In 1970, Intel released the first dynamic random access memory chip, the 1103. By the end of 1971, Intel had delivered the chip to 14 of the 18 leading computer manufacturers. It became the world's best selling memory chip and it took over the magnetic core memory industry.
But, before all this success became a reality, Intel faced many problems. Before the 1103 was created, the i1102 came into the computer industry. The product was unsuccessful because it needed substrate bias, yields were low, and it caused the devices to have very small operating margins. The 11003 was created as a solution to these problems; however, this chip had its problems too. When the package for the memory chip was heated, it produced moisture which in turn seeped into the inside of the package ruining the memory chip. Many people tried and failed at finding a soultion to this problem, one being Craig Barrett. Barrett, however, was asked to come back to the company after being fired and led the company to success. The 1103 and its manufacterers had a rocky path to success.

A "side problem" in the manufacturing of the 1103

During production of the 1103, Intel ran into a problem. The ceramic package necessary for the memory chip had a small, but costly glitch. When the package was heated, it produced moisture which in turn seeped into the inside of the package ruining the memory chip.This was a big problem because the 1103 had accounted for 90% of Intel's $23 million in profits the year before, and Intel had already made thousands of packets. Somebody had to find a solution.
Craig Barrett decided to step in. His idea was to create a hole in the packets with a laser and then bake the water out. His method failed miserably as it ended up destroying most of the memory chips. After this devastation, Barrett chose to leave the company.
But, Intel was so impressed with his intuitiveness that they asked him to come back. When Craig began work again, the Japanese were seizing control of the memory chip industry and he decided something must be done, as the original problem still remained with the the packets as well. Barrett's solution was to copy the exact methods of the Japanese because they were so efficient and were obviously working well. The company rose to the top again and the problem with the 1103 packets was solved with the new methods of manufacturing.

http://money.cnn.com/magazines/fortune/fortune_archive/2000/12/18/293148/index.htm

1103 Sales

The 1103 was the first memory chip that allowed for the huge growth of PC's. By the end of 1971, Intel had delivered the chip to 14 of the 18 leading computer manufacturers. Because the cost to make the 1103 was much lower than the costs of a core memory or a 1101, the 1103 could be put into market for consumer sales more rapidly then the other memory chips; it became the world's best selling memory chip and it over took the magnetic core memory.

http://content.cdlib.org/ark:/13030/kt0c6016vq/?layout=metadata&brand=calisphere

1103 - The World's First Available DRAM Chip

In 1970, Intel released the 1103, which was the first Dynamic Random Access Memory chip. The memory device became extremely popular, especially over the core type memory.The HP 9800 was the first computer to use the 1103.
Dr. Robert H. Dennard and his coworkers at Intel were working on early field-effect transistors and integrated circuits, and he got the idea for the creation of the DRAM. He then began to think of ideas to make a chip that had a single transistor and a small capacitor.
There are two types of memory, dynamic random access memory and static random access memory. Because dynamic random acess memory has to be refreshed thousands of times per second, SRAM is faster.
In 1969, William Regitz of Honeywell was looking for a company to help him develop his dynamic memory circuit based on a novel 3-transistor cell. Intel saw the many opprutunities that taking his offer held, so they decided to work with him on the idea. The company first created the i1102, but that had many problems such as it needed needed substrate bias, yields were low, and it caused the devices to have very small operating margins.
To solve these problems they created the i1102 alternative, which used different cell techinques.The i1103 basically did not function, until it was discovered that there was an overlap between the "PRECH" clock and the "CENABLE" clock. Nevertheless, the product was put on the market. It was John Reed's job to improve the product prgressively by doing things like setting the refresh time at two milliseconds. This chip became the best selling semiconductor memory chip in the world by 1972.

http://inventors.about.com/library/weekly/aa100898.htm