Last month marked the 25th anniversary of the MAX232, a single 5V-powered RS-232 line driver and receiver, quite possibly the “founding father” of analog system-level integration.
The MAX232, first sold in February 1988, skillfully married two seemingly unrelated functions, power and interface, and marked the beginning of what we integrate today at a much larger and more complex scale. When it was introduced, it was perhaps an even more innovative concept than the large-scale mixed-signal devices we see today.
The MAX232’s success was as much a tribute to the vision of its definer, Charlie Allen, as it was to the ingenuity of its designer, Dave Bingham. Before the MAX232, RS-232 interfaces consisted of two chips, a 1488 quad line driver, a 1499 quad line receiver, and a split bipolar ±12V voltage source. In older systems, the ±12V supplies were already needed for other parts of the system, so they were not much of an inconvenience. But by the 1980s, more and more hardware, including analog functions, could be powered from a single 5V rail, making ±12V less necessary.
Recognizing this trend was key to the birth of the MAX232. In fact, Maxim Integrated and competitors were hastening this shift by developing high-performance 5V-powered analog ICs, such as single-supply op amps, data converters, and analog switches. It would not be long before the RS-232 interface became the “high nail” — the last function to need a bipolar supply.
Like many innovations in history, when the MAX232 was announced, no one was asking for such a device. Combining a digital function with power was as far removed from designers’ consciousness as combining zombies with Pride and Prejudice . The MAX232 development required a system-level interest from an IC company that was very new at that time. Until then, analog IC companies’ technical interaction with customers consisted of helping select op amps, voltage regulators, ADCs, and the like, based on customer specs, and trouble-shooting customer problems. Not much time was spent on the “big picture” and system-level problems.
Maxim led the change to a system focus and away from treating analog functions like individual isolated devices in an effort to develop parts that were easier for customers to use. For highly integrated mixed-signal ICs, this once-novel approach is now mandatory.
At the IC design level, the concept of combining functions with their power supply was also nonexistent. In the 1980s, analog was a predominantly bipolar IC process world. CMOS existed for logic, but the best op amps and voltage regulators were built on bipolar processes. Not much analog performance was expected from CMOS in the 1980s. It took Intersil, Maxim’s progenitor in a sense (also founded by Maxim’s founder, Jack Gifford), to begin convincing people that CMOS could do heavy lifting in analog circles, but not yet with power. It took a second generation of CMOS design, at Maxim, to move that forward.
My hat’s off to the MAX232, one of the first of the “Integration Nation.”