Bootstrapping

It was 1984 and I was a Twin Otter flight and a white-knuckled chopper ride out of Khartoum and in the middle of the desert. It was my first day working in the Veribo. The Veribo was basically a big wooden box, the size of a small mobile home, with an aircon bolted on, and didn't look like much from the outside but the magic happened within.

Inside was the equipment for testing the cable telemetry boxes used in seismic exploration.

The telemetry boxes hooked into geophones and then hundreds of them were all connected up by cable sections over miles of exploration prospect. This network was known as "The Line".

The Line was connected at the roughly half-way point to something ignomiously named "the dogbox". The dogbox was the main command and control centre that committed all data to magnetic tape and printed the received seismic signals out onto thermally-sensitive paper.

Explosions were set off, shock waves travelled down through the Earth and bounced back, reflected by the layers of rock. Geophones (think microphones you stick in the dirt) converted the received sound waves into electrical signals, which were then amplified, filtered, digitized, encoded, time-stamped, synchronized and transmitted down The Line by the magic of the telemetry boxes. Each telemetry box also acted as a repeater, boosting and re-synchronizing the data as it moved down The Line.

In short it was all a miracle of analogue and digital electronics created by some very celever people.

The problem was the telemetry boxes would get thrown about like loaves of stale bread, despite the fact that they were several thousand dollars apiece. That's where I came in. When broken telemetry boxes came back in from the field to the main camp it was my job was to troubleshoot them and repair them at component level. This was back when component level troubleshooting was a Happening Thing - these days it isn't - you just throw out the old and in with the new. That wasn't economically viable for these boxes back in 1984 and that's why we had the Veribo.

The Veribo was essentially computer automated test equipment. In short you hooked up your telemetry box to a computer and it injected various signals, ran various tests, and lo and behold out came what the problem was. The trouble was the test didn't get very far if the power supply unit in the box was damaged or another major malfunction was in play, so this was where I dived in with a soldering iron and an oscilloscope and realized the three years I'd spent studying maths on my electronic engineering degree course was a complete waste of time.

But I am getting ahead of myself. Before any of that you had to boot up the Veribo at the beginning of the day, which was a bit more involved than just "switching it on".

First, the main computer controlling the Veribo had to be booted. That was a 4K Control Data Corporation microcomputer. First job of the day - enter the boostrap code. You had to open the manual to the bootstrapping page. Here was a listing of assembly language code, with machine code equivalents handily provided. Each line of the boostrap program had to be entered manually by flicking switches to the right position. You had a nice User Experience feature here because each switch had a corresponding red LED you could check against! So if the machine code was 0xF0 you'd have four switches up (on) and four LEDS lit, and four switches down (LEDs) off. You'd then press another button to enter that code in. At this point I can imagine today's Java programmers breaking out in a cold sweat. But we are just getting started here.

Once all the lines of machine code had been entered you now had a mini bootstrap program to play with. The code could be executed by pressing another button. There were no end of User Experience wins as this button was a large one with a green lamp in it!

Of course the boostrap program didn't do a lot - in fact it only had one function, to enable a paper tape reader to work. Yes, the main program that needed to be loaded was on a reel of punched paper tape! We had a box with about half a dozen or so of these paper tape reels in them, in various states of decomposition from heat, humidity and termites, but they basically did the same thing - bring up the Veribo's main computer fully.

So next job was to load the paper tape into the paper tape reader. Its little lamp could be seen quite readily and created a cosy glow in the Veribo. The paper tape reader worked by using a lamp and a series of light sensors so that the pattern of ones and zeroes on each line of the paper tape could be read. If there was a hole in the paper tape the light would shone through at that point and activate the sensor (a binary 1) if there was no hole you'd get a binary 0 (zero). Those Java programmers' toes are probably now curling up like bananas.

At this point you had reached the climactic event. Basically you pressed another button and all hell would break loose. The paper tape would go whizzing through the reader at what I can only call an alarming rate, and end up in a tangled heap on the floor of the Veribo unit. Next job was to wind up the paper tape, put it back in the box and hope that everything had worked. It might not have. The paper tape reader was a bit vulnerable to condensation. The high outside temperatures and aircon didn't help. Sometimes you made a mistake entering the bootstrap program and so you had to painstakingly flick switches, check LEDs and enter the minimal boostrap code again. All in all it was a right pain but could be a lot of fun too.

If it all worked then it was time for the real work to begin and that's when we'd whack a Bob Marley album into the cassette player, dig out the scope probes, and get the first patient in from the big pile of broken boxes outside the Veribo unit. Ah, I love the smell of solder in the morning...

Author's note: The telemetry system we used in Sudan was the Sercel 348BACKSLASH. A system that primarily consisted of discrete TTL components and analogue electronics. I later used the more modern Sercel 368 system in Nigeria. The telemetry boxes were more compact and used more modern types of components including more Integrated Circuits. You still had to boot up the Veribo though! :)

A nice page looking at Sercel's history can be found here:

The modern systems look very impressive indeed.