen plotters and oscilloscopes, and control of actuators such as servo-mechanisms (Paynter 1955). One author of a text on analog circuits suggested that a problem with their widespread adoption was the lack of a compelling and inexpensive visual output display (Smith 1971). By contrast, the analog computers at the heart of electronic music synthesizers produced compelling audification (rather than visualization) and thus became quite popular in devices such as the Moog synthesizer (Pinch and Trocco 2004).

In many ways, then, digital computing was a solution in search of a problem (Pinch and Bijker 2004). And even as they eventually supplanted many analog devices, digital devices continued to be shaped by techniques originally developed around analog machines (Williams 1991). This tendency is especially common in places where digital devices have to interface with users in a material world. Analog signals like sound, light intensity, and mouse movements are digitized by cell phones, cameras, and personal computers; that digital information, in turn, has to be converted back into the sounds we hear, the images we see, and so on. Today, solid state analog devices are ubiquitous (Turkel 2011).

In re-creating analog electronic computers, we attempt to understand what has been lost in the overriding emphasis on contemporary technologies and to experience some of the richness and complexity of a time when electronics were not exclusively or necessarily reduced to the digital. Additionally, the practice materializes one of the central arguments of technology studies from the last three decades — that the development of technology involves historical contingencies. The decline of analog was never as necessary, or absolute, as the standard story would have us believe. Rebuilding analog computers is an attempt to trace out what Andrew Pickering (2010) has called “sketches of another future.” Beyond the physical devices produced, our acts of construction generate insight into current scholarly issues and practice. We aim, then, to arrive at that long-standing argument about contingency in a novel way — through the performative engagement with electronic artifacts. In addition to contributing to the history of electronics and to broader methodological questions within the field of technology studies, both our subject matter and approach provide innovative opportunities for teaching within and beyond the classroom setting.

Analog computing is an exemplary subject for staging hands-on historical investigation, because there are often a number of different ways to accomplish a particular goal with analog circuitry. For example, one of the central components of analog computing (both before and after the invention of the transistor) was a device known as an operational amplifier (Frederiksen 1988; Jung 2004). Op amps, as they are commonly known, were created by combining a number of vacuum tubes or transistors into a module with certain properties. In the ideal case, the op amp was a kind of universal building block. By adding a few additional parts, it could be turned into any number of other useful devices to perform mathematical operations such as summing, averaging, integrating, differentiating, or computing logarithms (Paynter 1955). Op amps were also designed so that they could be substituted for one another without affecting the function of the larger circuits of which they were a part. Real op amps, of course, fell short of the ideal. Analog circuit designers faced opportunities and constraints that cannot be fully understood without exploring the electronic behavior of the circuits that they built.

To understand those opportunities and constraints, we designed and built a small analog computer loosely based on the Heathkit EC-1, which was sold by the Heath Company in kit form beginning in 1959. The original kit used lethal voltages and now-scarce components, so our re-creation is really an “analog” of the original. Working from the design, construction, and programming processes laid out in the schematic and instruction manual for the Heathkit EC-1, we substituted readily available, lower-voltage electronic components and integrated circuits to arrive at a system that works in the same fashion (Heath Company, Operational Manual; Heathkit Assembly Manual). Along the way we studied a number of contemporary physical artifacts and texts from mid-twentieth-century analog computing, solving typical electronics problems using the methods most familiar to analog computer designers and programmers of the 1950s and 60s (Johnson 1956; Korn and Korn 1956; Paynter 1955). Commercially available kits such as the Heathkit EC-1 generally attempted to teach electronics via moralized, highly scripted assembly projects. And somehow these kits conveyed the implicit knowledge essential to electronics construction. Our building of analogs, by differently skilled collaborators, helps us to understand the process by which kits taught the kind of “fingertip” specialized knowledge that cannot be learned from books.

Our performative engagement with analog devices is paired with representational work geared to both academic and broader audiences. Within the scholarly realm, we are interested in critically reinterpreting the role of analog devices in the history of computing and scientific instrumentation and expanding upon the value of hands-on methods within the humanities (Jones-Imhotep and Turkel; Turkel and Jones-Imhotep; Turkel, Muhammedi, and Start 2014). In addition, we are producing schematics for a series of open-source electronic construction modules that invite others to explore analog computing. These modules could be adopted within history and social science classes about technology, and perhaps even within electronics curricula. They could further serve as independent projects that would engage enthusiasts of making in both a hands-on electronics activity and thinking about social-historical issues surrounding computing.

Conclusion

In this chapter we have briefly described our experiences collaboratively reconstructing stage magic illusions, electronic musical instruments, and analog electronic computers in settings where the emphasis is on performative engagement with the world. In doing this work, we are not invested in making exact reproductions. Even if we could somehow create them, they would have very different meanings and affordances for us than for their original creators and users (Turkel 2011).[8] When we share these experiences with members of the general public, they seem to find them as fun and absorbing as we do. Academic participants lose themselves in these playful explorations, too, although some feel an obligation afterwards to try to make their new understandings more explicit. Here we outline four trajectories to connect our work to more traditional kinds of knowledge representation.

First, the affective. The fact that doing this kind of work is fun and compelling (to the point where some people have a hard time believing it is work) suggests that the scholarly study of play and games may have something to teach us about what is happening in these settings. Our view of our work as “ontological theatre” establishes some common ground with theories of performance. Our various audiences have found our model-scale stagings (of magic illusions, for example) to be particularly captivating, something we relate to a widespread fascination with miniatures, dollhouses, and puppet theatres, as well as an interest in microworlds more generally. On a more intimate level, we have often found that we experience a sense of loss or absence as we engage more deeply with forms of life we think are now gone. Jentery Sayers[9] suggests that this sense may be produced by the acts we are engaged in, rather than the things themselves, an idea we plan to pursue in future work.

Second, the artifactual. The philosopher Davis Baird (2004) points out that “‘Craft knowledge,’ ‘fingertip knowledge,’ ‘tacit knowledge,’ and ‘know-how’ are useful concepts in that they remind us that there is more to knowing than saying. But they tend to render this kind of knowledge ineffable. Instruments have a kind of public existence that allows for more explicit study” (18). Baird argues that an artifact may act as a model, as a device that creates phenomena, as an instrument for measuring or computing, or as some combination of the three. Like texts, models can provide representations. Physical devices, however, can also “constitute knowledge in a different, nonrepresentational way. Such instruments work epistemologically in a manner that draws on pragmatist conceptions of knowledge as effective action. A fundamental difference, however, is that with instruments, the action has been separated from human agency and built into the reliable behavior of an artifact” (12). Hands-on exploration provides technology studies scholars with a way to acquire technical and scientific knowledge that is tacit and embodied, and that can be learned best through experimentation, trial, and error (Collins 2009).

Third, the material. We have a preference for certain kinds of components and workflows. Whenever possible we work with modular and reusable systems,[10] we favor reclaimed and recycled materials, and we mercilessly compost, reuse, or continue to hack the things we have already made. This approach not only allows us to embrace the accidents and failures that are crucial elements of design thinking; it also enables us to maintain our focus on process rather than product and to offer frequent workshops. We document as much as possible, and we version and share our software on GitHub and elsewhere. Our work can thus be situated both as a kind of commons-based peer production (Benkler 2006) and as part of the so-called maker movement (Charny 2011).

Fourth, academic fellow travelers. Our engagements can be seen as a kind of “research-creation,” because they “typically integrate a creative process, experimental aesthetic component, or an artistic work as an integral part of a study” (Chapman and Sawchuk 2012). When we do make the detour through the space of representation, we engage in what Philip Agre (1997) calls “a critical technical practice — a technical practice for which critical reflection upon the practice is part of the practice itself” (xii). The artifacts we make might be seen as “provotypes” (Mogensen 1992). Although they are never exact re-creations of historical objects, provotypes have the ability to provoke “discussion about current practice through practice” (Donovan and Gunn 2012). In re-creating historical technologies, we raise both historical and historiographical questions about the role and salience of material cultures. Our practice of historical reverse engineering is related to the present-oriented scholarship of critical making. Like reverse engineering, critical making combines the methods of “two typically disconnected modes of engagement in the world — ‘critical thinking,’ often considered as abstract, explicit, linguistically-based, internal and cognitively individualistic; and ‘making,’ typically understood as material, tacit, embodied, external, and community-oriented” (Ratto and Hockema 52). In this way, reverse engineering electronic devices teaches us about their inner workings and, when complemented by careful historical analysis and traditional sources, gives us a glimpse into the decisions, constraints, and contingent choices of their designers and makers. But it also combines creation and critical analysis to expand our understanding of how scholarly knowledge is generated and communicated (Galey and Ruecker 2010). Each performative engagement opens new fields for play, experience, and discovery; teaches us things we can put into words and things we cannot; and creates shareable artifacts that can be mobilized and studied in their own right.

Notes

Bibliography