When Did Analog Steal Digital’s Mojo?
Name two semiconductor companies whose names begin with the letter T that used to mainly make digital ICs but now think analog electronics is more exciting and remunerative!
It used to be that the ability to scale digital transistors provided a seemingly unlimited upside to working in the digital domain. And there’s no doubt that students thought the simpler maths and logic of digital looked preferable to the differential functions used to describe some analog properties. That was the 1970s and 1980s when the twin Ms of monolithic memories and microprocessors grabbed the headlines, ruled the world and converted almost all academic and commercial electronics to the digital way.
It also lead to the rise of digital signal processing, which was discussed in these pages recently (see The cult of DSPism). That’s the idea that everything analog should be converted to digital as soon as possible and processed in the digital domain for as long as possible before being converted back to analog to drive some sort of display, speaker or actuator.
Well in those days power consumption wasn’t usually the first thing you thought about. Now it often is. And that can mean it is no longer acceptable to be profligate with digital transistor twiddling – and a return to the use of more analog circuitry or a move to neuromorphic-style computation. And now with global competition the financial margins in digital electronics, especially where it applies to the consumer, can make it a less attractive sand pit to play in.
So some companies say analog, MEMS and sensors is the place to be. But there’s no time to be complacent. Competition is coming much as it did in the digital domain. We are probably where digital was in the 1990s and 2000s.