Following the Paper Trail

The machines produced during the personal computing revolution of the late 1970s and early 1980s are what most of us associate with vintage computing. Of course these were not the first computers, however. There is some debate about what the earliest computer actually was, as some of the contenders were purely mechanical and there is no consensus as to whether or not they should truly be labeled as computers. But the massive vacuum tube computers of the 1940s and 1950s were undeniably true computers, as foreign as they are to a modern eye.

YouTuber Usagi Electric is no stranger to these pre-transistor era beasts, and in fact he has a few in his workshop that he needs to fix up before the end of the year. Having just wrapped up the repair of a Bendix G-15 — an industrial control unit from 1956 — Usagi Electric turned his attention back to a homebrew vacuum tube computer called the UE1. This machine may be big, but it is definitely not powerful with just a single byte of memory and a 1-bit processor. But archaic as it may be, Usagi Electric is determined to get it working.

The last major component that should be needed to get the UE1 executing instructions is a paper tape reader — a cousin of the punch card reader that preceded disk drives. Usagi Electric built a custom solution for the UE1, but as it stood, it had some issues. Each line on the tape contains 9 holes, which encode 8 data bits and a clock signal. The problem was that 2 of the data bits were incorrect, always showing low or high signals, and the clock signal was not looking like it should either.

After some troubleshooting, it was found that a simple mistake had been made. A metal plate that holds the photodiodes that read the holes (or intact paper) as the tape passes by them was installed upside down, which shifted all of the bits by one position. Flipping that around nearly fixed the issue, but a bad solder joint on a resistor was also causing problems. Correcting that had the tape reader humming along like it was the 1950s all over again.

With the reader working, Usagi Electric still needed to build an inverting amplifier board to process the signals for consumption by the computer. It quickly became apparent that the signals coming from the photodiodes were not uniform due to uneven lighting, and perhaps the peculiarities of each photodiode, so potentiometers were added to allow for individual adjustments, which could be made by checking the signals with an oscilloscope.

The equipment was mounted onto the UE1, however Usagi Electric still needs to make some custom PCBs to feed the signals from the tape reader into the computer, and to halt the reader when it is sent a specific signal. So we will have to wait until next time to see if the UE1 is able to successfully run instructions, or if more problems will be discovered. Be sure to stay tuned for more fun with the history of computing!