L.E.D.

by Bob Johnstone

  The first objection was glare. The light that the unshielded LEDs were shining was essentially too bright for pedestrians. It dazzled dog walkers and student cyclists alike. The second was light trespass. The new fixtures were installed on existing poles whose arms tilted upward. This was unfortunate. The angle ensured that light penetrated upper-storey bedroom windows. The effect, according to one vexed home-owner, was “more like a spotlight” than a street light: it was like living in a prison. Third, citizens objected to the color of the light, arguing that it was too blue. Los Angeles had chosen a cool color temperature of 4,000 degrees Kelvin. That was approximately equivalent to moonlight, which seemed appropriate for night-time illumination, and was markedly better than the orange of sodium-vapor. But Davis was not a rough-and-tumble inner city. Residents felt that LED light ruined the night-time appearance of their cozy little college town. Aware that LEDs for indoor applications were being marketed as “warm white,” a temperature of 2,700 degrees Kelvin, they wanted nice warm light outside, too. The city council quickly gave in to their demands, replacing the super-bright cool lights with lower-output warm ones. It cost them an additional $350,000.

  The moral of the story was clear: rather than rushing out to replace their street lights, cities should begin with careful planning that took residents’ requirements into consideration. But many municipalities lacked the resources and the technical expertise needed to design and implement a successful LED street lighting upgrade project. That was why the Municipal Solid-State Streetlights Consortium and its efforts to disseminate best practice were so important. Despite the hiccups, however, there could be no disputing the overall trend. By 2013, according to a DoE report, LEDs had captured almost 14 percent of the US street lighting market. The report predicted that solid-state street lighting would reach a share of 83 percent by 2020 and nearly 100 percent by 2030. Street lights would be one of solid-state lighting’s first big success stories. And there were also, as we shall see, others.

  C H A P T E R N I N E

  Chip Heads vs Metal Benders F or LED lighting, bulbs were just the beginning. Beyond bulbs lay the much larger fixtures market, which accounts for around three quarters of the lighting industry’s annual sales. Unlike LED bulbs, which were essentially a replacement market, the market for LED fixtures was growing and seemed likely to continue doing so for the foreseeable future. With bulbs, the same products were sold around the world. With fixtures, by contrast, the market was local and fragmented. In the US, lighting manufacturers introduced their first LED fixtures in 2008. The construction industry was then deep in the doldrums following the collapse of the housing market. Demand was initially driven by retrofits, which even in a normal year account for two thirds of lighting sales. Retrofit is a cyclical business: when the lease on a commercial property expires, the incoming tenant typically guts the premises, installing new fixtures and fittings. By 2013, with construction gradually recovering from its slump, new projects began to take up the slack. As the price of LEDs plummeted and utilities began offering substantial rebates, the return on investment that solidstate lighting could deliver became increasingly attractive to building owners. At the same time owners were obliged to comply with building codes, like California’s Title 24, that mandated energy efficiency. In consequence sales of LED fixtures took off.

  A fixture (or “luminaire,” the fancy French term the industry prefers) is defined as a complete lighting unit. In its most basic form it consists of a lamp, a socket to hold the lamp, optics (reflectors, refractors, lenses, and diffusers) to spread the light that the lamp produces, and wiring to connect the lamp to a power supply. Fixtures came in a bewildering variety of shapes: round, oval, square, rectangular, long-andthin, short-and-fat; also of sizes, everything from tiny pencil-beam spotlights to gigantic stadium floodlights. Some fixtures were hung from the ceiling, like chandeliers, pendants, domes, and track lights. Others were recessed, like downlights, troffers, and cans. Still others were hidden, like coves and under-cabinet lights. There were also decorative fixtures, like wall sconces, washers, and accents. And, for illuminating personal spaces such as desk tops, task lights.

  People sometimes talked about the lighting market as though it were a single entity. In fact, the market broke down into at least seven major segments. Of these, residential was by far the largest, accounting for around a third of total sales. This was followed in order of size by outdoor (streets and roadways, landscape), retail (stores), architectural (aka “specification grade”), commercial (offices), hospitality (hotels, casinos, cruise ships), and industrial (factories and warehouses). These broad categories could be further subdivided into scores of niches.

  In North America, four big fixture makers - Acuity, Philips, Cooper, and Hubbell - shared just over half the market between them. Acuity was the largest, with annual sales in 2015 of around $2.5 billion. The company claimed that it offered the industry’s most extensive portfolio of lighting solutions. It was made up of more than 1.7 million stock-keeping units - “SKUs” in industry jargon - which (confusingly) meant individual options, not inventory items. A downlight for example might come with a white or gold trim, a black or silver baffle, a warm or cool light source, each variation counting as a separate SKU. (Architects prefer their fixtures bespoke.) But even Acuity, despite the scope of its “SKUniverse,” could not cater to all lighting needs. Beneath the big four, the next layer of the pyramid consisted of literally hundreds of smaller fixture makers. Most were family-owned firms bearing brands that few outside the industry would have heard of. Many served highly specialized applications, like airport runways or aquarium lighting.

  In addition to being populated by minnows, the luminaire market is what economists call “high-mix, low-volume.” That is, a maker might sell a million fixtures a year in total, but the highest sales for any individual fixture might be as little as a few thousand units. There were reckoned to be over two and a half million commercial buildings in the US alone. Most were equipped with “legacy” lighting that was more than ten years old and in need of replacement. An office retrofit might call for hundreds of lights, but the purchase order would likely break down into two dozen of this, a dozen of that, ten of these, and just a few of those.

  Yet another peculiarity of the lighting industry was the way manufacturers sold their fixtures to clients. Over more than a hundred years, to connect upstream with downstream, the industry had carved itself a labyrinthine network of sales channels. Each fed a different type of customer. Acuity boasted that it served no fewer than 14 different channels. There were essentially three types of channel. The first two - direct sales, between manufacturer and end user; and retail, meaning relatively high-volume, low-price sales via big box stores like Home Depot and specialty lighting outlets - catered mostly to the residential market. The third type was wholesale, which served professional buyers like electrical contractors. In this latter, highly technical category, which accounted for perhaps sixty percent of the market, access to purchasing decision-makers was key. Channels also differed depending on the application. For example, though street lights and parking garage lighting both fell under the heading of “outdoor,” each had its own separate sales channel. In the former, the buying decision was made by municipalities, in the latter, by private owners. A manufacturer might dominate in street lighting, but that meant nothing in the channel that dealt with interior designers doing high-end residential. We shall return later to examine the puzzle that channels posed newcomers as they attempted to sell their stuff.

  In this well-entrenched good ol’ boys club where nothing much had changed in decades and inertia was the name of the game, the unlookedfor advent of solid-state lighting caused alarm. Especially since it was clear that henceforth, fixtures and lamps would be inextricably integrated. The paradox was that, though nominally known as “lighting” companies, most fixture manufacturers did not actually themselves make lights. Mostly what they did was to stick a new lamp in an old box. Traditionally these firms h
ad been in the business of shaping reflective steel or aluminum housings into which light sources could be bolted or screwed. They referred to themselves as “metal benders.” Few had electrical engineers on their payrolls. Nor were they accustomed to spending much on R&D. That made the metal benders very different than the chip heads from high-tech enclaves like Silicon Valley and Research Triangle Park, who were now aggressively attempting to elbow their way into the lighting market.

  For fixture makers the first sign that their cosy world was about to be disrupted came when salesmen from LED makers arrived knocking on their doors. The sales people glibly explained that, to make a fixture with LEDs, all you had to do was buy an off-the-shelf lens, a heat sink, some driver chips, solder them together on a circuit board, and you had what people had recently started to call, not a light source, but a light engine, because it was made up of many parts. And this was utterly confusing for the fixture folks, who were not accustomed to having to deal with electronics, certainly not at the component level.

  Chuck Berghoff, CEO of OptoElectronix, a San Jose, California solid-state lighting startup, first encountered this predicament while visiting the president of a small fixture maker. Knowing that the company had been showing interest in solid-state lighting for some time Berghoff naturally assumed that they would have something to show him. “What products have you designed with LEDs?” he asked the president. Whereupon the latter just looked at him sheepishly and said, “We understand screwdrivers and wrenches, we connect two wires with a plug at one end and a socket at the other end, but soldering is not a technology we have in our industry.” For Berghoff it was an aha! moment. Seizing his cue, he reached into his briefcase and pulled out one of the industry’s very first LED light engines. It was a round piece of white plastic about the size of a hockey puck with a standard-size screw-hole bored through the middle. Berghoff plugged the puck into the wall. It lit up. He pointed at the little glowing object. “You know, you could just mount this in any of your fixtures right now,” he remarked casually. The president’s reaction was immediate. He leapt from his chair, ran out of the conference room, and yelled at someone down the hall to go grab a fixture from the warehouse. Five minutes later a hastily-rigged LED light fixture sat shining brightly on the conference table. It was late November 2008 and the president wanted to be able to show an LED product to his customers at a trade fair the following January. Would OptoElectronix be able to modify its light engine for such a product in time? “No problem,” Berghoff assured him. “Just FedEx the fixture to my office, we’ll retrofit it and get it back to you by the end of December.” Which they did, the president took the fixture to the fair, and OptoElectronix won the account.

  This was Berghoff’s first experience dealing with a fixture maker. It soon became clear to him that the lighting and semiconductor industries were like chalk and cheese. The mismatch was particularly obvious with regard to their respective product cycle times. In semiconductors speed was of the essence. For more than half a century the chip heads’ mantra had been faster! cheaper! better!. In lighting, by contrast, nothing ever happened quickly. Lighting companies were used to selling the same fixture, more or less unchanged, for years. Makers like Berghoff’s new client would take orders from customers like big-box stores for a product in January, then spend the next few weeks negotiating contracts. The required quantity would be ordered from a manufacturer in China, where virtually all light fixtures are made. The Chinese would assemble the products, package them up, then ship them to the US by sea, a voyage that might take several months. By the time the product actually hit the shelves, it would be Halloween. Meanwhile, LEDs kept getting cheaper and more efficient. That October, before the new fixture had even arrived in the stores, Berghoff went back to the president and told him, “We could build that same product for at least a third less than you paid for it.” What should we do? the astonished executive asked him. “You gotta figure out a way to move things faster,” Berghoff replied.

  Though OptoElectronix was founded in 2006, the company’s origins dated back more than three decades. The “x” at the end of its name was a tip of the hat to Litronix, one of Silicon Valley’s pioneering LED startups. In the early seventies Litronix flourished selling displays in the electronic calculator and digital watch boom. But the firm over-reached itself and in 1976 it was acquired by Siemens, the giant German electrical engineering conglomerate. Chuck Berghoff arrived in Silicon Valley in 1973 to work for what was then a little outfit named Intel. The young man had never heard of Intel but his professor at the University of Minnesota told him the company was doing interesting things with semiconductors. Born and raised in the Midwest Berghoff was eager to try living someplace where it didn’t snow. Back in those heady days Intel was expanding rapidly, forever challenging its employees to take on new responsibilities. Berghoff learned a great deal during his seven years at the company, moving from one job to another. He ended up in marketing, but eventually got bored and quit. For a while Berghoff worked for a firm that supplied Siemens with driver chips. In 1986 he signed on as head of marketing for the Optoelectronic Division at Siemens, which was then one of the world’s largest manufacturers of LEDs.

  Over the next thirteen years, Berghoff turned the division into a multi-million dollar operation. His success was based on a canny strategy. Standard components like individual LEDs had a nasty habit of turning into commodities, where competition was based solely on price. “That got us into the business of taking LEDs, packaging and mounting the chips, then integrating the drive and control electronics into a single unit,” Berghoff told me. The advantage of fully-customized LED products had nothing to do with price and everything to do with their perceived value to the customer. The division’s largest customer was Motorola. Siemens made intelligent displays for Motorola’s StarTAC, the world’s first flipphone, a huge commercial hit in the mid nineties. In building up this custom business, Berghoff was supported by two powerful allies. One was a nationwide network of independent sales reps who brought in business. The reps were beholden to Berghoff who made sure his people responded promptly to their requests. More importantly, with full-custom products, profit margins - hence sales commissions - were much higher than with standard components. Over the years the reps made lots of money doing business with the Optoelectronic Division. Berghoff’s other ally was a remarkably capable individual who in later years he would refer to as the “secret sauce” behind the success of OptoElectronix. His name was Dato Yap Peng Hooi and he was based over eight thousand miles away, in Penang, Malaysia.25

  The manufacture of semiconductors divides into two parts: frontend, meaning chip fabrication, and back-end, meaning packaging the chips so that they can be soldered onto circuit boards. In the industry’s formative years packaging was done manually, by nimblefingered young women sitting shoulder-to-shoulder on assembly lines. During the early nineteen seventies, LED makers sought to lower their labor costs (and gain tax breaks) by moving back-end production offshore. One location that took their liking was Penang, a sleepy tropical island fanned by humid sea breezes a mile or so off the west coast of the Malaysian mainland. Legend has it that on a visit to Penang in 1972, HP co-founder Bill Hewlett was so impressed by the ability of a waitress in a local coffee shop to use his company’s scientific calculator that he decided then and there to establish a packaging plant among the coconut palms in what was formerly a swamp. Litronix soon followed suit.

  25 Dato is an honorary title bestowed by the Malaysian government, in Yap’s case in recognition of his services to the local electronics industry. The boom years of the eighties exhausted the supply of unskilled labor. That left chipmakers with two options: they could either relocate to another low-wage country, like Vietnam, or they could automate the assembly process. Automation would require lots of capital but would reduce the dependence on labor, whose costs always spiralled when demand outstripped supply. Automated manufacture with its highly specialized machines like pick-and-place robots required
expertise to keep the equipment up and running around the clock. That meant hiring and training skilled engineers and technical support staff. An entire infrastructure sprang up to service the chipmakers. Penang also added other competences, like design and testing, to the point where the island could legitimately claim to be the hub of LED production worldwide. “There are thousands of engineers in Penang now, and they all have experience in LED manufacturing,” Berghoff claimed, adding, “here in Silicon Valley, there aren’t any — they’ve all retired!” Though their corporate names have changed, the originals are still there. Today both Lumileds (formerly HP/Agilent) and Osram (formerly Litronix/Siemens) run large, ultra-modern factories on the island. In November 2015 Osram announced a one-billion-euro investment in Malaysia. The company claimed its new plant there would be the largest LED production site - front- and back-end - in the world.

  An accidental recruit to electronics, having graduated in 1972 in mathematical statistics at Melbourne’s LaTrobe University, Yap Peng Hooi had been in Penang since the beginning. Over the years he accumulated a wealth of experience in all aspects of LED manufacture. He ran factories, eventually rising to become managing director of all of Osram’s LED operations in Penang. “We have been living with LEDs since our early years,” Yap told me in 2013, “so we understand LEDs much better than most people,” he added with a chuckle. Perhaps the unassuming Malaysian’s most impressive attribute was what amounted to a genius for tackling technical challenges. You could take any problem to Yap and he would say, “OK, I’ll fix it.” Berghoff would challenge him, “How are you going to do that?” And Yap would reply, “I have no idea.” He would never say no. Berghoff was curious about his partner’s can-do attitude. After many years of working together, he finally came out and asked Yap what his secret was. “When somebody asks me to do something, I either know how to do it, or I don’t,” Yap replied simply. “And if I don’t, all I’m doing is committing myself to figure out how to do it.”