Half a century ago, the development of transistors and switching regulators
completely changed computer power supply design
Tech enthusiasts may know the model of a computer’s microprocessor and the size of its physical memory, but they are likely to know nothing about the power supply. Even for manufacturers, power supply design is often an afterthought.
This is unfortunate because designing the power supply for a personal computer requires a significant amount of effort, representing a huge advancement compared to the circuits used in other consumer electronics before the late 1970s. This breakthrough stemmed from the leap in semiconductor technology half a century ago, particularly improvements in switching transistors and innovations in integrated circuits. However, this is a revolution that is not known to the public, and even those familiar with the history of microcomputers may not understand it.
The power supply does have its advocates, and one of them might surprise you: Steve Jobs. Walter Isaacson, the authorized biographer of Jobs, said that Jobs had a deep affection for the power supply of the Apple II and its designer Rod Holt. Isaacson wrote:
Holt did not use a traditional linear power supply but instead created a product similar to the power supply used in oscilloscopes. It switches power more than 60 times per second, but thousands of times, allowing it to store energy in a shorter time and reduce heat release. Jobs later stated, “This switching power supply is as revolutionary as the Apple II logic board. Rod hasn’t received much acclaim in history, but people should remember him. Every computer now uses a switching power supply, and they all copied Rod Holt’s design.”
Jobs’s insights are profound, but I do not completely agree with his view, so I did some research. I found that switching power supplies are revolutionary because they replaced simple but inefficient linear power supplies, but this revolution occurred from the late 1960s to the mid-1970s. The Apple II, launched in 1977, benefited from this revolution, but it was not the cause of it.
Correcting Jobs’s insights is not just an engineering detail. Today, switching power supplies are ubiquitous, and we use them daily to charge smartphones, tablets, laptops, cameras, and even some cars. They power clocks, radios, home audio systems, and other small appliances. The engineers who truly led this revolution should be recognized. This is also a story worth telling.
The power supply for desktop computers like the Apple II can convert AC circuit voltage to DC, providing a very stable voltage for the system. Power supplies can be constructed in various ways, with linear and switching being the two most common.
A typical linear power supply uses a bulky transformer to convert relatively high-voltage AC from the power line to low-voltage AC, and then uses a classic bridge structure typically composed of four diodes to convert AC to low-voltage DC. Large electrolytic capacitors are used to smooth the output of the diode bridge. Computer power supplies use a circuit called a linear voltage regulator, which can reduce the DC voltage to an ideal level and maintain a constant voltage even when the load changes.
Designing and manufacturing linear power supplies is not complicated. They use inexpensive low-voltage semiconductors, but they have two main drawbacks. One of them is that they require large capacitors and bulky transformers, which can never be as compact, lightweight, or convenient as chargers for smartphones and tablets. The other is the linear regulator, a transistor-based circuit that converts excess DC voltage (voltage above the specified output voltage) into waste heat. Therefore, typically more than half of the energy consumed by such power supplies is wasted, and they often require large metal heat sinks or fans to dissipate heat.
The operation of switching power supplies is quite different: in a typical switching power supply, the AC line input is converted to high-voltage DC, switching tens of thousands of times per second. The high frequency used allows for lighter transformers and smaller capacitors. Special circuits can precisely measure switching times to control output voltage. Such power supplies do not require linear regulators, thus wasting very little energy. Their efficiency is typically between 80% to 90%, so they also release much less heat.
However, switching power supplies are much more complex than linear power supplies, making them more challenging to design. Additionally, they have higher component requirements, demanding that high-voltage power transistors can effectively switch at high speeds.
By the way, some computer power supplies are neither linear nor switching power supplies. There is a primitive but effective power supply technology that allows a motor to run without power and uses that motor to drive a generator to produce the required output voltage. Electric generators have existed for decades, at least dating back to the 1930s with the IBM card punch, and have continued to be used in Cray supercomputers of the 1970s.
Another scheme that was popular from the 1950s to the 1980s was the use of ferromagnetic resonant transformers—a special transformer that provides a regulated output voltage. There were also saturable reactors, which are controllable inductors used for voltage regulation in vacuum tube computers in the 1950s. It reappeared as a “magnetic amplifier” in some modern personal computer power supplies, providing additional voltage regulation. Ultimately, these quirky methods largely gave way to switching power supplies.
Electrical engineers have known the principles of switching power supplies since the 1930s, but their application in the vacuum tube era was limited. At the time, some of the special mercury-containing tubes used in some power supplies (known as thyristors) can be seen as primitive low-frequency switching regulators, such as the REC-30 teletype power supply from the 1940s and the power supply used in the IBM 704 computer in 1954. The introduction of power transistors in the 1950s led to rapid development of switching power supplies. Pioneer Magnetics began manufacturing switching power supplies in 1958. General Electric published an early design of a transistor switching power supply in 1959.
In the 1960s, NASA and the aerospace industry provided major impetus for the development of switching power supplies, as space applications focused more on the advantages of small size and high efficiency rather than low cost. For example, the 1962 Telstar communication satellite (the first satellite to transmit television images) and the U.S. Minuteman missile both used switching power supplies. Over time, the cost of switching power supplies gradually decreased and became more widespread. For instance, in 1966, Tektronix used switching power supplies in portable oscilloscopes, cutting off power currents or powering them with batteries.
As power supply manufacturers began selling switching devices to other companies, this trend began to accelerate. In 1967, ROAssociates introduced the first 20-kHz switching power supply product, claiming it was the first commercially successful switching power supply. Nippon Electronic Memory Industry Co. began developing standardized switching power supplies in Japan in 1970. By 1972, most power supply manufacturers were either selling or preparing to sell switching power supplies.
About this time, the computer industry began using switching power supplies, including the PDP-11/20 microcomputer from Digital Equipment in 1969 and the 2100A microcomputer from Hewlett-Packard in 1971. An industry publication in 1971 noted that companies using switching voltage regulators “read like a who’s who of the computer industry, such as IBM, Honeywell, Univac, Digital Equipment, Burroughs, and RCA.” By 1974, small computers using switching power supplies included Nova2/4 from General Data, 960B from Texas Instruments, and systems from Interdata. In 1975, switching power supplies were applied to the HP2640A display terminal, IBM’s Selectric Composer, and the IBM 5100 portable computer. By 1976, General Data was using switching power supplies in half of its systems, while HP used them in smaller systems like the 9825A desktop computer and the 9815A calculator. Switching power supplies also appeared in household appliances at that time, powering some color televisions.
At that time, electronic magazines widely reported on switching power supplies, both in advertisements and articles. As early as 1964, Electronic Design recommended switching power supplies that could improve efficiency. The October 1971 cover of Electronics World featured a 500-watt switching power supply and an article titled “The Switching Regulator Power Supply.” In 1972, Computer Design explored switching power supplies in detail and discussed their increasing popularity in computers, although it also mentioned that some companies remained skeptical. In 1976, the cover article of Electronic Design claimed, “Suddenly, switching has become easier,” introducing a new switching power supply controller integrated circuit. Electronics published a lengthy article on the matter; Powertec ran a two-page advertisement promoting the advantages of its switching power supplies, with the slogan “The Switching Revolution Changes Dramatically”; and Byte announced that Boschert would provide switching power supplies for microcomputers.
Robert Boschert is one of the key developers of this technology. He quit his job in 1970 and began making power supplies at his kitchen table. He sought to simplify the design so that it could compete with linear power supplies in cost. By 1974, he had mass-produced low-cost printer power supplies, and in 1976, he produced low-cost 80-watt switching power supplies. By 1977, Boschert had grown into a company with 650 employees. It manufactured power supplies for satellites and Grumman F-14 fighter jets, later producing computer power supplies for companies like HP and Sun.
In the late 1960s and early 1970s, companies like SSPI, Siemens Edison Swan Company (SES), and Motorola introduced low-cost high-voltage high-speed transistors, propelling switching power supplies into the mainstream. The faster switching speed of transistors can improve efficiency because heat is dissipated during the switching process, and the faster the device can perform this conversion, the less energy is wasted.
At that time, the speed of transistors was experiencing exponential growth. Indeed, transistor technology was advancing rapidly, and the editors of Electronics World even claimed in 1971 that the 500-watt power supply on the cover could not have been made with the transistors that had only been introduced 18 months earlier.
Another major breakthrough occurred in 1976. At that time, Robert Mammano, co-founder of Silicon General Semiconductor in the U.S., introduced the first integrated circuit to control teletype switching power supplies. His SG1524 controller integrated circuit greatly simplified power supply design, reduced costs, and drove sales through the roof.
Before and around 1974, anyone with a slight understanding of the electronics industry knew that a true revolution was taking place in power supply design.
The Apple II personal computer was launched in 1977. It featured a compact, fanless switching power supply capable of providing 38 watts at 5 volts, 12 volts, -5 volts, and -12 volts. It used Holt’s simple design, known as the offline flyback converter topology. Jobs claimed that every computer today copies Holt’s revolutionary design, but was this design truly revolutionary in 1977? Did other computer manufacturers really copy it?
The answer is no. At that time, companies like Boschert were also selling similar offline flyback converters. Several specific features of Holt’s power supply received patents, but these features were never widely used. The control circuit built with discrete components, which is the construction method of the Apple II’s circuit, also proved to be a technical dead end. The future of switching power supplies belonged to dedicated controller integrated circuits.
If there is a microcomputer that has indeed had a lasting impact on power supply design, it should be the IBM personal computer launched in 1981. At that time, just four years after the release of the Apple II, power supply technology had undergone a tremendous transformation. Both early personal computers used offline flyback power supplies with multiple outputs, but their similarities ended there. Their drive, control, feedback, and regulation circuits were not the same. The IBM personal computer power supply used an integrated circuit controller, with components about twice that of the Apple II power supply. These additional components better stabilized the output voltage and would signal “power good” when all four voltages were accurate.
In 1984, IBM released an upgraded personal computer called the IBM Personal Computer AT. Its power supply featured a variety of new circuit designs, completely eliminating the earlier flyback topology. It soon became the industry standard, until Intel introduced the ATX form factor specification in 1995. The ATX form factor specification defined the ATX power supply, which remains an industry standard today.
With the emergence of the ATX standard, computer power supply systems also became increasingly complex. In 1995, the high-energy Pentium microprocessor was introduced, requiring lower voltage and higher current, which the ATX power supply could no longer meet. In response, Intel introduced the voltage regulator module (VRM)—a DC-DC switching regulator mounted next to the processor. It reduces the power supply’s 5-volt output to the 3-volt voltage used by the processor. Many computer graphics cards also include voltage regulator modules to drive their high-performance graphics chipsets.
Today, fast processors may require voltage regulator modules to provide up to 130 watts of power, far exceeding the 0.5 watts required by the Apple II’s 6502 processor. In fact, a single modern processor chip can consume more than three times the power of the entire Apple II computer.
The increase in power consumption of computers has become a catalyst for environmental issues, leading to initiatives and regulations aimed at improving power supply efficiency. In the U.S., government Energy Star and over 80 industry-leading certifications have encouraged manufacturers to produce more “green” power supplies. Numerous technologies can achieve this: more efficient standby power supplies and startup circuits, resonant circuits that reduce switching transistor power loss, and “active clamp” circuits that replace switching diodes with more efficient transistor circuits. Over the past decade, advancements in power MOSFET technology and high-voltage silicon rectifier technology have also been driving improvements in power supply efficiency.
In other areas, switching power supply technology continues to evolve. Today, many power supplies no longer use analog circuits but instead use digital chips and software algorithms to control output. The design of power supply controllers is no longer just a matter of hardware design; it’s also a matter of programming. Digital power management supports communication between the power supply and other parts of the system, improving efficiency and logging. These digital technologies are currently primarily used in servers but have also begun to influence the design of desktop computers.
Jobs believed that more people should know about Holt, and he thought, “Rod has not received much acclaim in history, but people should remember him,” yet the true history is otherwise. Even the best power supply designers are unlikely to be known by outsiders. In 2009, the editors of Electronic Design welcomed Boschert into the engineering hall of fame. Robert Mammano received the Lifetime Achievement Award from Power Electronics Technology in 2005. For his innovations in switching power supplies, Rudy Severns received another Lifetime Achievement Award in 2008. However, these outstanding figures in power supply design are not even well-known on Wikipedia.
Jobs repeatedly emphasized that Holt was overlooked, and as a result, Holt’s design appeared in dozens of popular articles and books about Apple, including Paul Ciotti’s “Revenge of the Nerds” published in California Magazine in 1982 and Isaacson’s bestseller “Steve Jobs” in 2011. Therefore, ironically, although Rod Holt’s design for the Apple II was not truly revolutionary, he may have become the most famous power supply designer in history.
