Talk:Scientific Data Systems
From Wikipedia, the free encyclopedia
Greg
I would love to know who your father is. I worked on the entire Sigma family while employed as a field engineer, diagnostician and product engineer at Xerox from 1969 until 1978. During that time, I was fortunate to have mastered their design, became a specialist in their CPU, Memory and I/O components, and worked many challenging assignments, especially with the Sigma 9, which was my favorite. I even hold the distinction of having actually manufactured two of them. Long story, more on this below.
The Sigma family included ths Sigma 2 and Sigma 3 (16bit) and the Sigma 5,6,7,8 and 9 (32bit). Before Xerox left the main frame computer market, they also introduced a follow on family to Sigma, the 530 (16-bit) and 550/560/570 (32-bit) systems. Sigma systems were first introduced in the 1960's and were extermely advanced at the time. Designed specifically for real-time computing tasks, their execution was faster and more efficient than any competitors systems. Using multiple banks of registers and real-time interrupt levels, context switching for priority task execution could be performed with nanosecond precission. The Sigma 2/3 systems were typically applied to dedicated real-time tasks such as control and data acquisition. The Sigma 5 and 7 systems were general purpose 32Bit computing systems capable of simultaneous real-time control. Sigma 6 was a specialized version of the Sigma 7 for data processing. The grand systems of all were the Sigma 8 and Sigma 9, which supported parallel instruction pipelines with look-ahead analysis and pre-fetch, expanded system memory architecture, and exceptional maintainability. These were exceptional machines and were applied to high volume real-time data acquisition and control as well as multi-user time-sharing applications. The Sigma 8/9 ran the CP-V operating system, the only general purpose system executive ever written that fully supported high speed general computing and I/O tasks simultaneously with pricision real-time computing.
Sigma systems were widely used in research and development, education, process control and government applications, and I was exposed to a good many of them as a Technical Support engineer for the Western Region, based at the home plant in El Segundo California. Because of our direct association with the factory and design engineering staff, our office was often called to help resolve difficult system anomolies all over the world.
I was fortunate to be exposed to some of the most challenging applications of the Sigma 9 series, and became known as somewhat of a speciallist on them. As a result, I was assigned to support one of the most unique applications of this series at Nellis Air Force Base in 1975-1977. Sigma 9 systems were used to provide data acquisition and corelation for real-time air space management of the Nellis range complex. In addition, two speciallized systems, each consisting of 3 Sigma 9 CPU's tightly coupled as a single system, were applied in the monitoring and control of multiple aircraft engaged in air combat tactics test and evaluation. Commonly known as the Air Combat and Maneuvering Instrumentation (ACMI) system, these systems were capable of tracking and displaying 3-dimension views of up to 16 aircraft within a large block of air space in real-time, and also provided simulation of weapons systems to provide a true realistic combat environment. A unique design of the system was dedication of one of the CPU's to filter the real-time data and generate filtered projection of future events based on current and historical data for each aircraft. This allowed the system to provide accurate prediction of saftey issues, as well as seamless display even when data from one or more aircraft was disrupted due to dynamic manuevering. If you have seen the movie "Top Gun", there are scenes that show real-time air combat graphic displays produced by a similar system employed at Miramar Naval Air Station. That system, and several others like it deployed world wide, used 3 Sigma 5 CPU's.
After Xerox left the main frame computing business, support for all Sigma systems transitioned to Honeywell, where I continued to work as a systems engineer until 1978, when I joined Comshare Corporation, the largest single user of Sigma 9 computers. A combination of continued business growth, and the unavailability of more Sigma 9's in the near term threatend Comshare's growth, so while negotiating a contract with Honeywell to build more new Sigma 9's, we initiated an effort to do the same using available spares and scrap obtained from Xerox. Based in a rent-a-garage facility in Phoenix, I led a team that hand built from scratch two Sigma 9's, integrated them with a custom I/O processor that allowed use of IBM compatible Tape and Disk peripherals, as well as 3rd party Memory provided by Ampex. The result was two fully functional systems deliverd to Comshare in 1979 and 1980. Since we could not legally name these two system "Sigma 9", they were delivered with the monkiers "Kermit" and "Miss Piggy". They served Comshare customers for many years afterward.
Suffice to say, I had fun, and even though it was a long time ago, I can still tell what pins to hang scope probes on to get a synchronized view of an I/O transaction.
Today, I have a system in a small box next to my desk that is capable of processing instructions faster than a Sigma 9, and has a whole lot more memory and disk, but it will never have the class.
Regards Dave Phillips
Dlphillips1 (talk) 06:54, 14 April 2008 (UTC)
Bold textMy father was a field engineer and eventually field service manager for Scientific Data Systems (SDS) and the later Xerox company and then Honeywell. I believe this article to be fairly accurate.
Greg
I was involved with the second startup in 1979 and have added to some of the history on that company.
William L. Scheding, wls@wls.org
There was a Sigma series computer at Pacific Missile Test Center, Point Mugu, California. It was in the integration lab of the F-14 Tomcat fighter, I believe to control the simulation environment. This would have been from around 1969 to whenever the last software changes were made. —Preceding unsigned comment added by 205.175.225.22 (talk) 21:01, 5 October 2007 (UTC)
I worked with the 32-bit Sigma machines between 1972 and 1985 in the UK and Sweden. They were almost exclusively used for real-time processing, although a timesharing service bureau (Atkins) also used them. Their software was very advanced for the time. Chrisj1948 18:08, 26 October 2007 (UTC)
On the subject of known users; there are indications that a certain UK government establishment used a Sigma in the 1970's although, because of the nature of their work, this was never openly declared :-) Chrisj1948 (talk) 10:20, 30 November 2007 (UTC)
