Random Access Memory (RAM) (1968)

Dennard combines a transistor with a capacitor in a revolutionary memory cell.

   
RAM, or Random Access Memory, the short term, high speed "working" memory of a computer, has existed sine the invention of magnetic core memory in 1949. Modern RAM, though, owes its invention to Texan Robert Dennard (b. 1932).

     In 1966 Dennard was working at IBM's Thomas J. Watson Research Center. IBM knew that magnetic core memory was too bulky, power-hungry, and slow, and that transistors would be the answer to replacing it. They had reduced the problem of storing a single bit of memory to a cell that used only six transistors. Added to a silicon chip, this cell was already tiny compared to a magnetic core. Dennard, though, simplified the memory cell even more, to a single transistor and capacitor, a component that an hold an electric charge. The memory was stored as charge on the capacitor, and the transistor was used to read and write it. Capacitors "leak" charge, though, so the memory had to be continuously refreshed, many times a second. Because of this constant forgetting and refreshing, Dennard's system is called "dynamic" RAM, or DRAM.

    Despite its need to be refreshed, DRAM had a world-beating advantage. With only two components, which could be placed side by side in the thousands on a single silicon chip, it was the smallest memory ever made. The computer industry quickly took advantage of Dennard's invention, and fledgling company Intel released the first commercial DRAM chip in 1970. Magnetic core memory became the technology of yesteryear almost as soon as Intel began to ship its new chip.