old-computers.com/website
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Full content from the completed page crawl (media download still running)

Browse Source 
6,654 pages: 1,289 machines, 702 software, 460 companies, 64 history, 17 articles, 59 fun.
4,336 media files so far; references to files not downloaded yet are left out.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01YDRjtmAiNpo5mK5N3d7QwZ
This commit is contained in:
yann64
2026-10-12 07:55:22 +02:00
co-authored by Claude Opus 5.5
parent 785c4967be
commit 3eeabe9649
10027 changed files with 321638 additions and 24 deletions
Download Patch File Download Diff File Expand all files Collapse all files
content/museum/ibm-an-fsq-7.md
+290
View File
@@ -0,0 +1,290 @@
---
title: IBM AN/FSQ-7
categories:
- Computers
types:
- Professional Computer
origins:
- U.S.A.
years:
- '1958'
cpus:
- '55'
params:
name: AN/FSQ-7
image: /media/museum/photos/ibm_anfsq7_part_s1.jpg
company: /companies/ibm
company_name: IBM
specs:
- label: NAME
value: AN/FSQ-7
- label: MANUFACTURER
value: IBM
- label: TYPE
value: Professional Computer
- label: ORIGIN
value: U.S.A.
- label: YEAR
value: '1958'
- label: END OF PRODUCTION
value: '1963'
- label: BUILT IN LANGUAGE
value: None
- label: KEYBOARD
value: IBM consoles
- label: CPU
value: 55,000 vacuum tubes in each unit
- label: SPEED
value: 75 KIPS (KiloInstructions Per Second)
- label: CO-PROCESSOR
value: None
- label: RAM
value: Core 8892-word
- label: ROM
value: Unknown
- label: TEXT MODES
value: None
- label: GRAPHIC MODES
value: 256 x 256
- label: COLORS
value: Monochrome
- label: SOUND
value: Speaker (see the ''Read more'' page)
- label: SIZE / WEIGHT
value: 2000 square meters / 275 tons
- label: I/O PORTS
value: Tape, punched cards, modem
- label: BUILT IN MEDIA
value: Tape
- label: OS
value: None
- label: POWER SUPPLY
value: 3 MW
- label: PRICE
value: $238 million, printer around $200,000 to $250,000
tabs:
- photos
- links
photos:
- title: SYSTEM DIAGRAM
subtitle: ''
text: |-
An AN/FSQ-7 installed in a 4-story SAGE Direction Center building.
Photo from **[Mitre](http://www.mitre.org/about/photo_archives/sage_photo.html)**
image: /media/museum/photos/ibm_anfsq7_diagram_s1.jpg
links:
- title: Mitre Corp. museum
url: http://www.mitre.org/about/photo_archives/sage_photo.html
text: SAGE photo archives
- title: Radomes, Inc. museum
url: http://www.radomes.org/museum/equip/fsq-7.html
text: More information and pictures about the AN/FSQ-7
- title: Reminiscences of Jack Gilmore
url: http://ed-thelen.org/comp-hist/TheCompMusRep/TCMR-V02.html#Reminiscences
text: ''
- title: SMECC Museum
url: http://www.smecc.org/sage_a_n_fsq-7.htm
text: More pictures of the SAGE system
comments:
- author: Richard Wayne (United States)
date: Sunday 12th June 2022
text: I joined IBM in Kingston in 1959 and after 20 weeks of training I was assigned to the Duluth ADS SAGE location as a Central Computer Field Engineer. I was there for five years. I never saw anyone racing the aisles with replacement pluggable units (PUs). Nor did I witness large numbers of tubes failing daily. Most of the tubes were in the pluggable units which were repaired in the PU lab. A great time.
- author: Jacob Coleman (Unites States)
date: Wednesday 15th December 2021
text: Unfortunately my own father''s experiences directly contradict Glen Crandall''s statement. He absolutely witnessed people running up and down the aisles of tubes, replacing them with shopping carts, while he was an operator on the AN/FSQ-7 in Michigan. My assumption is that the load they were under simply wasn''t enough to trigger that many failures, because they deliberately induced them twice a day and still managed to need people running up and down the aisles replacing them.
- author: Henry Spencer (United States)
date: Monday 23rd August 2021
text: I programmed the AN/FSQ-7 computer to play music in 4-part harmony. I was a student in the USAF Command $ Control Computer Programming course at Keesler AFB in Mississippi in 1965.. I was a 2nd lieutenant at the time, later becoming an instructor. I have a recording of four pieces of music I programmed that was played at the Montgomery Air Defense Sector site in Mobile, Alabama in 1965. I can send you a copy of the music if you email me requesting it (henryspencer@gmail.com). I had heard some previously written tunes that had been written with one and two parts. I speculated that if played at the proper speed, I could program an accumulator bit to repeat four notes of a chord to simulate four part harmony. I wrote a test program to do that. After I picked the proper speed, I then programmed these four pieces of music.
- author: fred panter (United States)
date: Saturday 20th February 2021
text: taught the anfsq7 (bryan hall) kafb early60''s
- author: Gary Nelson (United States)
date: Monday 4th February 2019
text: I was a first generation USAF "BlueSuit" graduate of the IBM Kingston training program in 1961$my job code was 30551. I was a computer repairman and later computer operator of the Q7 at Duluth Air Defense Sector (DUADS). One topic I have not seen mentioned was the fact that many of the voltages that ran the computer came from motor-generators that were controlled by the computer itself. The pentode tubes were designed so that, when run at lower or higher than normal voltage, they would fail. By this means, the failure rate of 48000 tube system could be acceptably low.
- author: Michael Yourshaw
date: Tuesday 21st August 2018
text: I trained as a communications computer systems analyst at Keesler AFB in 1965. We had a field trip to the Q-7 in Montgomery where we each wrote, in JOVIAL, and ran a small program on the backup system. The CRT/light pen programming was fun. For some reason we thought the "NIFA" acronym was hilarious ("not in file area", a flip-flop that was activated by a light reflecting on a piece of aluminum foil glued on the magnetic tape near the end of the reel).
- author: Keith Krier (Pittsburgh, PA)
date: Wednesday 18th July 2018
text: Went to Keesler for SAC HQ training, but got assigned to Duluth from 1972 to 1975 Enjoyed the computer maintenance career 30594. Spent 34 years and retired as a Chief.At Duluth the shop manager was a civilian. He used to come at night to drive us crazy installing faults and making us troubleshoot them. Anyone remember 4cv was the main clock.
- author: Herb Pfeifer
date: Tuesday 8th November 2016
text: Obviously it was Keesler AFB, Biloxi, Mississippi
- author: Herb Pfeifer
date: Tuesday 8th November 2016
text: |-
I attended Q-7 training at Biloxi AFB in 1967-1968 but was never assigned to a Q-7 site. Instead I was assigned to the NORAD cave and later to the BUIC site at Havre AFS.
An interesting observation about the Q-7 was its extremely high reliability due to marginal testing of the backup system when the tubes were subjected to over and under voltage stresses. Vacuum tubes that would have failed during normal operation were "Pre-failed" on the backup system rather than being allowed to fail during critical online operation.
As you might imaging, the building housing the Q-7 required only air conditioning, but not heat as this was provided by over 100,000 vacuum tubes.
- author: Jeff Davis (USA)
date: Thursday 18th August 2016
text: |-
I worked on the Q7 from 1974-76...I/O and Displays, at Duluth, MN.
Was a lot of fun to work on! Had big and little memory, drums, reel-to-reel tape, card reader, printer (used relays...lots of them). Would boot up using a plug board(if I recall the name). The display console had a 10,000 volt power supply that had to be checked during maintenance. You have to turn off the power, attach your probes and then turn power on. If you attached the probes before powering off, the jolt would knock you back on your butt...or worse.
Toughest problem I worked on was an intermittent failure that we finally found by turning off the lights where the frames with pus and looked for bad tubes. We found the culprit...one blue tube.
The other cool thing was that there was a speaker that would emit these squeak-like sounds while the main program (active mode) was running. It was the same sound pattern over and over so you could tell by listening if things were running okay. You also would know when things were going down. Then the commander would call down from the fourth floor over the loud speaker. Hee, hee.
Fun times.
- author: Glen Crandall (Atlanta, GA)
date: Saturday 28th November 2015
text: |-
"There were usually several hundred tube failures each day, replaced by workers racing up and down the tube racks with shopping carts full of replacements."
This statement is incorrect. Because of the automated testing conducted on a daily schedule actual failures were extremely rare.
The only time large numbers of tubes were replaced occurred happened after about four years of operation. At this time predicted failure rates of vacuum tubes rose to the point where all tubes were replaced. Usually about 100 to 200 tubes were replaced at a time during this phase.
I joined IBM in September 1957 and went to Kingston, NY for a 6 month training period. My permanent duty assignment was the DC at Gunter AFB, Montgomery, AL. I was there from April 1958 to November 1961.
In November 1961 I transferred to the software development site in Santa Monica, CA. I was there until about June 1966.
- author: Richard Haskell (USA)
date: Sunday 26th April 2015
text: 'Listed email: railfanewmexico@hotmail.com'
- author: Richard Haskell (USA)
date: Sunday 26th April 2015
text: I was USAF trained at Keesler AFB 6/65-6/66 on the Q-7. First permanent party station at McGuire AFB (NYADS) as a 30553-A then Topsham AFS (BOADS). While at McGuire I was awarded Airman of the year and had my photos taken on the job. These are some of the best and most detailed photos of the Q-7 . Contact me at the listed email for copies.
- author: Mike Perzel (USA)
date: Wednesday 17th December 2014
text: I joined IBM right out of the Navy in 1957 and trained on the SAGE computer in Kingston NY. Our team installed the system at McCord AFB in Tacoma, WA. I recall during installation, the air-conditioning system was being tested and I do believe room temperature was around 50 degrees, and this was in the summer. In fact, I received several cash awards for changes to the 026 card punch manual. installati
- author: Gary Cartwright
date: Thursday 28th August 2014
text: |-
I was hired by IBM in 1958 to help install and maintain the ANFSQ-7 at DC-13 at KI Sawyer AFB outside of Marquette Mich. By that time the SAGE (Semi-Automatic Ground Environment) systems were being installed in 3 story buildings. The computer was amazing compared to what we have today. What we call RAM today was a 64K core memory unit (that was the "Big Memory Retrofit - the original was 4K)
I was part of the maintenance team until the USAF Operation Bluesuit, which turned over maintenance to Air Force personnel. I stayed on as Contract Technical Services until the site was closed in 1963. It was the basis for my continued career in IBM in "commercial" field service, eventually programing support, engineering and finally product pricing before I retired. Since then I have built PC''s as a hobby. SAGE was the great beginning of that career.
- author: Dr. Dave Felsburg (USA)
date: Tuesday 27th May 2014
text: I served as maintenance technician, crew chief and site trainer at Fort Lee AFS, VA (Washington Air Defense Sector - WADS) from 1966 through 1969. The Q-7 provided one of the best basic computer operations and air surveillance mission training available anywhere. Its maintenance console hosting over 10,000 neon lights allowed technicians to single step through machine operations to watch data transfers and troubleshoot with ease. Its size was equivalent to 23 ranch-style home with an internal telephone system to match. It had 123 miles of internal signal wire and 75,000 vacuum tubes. It was so impressive that Hollywood bought several of the surplus machines as they were shut down or never installed. They were featured in "Lost in Space," "Journey to the Bottom of the Sea," "Time Tunnel" and many more. The expertise I gained at Fort Lee made it easy to become site computer maintenance trainer for AUTDIN at Clark AFB, RPI and complete an electronics engineering degree. As an officer,
I led a special team out of the 4754th Radar Evaluation Sq at Hill AFB, UT to respond to a disabling fire at the Fort Lee Q-7 by redirecting all air surveillance inputs and response capabilities to two neighboring Direction Centers. The severely limited data handling capacity was addressed by focusing computer processing on only those aircraft of interest instead of the entire civilian air traffic load. The basics of computer operations and maintenance learned at Fort Lee led me into over 20 years of active USAF time in computer systems design and engineering culminating with 4 years at the USAF Academy leading a mathematics department and teaching engineering application to thousands of Academy cadets. I have always been grateful to the knowledge and experience I gained through that old vacuum tube computer!
- author: Jim Bellah (Oregon/USA)
date: Sunday 2nd March 2014
text: I was computer maintenance at Adair AFS, Oregon from 1965 to 1968. Started out as display, cross trained to central computer later. In about 1967, some "genius" decided we should change all the Rayethon (5616 or 5818?) tubes which were the meat of the computer. We protested, but had to do it anyway. About 5$ of the new ones were bad, had a heck of time getting it back up. They closed Adair July 1969.
- author: Bill Ramsdale (Truax AFB Madison wi.)
date: Tuesday 18th February 2014
text: Just out of high school I was hired as a Multilith operator at truax AFB Air Force Base in Madison Wisconsin I was given a confidential security clearance. About 1950 I was promoted to process the programmers programs into computer via punchcards and return the results of their efforts to the programmers. I was employed by the systems development Corporation a division of RAND Corporation. I''ve also and my clearance changed to secret. IWorked at this position for about a year and 1/2until the Air Force took control of the facility.
- author: Mark Avant (Yucaipa, CA)
date: Thursday 7th November 2013
text: |-
I was a crew chief for a maintenance crew on the Q7 at Ft Lee Virgina from 1979 to 1982. we called the core stacks big memory and little memory. A couple of years before I arrived they had a fire up in the weapons area. Soaked the entire computer. We used to have issues with the wire relays all the time. take them apart, clean them with an eraser and viola, problem solved. Troubleshooting with an oscillascope and logic diagrams. Those were the days.
After I left the Air Force I worked as a field engineer for WANG Laboratories and then went on to Network administration after the colapse of the mini computer industry. I am retireing in 8 months from my current position as a Network Adminsitrator for a School District. No more banging my head against the virtual wall.
- author: Dave Bignoli (Phoenix, Az)
date: Tuesday 22nd October 2013
text: Went to Keesler AFB for BED (Basic Electronic Development) and was quartered in the Triangle area May 1962 to May 1963. Worked as a central computer tech at McGuire AFB New Jersey thru Feb 1966. There were not very many central computer techs around, we just had 4 at McQuire. Went to work for GE upon discharge as a field engineer thru 1970 then got out of the field and into real estate in Phoenix area. In 1973 decided computers and real estate were something that needed each other. Currently operate a system that evaluates real estate. www.propertyvaluecentral.com . The AF provided me with a great living my entire working life.
- author: William Lagasse (Richmond, VA)
date: Monday 13rd May 2013
text: My father, Conrad Lagasse was assigned at Tomshom AFS from 62-69 working on the AN FSQ 7 / SAGE, still knows a great deal about the system, anybody who was there then please let me know..
- author: Mac McLemore (Oklahoma, USA)
date: Monday 19th November 2012
text: |-
This is for Dean from Florida. My experiences parallel yours. I graduated Keesler and reported to the Q-7 shop at Topsham AFS, Maine in September 1965. That means we were in Bryan Hall at the same time. Stayed at Topsham until 1968. Went to Udorn where I worked on the Univac 1050-II Base\
Supply Computer. Came back to the States on SAC''s 465L system . Worked on it until 1982 then went to Tinker as an Airborne Computer Tech on the E-3 AWACS aircraft. Stayed with that until 1994, at which time I had 30 years service and retired.
- author: Dean (Florida) (Florida, USA)
date: Sunday 20th May 2012
text: I went to tech school at Keesler AFB in 1964/65. A 49 week basic electronics course and advanced training on the AN/FSQ-7. In Aug 1965 I arrived at McChord AFB as a computer maintenance tech. I worked on the computer for two years. At that time there was a Direction Center (DC) and a Command Center (CC) colocated in the same building. In 1966 one of the computers, I think the CC was shut down and sold for scrap. I helped remove the equipment racks and consoles. In later years on a TV show called "Time Tunnel" had one of the main computer control consoles to control the time tunnel. Another TV show called "20,000 Leagues Under the Sea" had one of the drum control panels as a control panel on the submarine. I left the vacuum tube computer in 1966 and was assigned to a transisized version called Back Up Intercept Control (BUIC). The new computer was a back up to the Q-7 but with all the same plus increased capabilities in a much smaller package. It was housed in a building of 4 or
5000 SF on a raised floor with cabling under the floor. After a few months of OJT on the new computer, I was transferred to Udorn RTAFB, Thailand in 1967 to work on the BUIC which was used to control and direct fighters and bombers flying from Thailand and South Vietnam in the Vietnam war. I''ll be happy to correspond with anyone about the AN/FSQ-7 and my experiences.
- author: Steve (Vermont)
date: Sunday 1st January 2012
text: I worked at the 24th NORAD site at Malmstrom AFB, Great Falls, Montana from March 1981 - June 1983, when we shut it down and locked the door. I was teh tech that worked on the TDDL (Time-Devision Data Link) system that the computer communicated through to the Guard jets. I worked in the computer repair lab as well.
- author: Tarbet
date: Tuesday 26th April 2011
text: |-
Questions on Q-7 Speed, Memory $ Storage ...\
Of course, the Q-7 was duplex - 2 independent machines interconnected.\
Speed - 2MHz clock, 6MS memory cycle time, 12MS add time (2 cycles)\
Memory - 64K and 4K of 32-Bit words (not bytes).\
Drums - 12 (originally 6) with 6 fields each. Fields contained 2048 locations.
Manuals are available on Bitsavers.org/IBM/SAGE/
The system was specialized enough that an emulator would do you no good. It was never used in a data processing environment.
- author: Tarbet (US)
date: Tuesday 26th April 2011
text: Pete''s comment on "first ground control of interceptor aircraft" relies on a mixed metaphor. "GCI" in the 30''s was manual (radio to pilot). Under SAGE, it was actually controlable from the ground.
- author: packetStar (Califoria, USA)
date: Saturday 9th April 2011
text: |-
I agree with Crystal Campbell, SAGE was running in the northwest US, the 25th ADC, until at least the end of ''83. I worked at the other end of SAGE, at a GATR site, ground to air transmitter receiver site, where we had the AN/GKA-5 Time division data link that the Q7 at McChord sent data to. The GKA-5 then transmitted data to the interceptors, coded messages that vectored the interceptors to their targets. The K-5 was still running when I left the AF at the end of ''83. FYI - I was stationed at the 777th radar squadron, Klamath, California.
peace
- author: royal prentice
date: Tuesday 11th May 2010
text: I was on a maintenance crew on a Q7 during 1971 72 and I never saw anyone running around with a shopping cart full of tubes. The machine used pluggable units that were removed during marginal checks and repaired on a test bench.
- author: Pete
date: Monday 3rd August 2009
text: The article claims "First ground control of interceptor aircraft". The Royal Air Force were doing ground-controlled interception in the late 1930s - it''s largely this that allowed the Battle of Britain to be won and hence German invasion prevented. SAGE was extremely significant in many ways, but not this particular way.
- author: Crystal Campbell (USA)
date: Friday 17th April 2009
text: |-
The dates concerning the shutdown of the AN/FSQ-7 are not quite correct. I joined the Air Force in 1981, and I worked on the McChord computer until we shut it down and dismantled it. I did not actually arrive at McChord until after tech school, so the AN/FSQ-7 was definitely operational until 1982. I do not remember exactly when the mission switched over to computer replacing the Q7, but it would have been in 1983. I had heard that there was one still active in 1985, but I can''t confirm that.
It was a fascinating computer to work on.
- author: Programer Learning Ti-Basic (Aplace, CountryWorld)
date: Sunday 15th February 2009
text: I really wish there was an emulator for the IBM 305 RAMAC. That would be a step in preserving the history of one of man kinds most astounding achievements, the computer. I myself love the RAMAC''s unique grey case. It RULES!!!!!!!! I want to own one some day, along with an Apple PiP!N game system.
- author: B (USA)
date: Friday 24th August 2007
text: To give some perspective, a $238 million computer in 1958 would be the equivalent of a $1.6 BILLION computer today.
- author: Alexei (Seattle (USA))
date: Wednesday 3rd May 2006
text: |-
So, can we get an emulator for this? I don't see it in MESS.
As for long-term storage, probably punch-cards or paper tape. Maybe magnetic drum?
- author: Tommy Mason (playing halo on my pimped-out 386)
date: Thursday 17th November 2005
text: |-
cathode-tube's where pertty much fazed out around 1983.
although there is some talk about one still being fired up in New york for shift keeping.\
but I don't think that's true, still- that would be neat to see it!
- author: Computer Guy (Garage^^)
date: Tuesday 15th November 2005
text: |-
How fast was it anyway?\
And how long did computers use cathode-tubes for?!
- author: ClAwHaMmEr (UK)
date: Saturday 16th July 2005
text: |-
err, i would estimate around a few kilobytes, as this big boy was created merely two years after the first HD was built by IBM, and as such uses FSQ-7 magnetic core memory circuits instead.
even if IBM decided to include hard drives in the design, i seriously doubt the lowly cathode-tube CPU would be fast enough to access data at an acceptable rate. Apart from this, IBM's first HD, the IBM 305 RAMAC required 50 24-inch disks to store five megabytes of data, which would be prohibitive considering how cramped the storage building was already
hope this answers your question, linkman ;)
- author: Linkman (Slovak Republic)
date: Friday 15th October 2004
text: Wow - that is hard machine :))) tell me how big disc space it has ?
oc_id: 1050
source:
url: https://web.archive.org/web/20230603204628/http://www.old-computers.com/museum/computer.asp?c=1050
date: '2023-06-03'
---
The AN/FSQ-7 was by far the largest computer ever built, and is expected to hold that record. It consisted of two complete [Whirlwind II](/museum/mit-whirlwind/) computers installed in a 4-story building (See the impressive diagram in the 'More Pictures' section).
Each AN/FSQ supported more than 100 users. IBM had about 60 employees at each site for round-the-clock maintenance.\
Keeping one unit operating and one on hot standby (to allow for switchover when vacuum tubes failed) resulted in better than 99% uptime. The roles of the two units were reversed at regular intervals, allowing diagnostics and maintenance to be carried out on the standby unit.
There were usually several hundred tube failures each day, replaced by workers racing up and down the tube racks with shopping carts full of replacements. Automated tests run by the computer itself would cycle the voltage to the tube racks down and back up to induce marginal tubes to fail early, so that the computer would normally run correctly for the rest of the day. Without this process, the MTBF would have been a few minutes.
By the time SAGE was deployed (22 or 23 stations in the period 1959-1963; sources disagree) it was nearly obsolete, since it was designed to detect bombers, not the new ICBMs. Nevertheless it was operational until 1979, when the ROCC (Regional Operations Control Centers) system took over, using much higher-speed computers. One SAGE station continued operating until 1983. This last unit was donated to the Boston Computing Museum, since relocated to Moffett Field in Mountain View, California. The museum also has a tube panel from the Whirlwind I. Whirlwind II consoles turned up in the TV series *Battlestar Galactica*.\
In spite of its limited military value, the SAGE system served as an excellent prototype for an air-traffic control system. The FAA operated its own AN/FSQ-7 systems for many years after SAGE was shut down. IBM's experience with these systems had a great deal to do with its later success in computer systems, and its dominance of the market for large computers. The IBM 7090 was essentially a solid-state version of the AN/FSQ-7/8. (The 7090 has its own rich history, including hosting the first-ever multiuser APL system.)
• First CRT-based real-time user interface,\
• First use of light gun to pick an item on the screen,\
• First wide-area modem communications (1300 bps),\
• First hot standby system for maximum uptime,\
• First ground control of interceptor aircraft,\
• The first in line microfilm fast processed 35 mm projection displays, preceding printer/plotters. A screen capture could be displayed within 30 seconds, • First two-pass assembler, permitting symbolic addresses.
Thanks to **Edward Cherlin**, [Simputer Evangelist](http://www.ryze.com/go/Cherlin), for all this information.
The photo (from [Mitre](http://www.mitre.org/about/photo_archives/sage_photo.html)) shows the rear panel wiring of FSQ-7 arithmetic element frame.
We need more info about this computer ! If you designed, used, or have more info about this system, please send us pictures or anything you might find useful.