ARM Revolution: How a Chip Changed Tech

It’s late at night when two legendary computer geniuses receive a phone call. What they are asked seems impossible: to build a computer that will change the fate of an entire generation within a week. However, they have yet to learn that this will only be the beginning, initiating a chain of events that will lead a small group of visionaries from Cambridge to lay the foundations for a technology that will likely affect the lives of everyone on Earth. This is the story about ARM Revolution: How a Chip Changed Tech. 

It’s the late 1970s in Cambridge, UK, and the co-founders, Hermann Hauser and Chris Curry, are trying to figure out how to get their small business involved in big projects. Until recently, Chris had been the second in command at a local consumer electronics company run by the legendary Claire Sinclair, which we covered in a previous video with its own set of strange and slightly odd events. But at this time, if he intended to enter the world of microprocessors and home computers, he needed something significant: talented technical staff to make his laptop dreams a reality. Fortunately for them, Hermann knew precisely where the best talent pool was—the legendary University of Cambridge.

Within the University of Cambridge, a specific club piqued their interest: Cambridge University Professor Group. This society was deeply engrossed in the promising world of processors and the computer revolution. After some inquiries, they were introduced to the most promising students in the group, a doctoral student named Steve Furber and a remarkably young student, Sophie Wilson. This encounter would prove to be a turning point in their journey. 

Furber was immersed in his doctorate in aerodynamics and had developed extensive mathematical knowledge, and Wilson, for her part, was a wizard of technology. Her fascination with microprocessors had begun after some friends told her about an astonishingly cheap microprocessor that had just come onto the market, the MOS 6502. Now, with new talent under their wing, they searched for profitable ideas. Having witnessed Apple Computer’s growth, Chris was inspired to choose a name to put it above Apple in the telephone book and referenced an old saying, “Great oaks from little acorns grow,” and Acorn Computers grew into a big tree by 1980.

As a new decade begins, things at Acorn Computer are heating up. Yes, a home computer packed with everything is ready to plug in and use like those manufactured in the United States. There are no kits for assembly for enthusiasts; we can put a beautiful case around our computer and make it a reality—a gray case. How the hell are we going to finance this? We don’t have the money for that level of production. Don’t worry, we’ll make one, take some good photos, put a full-page ad in the best electronics magazine in the country, and people will love it so much that they’ll send us checks.

They were the first orders for which we will have the funds for production in no time. This methodology of selling first and producing later was something Curry had learned from his old boss, Sinclair. Although Hauser had doubts, the growing computer market in the UK thrilled him. They soon obtained the funds to make the Acorn Atom, a wholly pre-assembled and packaged home computer designed around Wilson’s favorite MOS 6502 and powered by Acorn BASIC, a compelling interpreter for the BASIC programming language that Wilson had programmed herself, with additional features such as networking capabilities and the ability to mix machine code with BASIC instructions. In many respects, it was superior to Microsoft BASIC, which most of the industry would come to rely on. 

The product was modestly successful despite competition like the ZX 80 from Cris’s former boss. When the employees of the still-small company discussed what to produce next, they encountered an unexpected disagreement. Should we build workstations? Computers will be revolutionary in the workplace. Workstations, look at all the atoms we sold. We should make another home computer for people. What do you think about this? Can you design a computer that uses two microprocessors? One microprocessor will connect to the IOY and be sold as a home computer. Still, when connected to the second microprocessor, it will become a much more powerful machine that could be used as a workstation. Pau, can you do that? Then, the development of the next project began. It would take the first half of Sophie’s idea and turn it into a home computer with the code name Proton, planning to sell the other half as an expansion to turn it into a powerful workstation.

However, the timeline for this project was about to be radically shortened thanks to an unexpected organization that would change everything. For the country, the BBC, the public broadcaster of the United Kingdom, was producing a television program to educate the general public about computers. At that time, there was no unified operating system on computers; they were a messIt appears there was an error in processing your request of incompatible systems with somewhat unique quirks. So, for the program to work, the BBC needed a single computer to focus on. This presented an opportunity. The entire UK computer industry understood what this would mean. Whoever secured this lucrative government contract would receive millions of pounds in free advertising and sell millions of computers to schools and universities.

Sinclair, Chris’s former boss, pounced on this with his characteristic arrogance, presenting the Sinclair ZX 81 as the obvious candidate for this competition. However, its flat keyboard, lack of expansion capability, and Sinclair’s overall attitude immediately turned off the BBC, and Chris saw an opportunity to reapply the sell-first, build-later approach. “Our next computer is exactly what they need. With plenty of expansions and ports, a high-quality keyboard, high performance, we can easily adapt our Proton for everything the BBC needs. We can visit the office next week to see it next week. Ah, yes, of course, we’ll be ready.”

Now, the problem with this promise was that the Proton was at best in the ideation stage, and convincing the team to undertake the madness of implementing a prototype in a week would need almost a miracle. But Hauser had an idea of how to motivate his workers, something that would become a legend within technology history circles. To tell the story, I thought it would be appropriate to have someone who was there. This was quite late in the negotiations with the BBC; they planned to visit Acorn on Friday, and that phone call was the weekend before that Friday. We hadn’t even designed it, I mean. I had a design on paper for the machine, but not a detailed design. 

So, Hermann called Sophie, told her, “Hello, Wilson, can you create a prototype for production by the end of the week for the BBC to review by the end of the week, are you crazy? That’s impossible. What weird way he said he could, wait he, maybe knows something I don’t? Okay, then Hermann called me and said, “Hello, Steve, can you create a prototype for production by the end of the week? Are you drunk? Hermann, Wilson said it was feasible, how? Well, maybe you know, it was a long time ago, and I don’t trust my own memory and the little details, but. Certainly something like that happened. It took the pair a while to realize the trick, but by then it was too late they were up to their knees in the project. What followed was one of the most intense weeks in computer development history. The whole team worked tirelessly to have the prototype ready for the BBC’s visit. They worked until the last seconds, with Wilson literally programming the last bits of the CRT controller while Curry and Hauser delayed the BBC from coming up the stairs, but the BBC was so impressed to see that this quirky team had made more progress in one week than other companies had in years that they awarded them the multi-million pound government contract. 

And now, with the proper time and money and prototypes needed, it could become a complete product. The enormous amount of chips used for this computer could be simplified with the help of a couple of custom-built chips using what was called a gate array, these chips. 

The problem was that the video processor ran at speeds quite close to the limits of the technology, one of the features of the BC Micro was that it worked well when turned on, but 10 minutes later it warmed up, and you started to have pixel breakups on the screen due to the video processor, and so they were saved by 1 second supplier of similar technology. E N S I 1, an American company using a different manufacturing technology, which was much more resistant to overheating and soon became a powerful ally. All the pieces were in place to finish the new computer that would be called the BBC Micro, the official computer of the BBC’s educational program, which would be sold to millions and become a platform that would teach the entire UK population about computers, but if you remember what had been proposed before it was a 2-part computer, where the BC Micro was only the first part ready to be connected to 1 second processor to become a powerful workstation, and now that they had the money, the momentum, and the name to make the second part, they began experimenting with connecting a second processor to the Micro.

They discovered something unexpected, the newer and supposedly superior processors like the Motorola 6800 and the National Semiconductors 32016 were super complicated devices with extensive instruction sets. Wilson and Furber even traveled to National Semiconductors’ lab in Israel and were shocked by the complex facilities and all the staff needed to make this beast work. Israel was basically what you’d expect from a big American company, with a huge structure and a huge team, but they had recurring problems debugging and yet, the practical improvements they offered in performance were minimal, especially in terms of making use of faster memory. The only test that really gave them positive results was another 6502 running on a faster clock they were overlooking something given these results, all hopes fell on a successor to the 6502; Wilson and Furber traveled to Arizona to visit the Western Design Center labs, run by one of the original design team members of the 6502. They were surprised to find something different there. Yes, that was very interesting. We went to Phoenix, Arizona, to the Western Design Center, expecting to find the typical shiny American offices with big windows. And what we discovered was that the Western Design Center was operating out of a bungalow in the suburbs of Arizona and they were using students to design basic gate cells and others on Apple computers. 

Their entire operation was less sophisticated than anything they had with them; maybe they had seriously overestimated the technical capabilities needed to create a simple microprocessor. So, what would happen if they could design their own microprocessor? That would finally satisfy them. When this idea was communicated to the rest of the team back home, they realized that they had recently read something that fit perfectly with this problem. What is this? I’ve been monitoring documents and press releases coming out of Berkeley about a new approach to processor design. The approach was, of course, the computer with a reduced instruction set, the idea that instead of making a processor more and more complex with many instructions, to do something similar with a smaller but much faster set of simple instructions, and then use those simple instructions to create more complex instructions. 

For a simpler and faster design and the simpler instructions meant a simple processor that could be designed by a small team. Oh my God, they could actually achieve this, they could say a chip just as they had done with the ABC micro chips and have PLSI manufacture them, still believing that they would encounter some inevitable obstacle. The team set out to design what they would call ARM Acorn Risk Machine. Yes, that’s absolutely what the acronym meant. Initially, Wilson designed the instruction set, wanting it to be powerful but simple, and then Furber implemented these instructions as circuits, carefully simulating the results with a reference model programmed in 850 lines of BBC BASIC. However, they encountered an additional requirement that presented an unexpected limitation. Look, if we want any chance of keeping the educational market happy, we need to keep costs as low as possible. What do you mean? What I mean is that the chip has to fit into a 50-cent plastic coating if it can’t have additional expensive cooling. Any other coating will cost 20 times more, do you understand? 

Furber was shocked by the draconian requirement that the chip had to operate with a plastic coating and produce less than a watt of power. Moreover, the tools and models they had to calculate how much heat each design element would generate were really archaic and inaccurate. So, he would have to estimate and leave a considerable safety margin. April 26, 1995, finally, the day everyone in the company had been waiting for arrived. The first samples of the first ARM chip had arrived from PLSI Technologies. After solving a couple of minor issues on the test board without plugging it in, the computer starts running. Invisible and collectively holds its breath as a test command is typed, the command executes, and out comes the result, Hello world, I am an ARM. 

A cry of victory resounds in the office. A bottle of champagne is opened, it works and it’s running extremely well, given the critical heat limits, Four estimates that by definition, it would also have to be very energy-efficient. Curious, he connects a meter and measures how much energy it is using zero watts that doesn’t make sense. Flustered, he discovers that the processor’s power supply wasn’t even connected, but how? The processor was unequivocally working, so where was it drawing energy from apparently, the chip was using zero power. And clearly, that can’t happen. A little investigation showed that I actually hadn’t connected the chip, but it was still happily executing the code, YY the only explanation for this is that the diakos chips have diodes on the pins. I as protection against electrostatics in case you walk on a nylon carpet carrying the chip it’s less likely to be damagedIt turns out the processor was so extraordinarily energy-efficient, operating at levels far beyond expectations, that it was functioning using only the trickle of current supplied through protection diodes meant for static discharge.

This remarkable efficiency set a new standard in the tech industry, allowing ARM to transcend its original purpose and lay the foundation for a range of applications in consumer electronics and beyond. This accidental discovery underscored ARM’s potential not just as a component in educational computers but as a pivotal technology in the broader computing landscape, potentially transforming ARM into a major player in the global market.