Inside the competition for the first PLC
The race to develop the first programmable logic controllers was underway inside General Motors’ Hydra-Matic Transmission Division in Ypsilanti, Michigan, in 1970. Three finalists had very different architectures.
In 1970, a fierce contest was being waged inside General Motors' Hydra-Matic Transmission Division in Ypsilanti, Michigan. At stake was program ownership of what would become the first programmable logic controllers (PLCs), and the outcome would shape not only the design of the first PLCs, but also the success or failure of their advocates.
It was natural that GM's premier manufacturing technology group, Hydra-Matic, would be the incubator of the PLC. A nationally recognized technology leader, Hydra-Matic had installed one of the early IBM 1800 computers, the IBM 1801.
Richard Lundy, an owner of one of the original PLC suppliers, tells the story of when Ed O'Connell, a Computer Group engineer, didn't think GM was getting enough support from IBM. "An entire test line was running off of the one computer," Lundy said, "which made it very vulnerable, and Ed didn't like the support he was getting. So he called Watson." That's Thomas Watson, Jr., the son of IBM's founder and company CEO. Support for Hydra-Matic's IBM system quickly improved.
Fuel: Hydra-Matic internal competition
In April of 1968, a young Hydra-Matic engineer, Dave Emmett, proposed the development of what he called a "Standard Machine Controller." The controller he envisioned would replace the relay systems that controlled machine operations. Emmett was in charge of the Circuitry Group, and he envisioned a machine that would reduce maintenance costs, improve machine diagnostics, and decrease panel space.
Randy Brodzik, who later worked for Emmett, remembers that Emmett had a clear vision of what was needed: "Dave said the goal was to develop a technology that would significantly reduce the time it took to make changes to a machine control sequence. With relays, first you had to do all the documentation, and then change all the wiring. It would take hours."
At about the same time, another group in Hydra-Matic was envisioning a different type of machine control system for GM. The Computer Division had hired Information Instruments, Inc. (3-I) to create a computer controlled assembly machine for the forward clutch line. This new control system had no limits to the number of elements in a ladder diagram, included parallel processing, and could incorporate complex Boolean equations.
Richard Lundy, a program manager for 3-I at the time, said the competition between the two Hydra-Matic Groups was "pretty dramatic." He described the Circuitry Group as wanting to duplicate existing ladder diagrams, while the Computer Group wanted to use nonsequential programming, similar to that used in current end-of-line test systems, to provide a more robust instruction set and reduce processing time. The advantages and limitations of each approach would become clear as the project progressed.
3 finalists, different architectures
In April and May of 1968, work was done on a request for proposal that is remarkably simple by today's standards. The RFP, issued in June 1968, included only four pages of design specifications, including a requirement that "memory word length shall be at least eight bits." Douglas Brant worked on specifications for equipment coming into Hydra-Matic. He said that there was a strong rivalry between the Computer Group and the Circuitry Group in developing the specifications. "I guess it boiled down to people's territory," Brant said.
Of the many companies who received the RFP, three were selected for evaluation: Digital Equipment Corporation (DEC), 3-I, and Bedford Associates, a consulting firm. Its products were the DEC PDP-14, the 3-I PDQ-II, and Bedford Associates' Modicon 084. (The Modicon name came from MOdular DIgital CONtroller.)
In June of 1969, DEC's PDP-14 was installed to control a gear grinder. Richard Lundy says that in spite of a very strong GM-DEC relationship, DEC's concerns about memory failure significantly limited the PDP-14's competitiveness. To make a change to the DEC program, "you had to send the client program in to them, and they would send you back a hardwired memory board," Lundy said. Having to remove a controller's memory from the plant to make changes would prove to make the DEC option unsustainable.
2 Hydra-Matic Groups, alliances
Richard Lundy's company, 3-I, was aligned with the Hydra-Matic Computer Group. Lundy said that during the evaluation period, the rivalry between the Computer Group and the Circuitry Group escalated. To the Computer Group, the advantages of higher level logic capabilities were clear and the selection of the PDQ-II was an obvious one. The 3-I solution for the forward clutch line, with its complex programming, had performed beautifully. Lundy said the PDQ-II quickly became known within the Computer Group as "pretty damn quick."
Emmett and the Circuitry Group advocated strongly for the Modicon 084. The programming that was seen as a benefit by the Computer Group was seen as a detriment by Emmett's team. Because Modicon's programming language was similar to the Group's familiar relay ladder diagrams, it was expected to provide the smoothest transition and lowest cost for training and support. Randy Brodzik recalls that Emmett had an implementation view of the project: "He said that in order for this new technology to achieve wide acceptance, it needed to emulate what was already in place."
As an added bonus for Modicon, the 084 was the only controller built into a hardened package, providing plant floor protection that the other two options did not.