Honeywell and the 512 bit MOS memory device

Honeywell Former Headquarters

While at Honeywell, Bill researched semiconductor memories in products. His job was to research and then apply the knowledge he gained to create a low-cost solution. Along with John Defalco, he conducted simulation tests to see how the MOS worked. So, with tests conducted and the ability to navigate the MOS, Regitz began building a circuit simulation program to analyze the MOS in more depth. 

Bill's research on MOS design and MOS integrated circuits led him to request funding from Honeywell to develop the 512-bit memory-MOS memory device. While working on this, he incorporated a three transistor in the MOS. For his work on the MOS Memory, Regitz won an H.W. Sweatt Award in 1969. He presented his work on the MOS memory at the IEEE Solid-State Circuits Conference in February 1970. 

Willliam Regitz proposed a 512 bit MOS memory device. In October 1968 Regitz published the "Three Transistor Cell MOS Memory Array" report, which became a cornerstone in the semiconductor industry. The report was conducted on his findings for the transistor-cell 1024-500 nS MOS RAM. This was the first semiconductor memory that could compete with the low costs of cores.

Intel HQ (2017)

1968: “Intel was the driving force to modify the proposal to a 1k-bit chip instead of a 512. They said the density was not aggressive enough” for the 1102. 

Regitz interviewed with Intel and was given a job offer. He was tempted to stay at Honeywell but ultimately accepted the position at Intel to work more intimately within the semiconductor industry and chip design process. 

In 1970, Regitz gave a joint paper with Joel Carp at the Solid-State-Circuit Conference titled "A Three-Transistor-Cell 1024-Bit 500 nS MOS RAM." This was a goal of his for a while. 

At the Solid State Conference, his boss, Bill Jordan, met Regitz to convince him to stay at Honeywell. Not only did Jordan not convince Regitz to stay, but Regitz convinced Jordan to join him at Intel. After the conference, they “agreed to put a business plan together to convince Intel to go into the memory systems business.” 

Regitz sent his job acceptance to Intel along with his and Bill Jordan’s proposal for Intel to join the memory systems business. 

In 1970, Intel wanted Regitz to lead the DRAM dev efforts and “do whatever was necessary to get the 1103 into high volume production”, and with the help of Bill Jordan, that is exactly what he did. He not only accomplished this but was a key contributor in the high volume production. 

Early 1971, Intel dropped the 1102 and went into production with 1103. Though Regitz was disapointed, he understood the larger goal was to “set an industry standard, not develop a custom part for the sole use of Honeywell. We truly understood the power of buying standard parts and the economic impact on pricing.” 

In 1971, Regitz helped develop the 4 Kb 2107 DRAM which has “many improvements including on-chip intermediate-voltage generation for simpler system interface”. 

Regitz had the choice to either become a project manager “of a 4k part for use in Intel's Memory System’s IBM add-on memory products or move over to manufacturing set up and run the newly formed Product Engineering group”. He chose to become a project manager.

Regitz was split into the DRAM division with George Schneer and Larry Regis. He ended up becoming the manufacturing manager for DRAM. 

Using his system training from Bell Labs and Honeywell, Regitz’s “first accomplishments at Intel was characterizing the 1103 and solving many customer problems while enabling the 1103 to achieve high volume production.”

In February 1972, Bill Regitz and Joel Carper presented Intel’s 2107 4Kb DRAM at the ISSCC. 

The low-cost ROM invented by Regitz was officially patented in November 1973 under the number 3,774,171. 

Bubble Memory

Regitz, in 1977, worked with the Bubble Memories Division. His assignment was to “define a series of system products to be designed, built, and marketed by [the] Memory System Divisions”.

Regitz redesigned the 1103 and made the 1103A, the “only chip that [he’s] really designed personally and put into production” in 1978.

By 1998, Regitz was co-manager with Jeff Citta of Mobuild Module Manufacturing. They “ramped the Pentium II into production" and "went from almost 0 productions at the beginning of the year to about two million modules a quarter in Q4”.