I must be getting really old. I was just browsing around on the Web the other day, looking up one thing and finding 47 other unrelated topics. At any rate I came across one of the main computers I used to be a computer service technician in the 70's and 80's. In those days I was a top computer technician and we were good. The companies IBM, Univac, Burroughs, RCA were all in heavy competition and produced brilliantly engineered computers and periferals. They also trained the technical staff and all the personnel to be the very best of the best. The components were all highly engineered to NASA like standards. The main customers were the Military, Government and Educational institutes. I guess that was because they were the only ones capable of affording these machines. Anyway those were the good old bad days and it brought back a flood of memories.

The 9400 computer described below is in working condition and resides in Germany in a Museum. Thats the part that really makes me feel old, I actually worked on something that is now in a museum, whew! I have lent my comments to those from the Web Pages in Germany. Visit the museum here.

Univac 9400 Computer 1970-1980

Light permitting diodes (LED's) had not been invented yet. Little glow lamps were used to display the binary states because filament bulbs would have been too lazy to perform the rapid transistions in logic states. The display features four rotary switches so it can be switched between eight different state arrays with 18 bits.

The Console with the keyboard was not the main control of the computer like we do now with the keyboard and mouse. The operator would use all those switches to throw different things into gear. He would walk over to the tape drives shown and load a tape into one of the units then walk back and flip a few switches on the front panel and the tape drive would kick in load the data up. The Console would then start typing--slowly like a teletype machine--. It was a glorified typewriter with a ball.

The Operator was a very busy fellow running around and feeding the peripherals with data. He would have racks of 9 track tapes that he had to pick and choose from. The Disk Drives were monster plastic cases with 9 -16 metal platters that had to be loaded into the disk drive. And that printer could print out a box of paper in 5 minutes flat on some jobs. As well the main input was from punch cards, Hollerith coded with all the new information such as your telephone bill had to be loaded by the box full into the card reader and transferred to the tape and Disk Drives. A operator would go through nearly a pallet full of cards in a shift. The room itself had a raised floor and underneath was literally miles of cable to feed all the power hungry lights and heavy duty motors that were in use. Most installations had to have special power feeds into the office building for the computer. The room was air conditioned and kept real cool. It was tough in Toronto running from one building then outside into the hot humid summer weather and then back into the next building. I was constantly fighting a cold. I felt sorry for some of the gals working in the computer room. It was the time of the mini skirt, great for us guys but I sometimes wondered about the poor girls.

Front Panel Univac 9400

The Main Council was much more than a bunch of lights.The Big Knobs were mainly for us as service technicians. We could turn it and view all the working registers and data flow from anywhere in the computer. We did regular PM's on the machines and would run certain batches of tests regularly to check the condition of the machine. The machine was always having memory problems so we did overnight tests on it and examined the log file for errors. It never seemed to catch any problems as cycling the memory seemed to make it work better. I always remember that little thing about computers and their memory. After we finished the PM the machine would purr away for a few more days and the operators would think we were great technicians for making their machine run so well. Then it would start to fail again. So this time we would come in and change some boards around to see if that make it work better. At any rate communication was great between the technician and the operator and we kept it running no matter what it took. Sure would be nice to have that kind of service these days.

The Big White Rocker Switches as I said before were always being used by the operator to run the computer not from the console with the chair. The rocker switches activated a reed switch behind that made the contacts. An example of the engineering that went into the machines in those days. A mechanical switch would wear out in no time at all with all the pushes that went on those rocker switches.

Also note at the bottom left is a single toggle switch, cheap and dirty out of an old car it seems. That was the main power switch, turn that on and lights and motors fired up like you were on a rocket ship. We often wondered about that wonderful engineering team that designed the computers in those days. So good in some ways but just crazy in other ways. The IBM computers for years had a great big red knob very conspicuously placed. This was the emergency off switch. Now why you would have to shut down a computer in a emergency situation I can't image.

9 Track Tape Drives

The tape drives UNISERVO 12 (at the left and the right side) and UNISERVO 16 (center). Reading and writing is performed in blocks. To avoid data corruption (overlapping), the timings of the accleration of the band has to be very correct. While the "cheap" devices UNISERVO 12 control these operations electromechanically, the UNISERVO 16 features a pneumatically driven system. This is very heavy and expensive, but the drive is three times faster than the other.

These Tape machines were wonderful devices to work on. They were like glorified tape recorders. They recorded nine tracks of data (one byte of data) and a check bit per thousands/inch . There were two methods of recording the data one was NRZI (non return to zero) and I forget the other. At any rate they were extremley reliable and could store a massive amount of data for that time. I think it was something like 10 Megs. All Input Data was first transferred to the tape drive not the Disk Drive for storage.Large Computer Centers would have a huge room with rows of tapes for arcival purposes and the running of different "Jobs" as they were called. Hardware wise they were wonderous also. The Uniservo 12 was the workhorse. You would wind the tape through these rollers like putting a film in an old projector. Then you would hit the Load button and these giant sucktion motors would fire up and suck the tape up into the columns. This suction would keep the tape tensioned "just right to glide into the pinch rollers and not put any pressure on the tape to stretch it. It was neat to watch it load. Now that fancy one the Uniservo 12 was for the lazy operators and had lots of problems loading right. We were always tweeking it up. The Idea was you just threw the tape in and locked it, and hit the button and the tape was supposed to be sucked into the box and fed up to the other reel automatically. It was sometimes like your car trying to start on a cold day. The motors would come up to speed and then it would miss catching the tape and you could hear the motors winding down. If it happened more than once the operator would get really pissed off and simply load it on to one of the other drives. They were fast threading that tape in and could often do it faster than the fancy autoloading tape drives. And naturally They always bitched at us service technicians like we designed it. The darn thing cost twice as much as the other standard drives so everybody compained to us about all the money they had to spend and what a horrible useless machine it was.

Disc drives have a shorter access time compared to tape drives. On the other side, the controlling of the drive and the 10 heads is very complex, so they needed a whole cabinet with electronics (in the photo only particulary mapped). The storage capacity is 7.5MB per disc pack. For the time of 1969, this were incredible dimensions.

Hard Disk Drives for the Univac 9400Now these hard disk drives as we called them were another engineering marvel. You can see the disk Pack enclosure sitting on top of the cabinet. The Disk was removable and if you wanted different storage You just went and got another pack. You would turn that handle on the top a quarter turn and then the top plastic part would come off. You then lifted the hood locked the platters into the Drive Cabinet. You then hit the button to load and your drive would spin up so it was ready.

The tape drives and the Hard drive were both fed into this giant selector/multiplexor channel. So both devises could be sending data at the same time to the CPU. This was all done in hardware not software so it was fast for the time.

Interesting to note is that the 19 inch rack on the right was not for control of the hard drives in normal day to day operation. They were for us to do maintenance and check out the operations. Note the "Hazardous Area" Warning Sign. We also had another magic box that we could do maintenance with that had a hajillion switches on it. You will note the large circuit breakers on that panel also. That outfit drew a lot of power.

What was really neat was that the actuator head floated above the surface of the disk on the cushion of air provided by the spinning disk. So the speed control of the drive motor was crucial. The term head crash came from this fact. If the platter was not spinning fast enough the head would crash into the surface and that grove would sure put a dent in your data retrieval. The motor that spun the platters was huge, around two horsepower and toke a lot of electronics to control. I never saw a crash in the 10 years that I worked for Univac.

The other fantastic thing was that the actuator that moved back and forth across the platters to pick up data was hydraulic. So there was this oil tank and all the periphery that is needed for a hydraulic cylinder. The position of the heads out on the Disk Platter was controlled by a glass graticule that we didn't dare touch as it had fine laser carved lines.

Hollerith Card ReaderThis heavy card reader can read 10 cards per second. It pulls cards with low pressure in the transport band. After reading, they can be sorted with the two stacking trays. You can see the card's complex way from the optical reading station to the tray.

Hollerith Punch cards were actually the heart of the computer system. For those of you that don't know what a punch card is here is a short explanation. It was a hard paper stock card about the thickness of a playing card and measured about 3"x 6". A keypunch machine would punch 8 square holes vertically on the card. It would do this for 80 columns on the card. Thus a card could have 360 holes in it. The eight holes vertically was equivalent to a byte of data and it was hollerith code the forerunner of ASCII code. This card was how data was input to the computer. (80 bytes per card)

I worked on keypunch machines for quite some time. For instance Bell Telephone had a room full of keypunch machines and operators to enter customer billing. A gal could fill in the information for a customers bill in about 2 minites flat. That would consist of about 3 punch cards or about 250 bytes. I think a box of 1000 cards was about 200 customers monthly billing entry. So you can imagine how many boxes of cards were produced just to do the monthly billing for the telephone company. The keypunch machine was quite the marvel in itself. One of our models actually printed the data on the top of the card. It was like a stand alone work station with a keyboard. You would put a stack of cards in the input tray, then type the data in, and boy could those gals could type, 60 words a minute or more: you would then hit the enter key and the machine would whomp away and punch up to 360 holes in the card. And was it ever noisy. Now can you imagine a room full of operators 30 or more and you get the idea of the noise and the production of punch cards.

Now all these boxes of punch cards ( pallets actually) were brought to this machine for input to the computer. Usually read to the Tape drive first for backup. The card reader shown here would correctly be called the card sorter. The cards were ran though the machine multiple times to sort the cards according to certain criteria. For instance it would make sure that each customer had 2 or three cards in order and then you would run it through again and sort it alphabetically. It was a marvel of mechanical engineering. As said above 10 cards per second is very very fast to watch. The cards floated through on a path of air and belts that was extremely complicated. Ask me as I have replaced many a belt in these machines.

Work flow was really organized though with this machine being the key to how the shift went. Most times there were two operators on duty and the main job or batch as they called it was to get the input data for the day into the computer. Once the card data was in for the day and stored on the tape drive the rest of the batches could proceed. Usually the business computers were running the Cobol Language. So they would run a batch job to take the data from the tape drive and run it into the program and store it to the Hard Drive. So a billing batch would get a new disk pack remember they were removable. Then they would backup the new data again on to the tape drive and archive it. Then the next batch job would come up Change Disk packs, change tapes, Run the Batch job recievables. Archieve the data back on to the tape. Etc Etc. The last batch job of the day was usually the print job which was like a tally of how you did for the day. Oh I should mention that this high speed card sorter was a luxury and really expensive. The 9200/9300 lower level but common models, had a single card reader that would feed in about a half a box of cards at a time. The 9400 model shown here was a higher level model meant to compete with the good old IBM 360 which was the workhorse standard in the computer industry for years.

Univac Drum PrinterEven small mainframes had a high-speed printer. This monster has a printing capacity of 26 lines per second and features 635 kg.
As you could play "music" with old chain printers by printing a special typeface (the River Kwai March was a typical sound), these jokes are not possible with this barrel printer.

We Called this a Drum printer. Some of the guys used the drum from old printers as a Lawn roller as it was so heavy, about 300 lbs. The drum had 132 characters per row on the drum. It was raised type like an old style print shop would have. Thus it would have a row of Capital A's then a Row of lower case a's. The drum had all the possible characters it would print on the drum and a row of 132 of them. Behind the Drum was a set of 132 hammers. (Little coil actuators). It would print a line of 132 characters in one shot . The Drum would be spinning around at a tremendous speed I forget what it was. At any rate it would wait for the first character it needed to come up on the drum and fire the hammer. It would press the paper forward on to an inked ribbon and the character would be on the paper. It would then wait for the next character to come up on the drum and fire the actuator for that character in the #2 position on the drum. It would thus proceed through all the characters it needed to print on that line and in the position required. After up to 132 characters had been printed the paper would be moved up one line in readiness for the next 132 characters. To see this in action was amazing as it was so smooth and fast. They had this one really high speed model, I forget the number on it, but it would use up a whole box of paper in less than 3 minutes. I remember testing it out on a snoopy picture made out of x's and o's. A whole box of those pictures in 3 minutes. I was handing them out to friends for years.

I think Univac was a leader in neat printers in those years. The bar printer or chain printer as some people called it was another neat printer. It was a long flat bar about 3 feet long with all the characters needed embossed print shop type on the bar. This bar was moved back and forth horizontally by a big chain about the size of the timing chain in the engine of a automobile. The chain moved around in a tank of oil for lubrication. One character was printed by a hammer firing the paper unto a inked ribbon, per pass of the bar. Again it would print a line at a time. I liked that printer too.

There was constant maintenance on these beasts of printers and special service needed for cheque runs tax notices etc. All the forms were special, printed sequentially so that old printer better not jam or miss characters. I remember one run for the city of Kelowna where I was on standby for the whole 6 hour run of Tax notices.

Revised 2013 by Larry Gentleman