XDS Sigma 5

To take full advantage of Flickr, you should use a JavaScript-enabled browser and
install the latest version of the Macromedia Flash Player.

Comments

SkiBum says:

That's Xerox Data Systems, not Scientific.

My Jr college in WI had a 300 baud link via ASR-33 tellitype to a Sigma-5 in Green Bay for Fortran and APL.

300-baud was a big step up from the previous year, when we'd punch cards and give them to a corrier who would drive them up. The data center would run our jobs, print the results and return them to us in 2 days. Imagine getting anything worthwhile done with 2 day turnaround for the edit/test cycle. 1974/5
Posted 47 months ago. ( permalink )

Laughing Squid says:

I did a little research and it turns out that Xerox bought Scientific Data Systems in 1969, which would explain why there are two different company names on this machine.

en.wikipedia.org/wiki/Scientific_Data_Systems
Posted 47 months ago. ( permalink )

bhurley317 says:

SDS (XDS) are the same thing ... we were using them for store and forward message switching and using the backup system for time sharing in the late 60's early 70's ...
Posted 38 months ago. ( permalink )

samyers124 says:

Sigmas also used old Wang 7 and 9 track tapedrives for backups
Not a chip one in these machines...
Posted 28 months ago. ( permalink )

NanooGeek says:

Heck, Rockwell, in Downey CA, used a couple of Sigma computers for engineering flight simulation of SpaceShuttle back in the 1970s.
Posted 24 months ago. ( permalink )

pshisbey says:

That's a beautiful machine.
Posted 23 months ago. ( permalink )

pcr1424 says:

The Sigma was indeed a beautiful machine in more ways than one. I was a Sigma 5/7 field engineer in the '70s/'80s and knew these machines well. SDS designed and built the Sigma 1 through Sigma 9 series in the 1960s which were based on diode / buffer amp logic (questionably third generation) and core memory. There was also a 500 series (530, 550 etc) having the same architecture but using large boards with integrated circuits and semiconductor memory in an attempt to reduce the size of the same machines but these machines were less maintainable than the Sigmas. The even numbers were development models and odd numbers were the production models. Sigma 1-3 were 16 bit machines, 4/5 was 32 bit with hardware floating point added, 6/7 added hardware BCD math and a new memory system for business purposes and 8/9 added instruction lookahead. XDS (El Segundo, CA) bought out the Sigma and continued production until Xerox closed their computer division and sold the Sigma to Honeywell in 1976 or 1977 (Phoenix, AZ, Camelback Rd at 35th Ave which is where I got my training).

The Sigma was half the speed of my first computer (a Mac II) but had an elegance of design that in my opinion has never been matched. It was used for timesharing and business purposes but its best application was for realtime tasks and was preferred by power companies and for railroad control and data acquisition applications. It was also the preferred computer used in the '70s/'80s by most NORAD spacetrack radars worldwide which is what my site was. The reason for its suitability for realtime uses was its incredibly simple yet efficient hardware interrupt system. This interrupt system could perform a complete machine context switch in only 50 microseconds by executing a single instruction. We had over a dozen realtime interrupts going on simultaneously with some happening as rapidly as 1000 to 2000 times a second but the Sigma didn't even flinch and typically ran at no more than about 70% of its capacity. When we replaced the Sigma in 1987 with a Gould machine that was twice as fast it took an entire millisecond for it to respond to such interrupts so it was necessary to reduce this dozen or so interrupts to only one and to add a separate hardware box to deal with all the realtime tasks that Sigma did by itself. Progress? I don't think so. These days computers can accomplish that same thing much faster but they must all use brute force to do so and none have the same wonderful elegance that the Sigma did. Today's computer designers could learn some very valuable lessons by studying the old Sigma design. Oh, and did I mention that the resident portion of the OS only occupied a total of 6044 words of memory (about 24 KBytes)?
Posted 10 months ago. ( permalink )

tony.burton43 says:

Not quite!

"Sigma 1-3 were 16 bit machines, 4/5 was 32 bit with hardware floating point added, 6/7 added hardware BCD math and a new memory system for business purposes and 8/9 added instruction lookahead"

I worked for tech support on the 32 bit Sigma series and was also familiar with the 16 bit series. To my knowledge there was no Sigma 1, although there was a 16 bit Sigma 2 (and a Sigma 3).

With the 32 bit systems, SDS developed the Sigma 7 first, it was built with floating point as a standard and had the earlier style of core memory and front panel.

Not sure of the subsequent chronology of development but the Sigma 6 was very similar to the 7, but without floating point. It had a BCD (Binary Coded Decimal) unit instead. SDS/XDS's answer to those customers that wanted to do some commercial computing.

The Sigma 5 was an entirely new breed, new, (electrically simplified) control panel and new memory system. It was also a cost reduction compared with the Sigma 7.

The Sigma 8 and 9 were the last of the Sigma range, I don't remember working on a Sigma 8 but spent a long session fixing a broken Sigma 9 in Milan. If memory serves me correctly, the 9 was available with both floating point and with BCD.

As far as interrupt response was concerned, yes, those systems coulld do a context swap very quickly, having 4 to 16 sets of general registers was part of the reason for the speedy response.

The European Tornado swing-wing aircraft was developed on 3 Sigma 5 systems, one each in England, Germany and Italy. BAC in England added a second system part way through the program.

I did my initial training on Sigma 5, 6 and 7 at XDS headquarters on Aviation Blvd in El Segundo in 1971 and '72. My second training session was on the additional telemetry equipment used for flight test systems in Europe and Israel.

The "Wang" tape drives were designed by an engineer called Wang at SDS, not to be confused with Wang Computers. There were some specially modified 7 tracks in the field, modified to read each individual track serially. These specials were used to read the 7 track field tapes created during the search for oil. SONATRACH in Algeria had such a device, fortunately it didn't break on any of my visits to that lovely country.

"That's all folks!!"\
Posted 7 months ago. ( permalink )

Would you like to comment?

[?]

	Laughing Squid's photostream
		27,642 uploads

This photo also belongs to:

	Computer History Museum (Set)
		83 items

3 people call this photo a favorite

Additional Information

Some rights reserved Anyone can see this photo

Taken with a Canon EOS 5D.
More properties
Taken on February 21, 2006
Viewed 2,142 times

Add to your map

Yay location removed

...

(Some formats are OK.)

Or, remove it from the map.

Go to the map

You	Sign in \| Create Your Free Account
Explore	Places \| Last 7 Days \| This Month \| Popular Tags \| The Commons \| Creative Commons \| Search
Help	Community Guidelines \| The Help Forum \| FAQ \| Sitemap \| Get Help