Go to Antique Computer home page
back to Stories
back to Stories by John Van Gardner

7090 Adventures, toward restoration of a 7090.

by John Van Gardner, Nov 15, 2010

Robert,

I worked extensively on the 7090, 7094 and 7094 II at Lockheed Aircraft in Marietta, GA and took assist trips all over the southeast and as far north as Chrysler in Detroit, MI.
Here are some thoughts I would like to pass on to you.

When Lockheed ordered their 7090 my branch office arranged for me to go to a CE school in Poughkeepsie that would end as our machine was in final test at the plant. After school I worked on our machine until it was ready for shipment then went home to install it. This was a great learning experience for me.

One of the things I learned here was that there were a lot of delay cards in the system that had to be adjusted to the individual system. All of the line technicians I worked with had already memorized the locations of these cards and how to plug them for the delay they wanted. There was no such thing as a square pulse in most of these machines until the 7094 II came out with different technology cards. I attended a meeting with the quality control people that said 40% of the cards replaced in final check tested good on the card testers that manufacturing was using. These were very sophisticated card testers and the problem was thought to be delay buildup in the circuits even though each card met its specs.

When I told the quality control man about having to hand pick cards to get a machine working the manufacturing people denied it so we set up a challenge. We took a 7302 Oil cooled memory frame that was checked out and ready to be put on a system. The 72 sense amplifier cards were removed and placed on a table. They were scrambled around and the machine repopulated. To get the machine to run all the test again 6 cards had to be replaced. These cards were position sensitive.

Current mode SMS cards are going to be hard to come by. There were only about 250 7090s built. Lockheed's 7090 was converted to a 7094 and finally they gave it to Georgia State University in Atlanta. I helped maintain that system for over 14 years. In the end all the parts we could get were used parts such as the heat exchanger unit for the 7302 memory and cog belt driven squirrel cage blowers that were always going bad.

The power system is like nothing else. The 7608 Power Converter frame has a General Electric 400 Hz 3 phase generator driven by a 50 Hp electric motor. There is another frame called the 7618 Power Control Unit. It has an exciter panel that regulates the field current in the generator and supplies a field flashing voltage. I went to Eglin AFB in Florida once when the field flashing supply had died and no part was available locally. We went to Sears and bought the biggest 12 volt car battery they had and hooked it up to the 7818 and ran that way for 2 days. We had a lot of fun showing everyone there the battery and telling them the system was running off of it. The regulator panel had a VR tube to set the reference voltage. This was the only tube in the whole system as zener diodes had not been invented at that time.

The 7090 was developed from parts of the 7030 Stretch program for the BMEWS (Ballistic Missile Early Warning System) on a rush basis. It was so fast that when the first two systems were sent to Alaska the CE schools had not been developed. Engineering had to send people to maintain the systems for the first 6 months. This turned out to be a good thing for future CEs. The original machines had the power supply tubs opening to the rear of the machines. This meant all the cables had to be removed from the tailgate to gain access to the power supplies. The first thing the Engineers did when they returned was to redesign the tubs to open to the front. Al Broadwater was one of the original CEs sent to Alaska and he later became my Field Manager at Lockheed.

The 7090 had a very elaborate remote control motor driven marginal check voltage system . You could select which power supply you wanted and vary it from the 7151 main console. This was our main method of picking off troubles before they got bad enough to impact the customer. Attached is a picture of the console controls

This system is much more complex than a 1401. Trouble shooting can be a real pain as there was no parity checking anywhere in the system except data from the tape drives to the TAU. You always started with one of three symptoms, incorrect answers, program loops or program stops when a word executed as an instruction had 00 in the op code field. You had no idea which box to work on unless some diagnostic program would also fail. You get complaints like, "Every time I run my matrix inversion program it gets different answers, What's wrong? I used to tell people the 704, 709 and 7090s separated the men from the boys. Getting one to work is going to be like going from a Mercury capsule to a Space Shuttle.

I better quit now. I had cataract surgery on my right eye last Thursday and it's been raining for two days so I'm bored.

Van Gardner
Smyrna, GA