This page last updated on 04/17/2017.
Copyright © 2001-2017 by Russ Meyer
Over the summer, my family moved to Walla Walla, Washington. I entered 5th grade and found that all the kids had their multiplication skills down cold. They had been fully drilled in the tables during the prior year. I was woefully outclassed and fell behind in math. Eventually my teacher, Mrs. Snider, discovered my shortcomings and enrolled me in a remedial multiplication skills class. I eventually re-entered the normal math curriculum, but never really felt like I was on equal footing with my peers. My confidence in doing math had been shaken and this worrisome doubt nagged me until my last day in college. In many ways, this became my Achilles heel in college technical courses.
By 7th grade, I was only hesitantly confident of my multiplication skills. I would remember that 6x7 was 42, but so deep was my lack of self-confidence that I frequently felt compelled to add six 7s together to verify the result. I know it sounds neurotic and I suppose it was. Anyway, this made doing math homework very slow and tedious. My Dad had a calculator, but it was a prized family possession. It was not really available for use on mundane things like math homework. This was the mid-70s, and calculators were a fairly new invention with a high price tag to match. There was no such thing as a personal computer. Man, I sure could use some kind of apparatus to help get my math homework done faster. I began trying to think of a way to build a machine that would multiply.
I went to the library and scoured the shelves for anything on calculating machines. I found a book on analog computers. The book contained an illustration of a simple analog computer that would theoretically multiply two numbers. I was thrilled! The thing only needed three potentiometers, a voltmeter, and a power supply. I decided to try building it.
After a lot of work, I managed to secure the parts I needed and wired the thing up. I had a lot of trouble getting it to work right. You were supposed to set the potentiometers to correspond with the two operands. The product was supposed to appear on the voltmeter. The biggest problem was calibrating the potentiometers and voltmeter. I just couldn't figure out how to do that. I floundered with analog computer theory. Understanding how it worked was at the perimeter of my abilities, and I was hopelessly lost when it came to a finely detailed skill like calibrating it accurately. It was just beyond me. I didn't see any way I could get the thing to work accurately enough to help me with my homework. Frustrated, I tossed it aside and started over.
I decided to sit down with a clean sheet of paper and make up something myself to do the job. I struggled and struggled with this for months. I thought of elaborate banks of relays, switches, dials, meters, etc. I groped and groped for a logical approach. I hit upon an idea that used a bunch of relays to perform logic operations with the result represented by a series of illuminated ping-pong balls. It looked like the theory would work, but there was no point in pursuing it because I did not have the money to build it. Relays cost a lot of money and require big power supplies. Finally, on Thanksgiving day, 1976, I hit on an idea that used only switches...preferably rotary switches. Rotary switches were much more affordable than banks of relays!
Here's how I evolved the idea. I decided to use gas discharge tubes to display the answer. These things are called "nixie" tubes and have ten numeric shaped electrodes that glow when 250 VDC is applied. I started by thinking of how to control one nixie tube with a single rotary switch. I could easily connect a rotary switch in series with the electrodes. By doing this, I could make any sequence of numbers appear in the tube as I spun the switch. I could use this to present one digit of a product. For example, the "6" series multiplication table increments like this: 6, 12, 18, 24, 30, 36, 42, etc. Just looking at the "units" digit of that series gives: 6, 2, 8, 4, 0, 6, 2, etc. I could wire a rotary switch to present that sequence as it was successively clicked from one position to another. (see illustration) To display the "10s" digit, I would need a second rotary switch mechanically connected or "ganged" with the first switch. I could then wire this second switch to present a sequence of numbers appropriate for the 10s digit. For the "6 series," this second rotary switch/nixie tube combination would be wired to display <blank>, 1, 1, 2, 3, 3, 4, etc. Turning the knob on this one ganged rotary switch would then present the two digit sequence of numbers representing the "6 series" of the multiplication table: 6, 12, 18, 24, 30, 36, 42, etc. I could have a bunch of switches like this, one for each of the multiplication series I needed. To select which series switch was activated, I could put another rotary switch in series. This other switch could be used to activate one of the multiplication switches at a time. Setting the switch would basically correspond to entering one of the operands for multiplication. Sheezh, I had it...I had actually figured it out! The circuit was fairly simple and could be extended to handle as many input operands as needed.
So, summing up, here's a basic explanation of how the user operated it. There was one master switch called the "series" switch. The first operand was dialed into this switch. This setting would activate a corresponding "series bank." The "series bank" was a set of switches where the second operand was dialed in. When the two operands were dialed in, the product would appear in the nixie tube pair. In the example below, the switches are set to give the answer for 6x7.
I laboriously drew the entire schematic diagram of the thing and set to constructing the gizmo. As usual, the main limitation was money. I had enough to construct a power supply, buy a pair of nixie tubes and a few rotary switches, but not enough to finish the project. I pressed ahead anyway. There were a lot of interconnections required, and the hardest part was getting all of them soldered up. I worked on the thing several hours per day, day-after-day. Finally, in early December, I had the first few banks of switches wired up. It worked flawlessly, and I was very pleased! I needed gobs more rotary switches, and that is what I put on my Christmas wish list. Lo-and-behold, Santa came through with most of the remaining parts I needed. I was short several switches, but I had enough to finish construction of most of the rest of the machine. Here's a photo of the device under construction:
A few months later, I had acquired the last few needed rotary switches and wired them up. It was now a full fledged, two digit multiplying machine. I actually used it to help me with my home work a few times. Here are a few other photos of the thing:
After that project, I started thinking of other automatic information processing devices I could build. I wanted to make something that I could enter my friend's names and phone numbers into. It would work like an electronic rolodex. I would enter a friend's name and it would find his phone number. I worked on this for some time, and believe I figured it out. I used thin, enameled copper wire stretched across arrays of terminal strips. These acted as fusible links. To program a friend's name and number into the machine, I would select an un-programmed slot, configure some switches to represent the data, and throw another switch to run a bunch of current through the corresponding links. The selected links would vaporize creating a pattern of opened and closed circuits. To find an entry, I would re-enter the friend's name I was searching for into a series of switches. Throwing another switch would turn on an electric motor attached to a disk. Affixed to the disk was a wiper, and as the disk turned this wiper would brush over contacts underneath. These contacts were wired to the fusible links and then on through the switches representing the friend's name. When a set of fusible links was energized and that pattern matched the pattern set into the switches, a relay would close turning off the motor. This stopped the disk and kept the target contact energized. A second set of fusible links representing the phone number were used to route 250 VDC via relays to the output "nixies." This caused the friend's phone number to appear.
I pretty much finished designing this funky gizmo, and started building it. I didn't get too far before I abandon the effort. It was significantly more complex than the multiplying machine, and would have taken a lot of money and work to complete. To top it off, the whole problem I was trying to solve with it was pretty contrived. I actually could remember my friend's phone numbers well enough, and really didn't need the machine. It was just a novelty. I moved on to other things.
Now that I've been working with computers for years, I look back on those projects and recognize them for what they were. The multiplying machine was really just a Read Only Memory (ROM) device with a novel input and output method. At its heart, the rolodex gizmo was just a Programmable Read Only Memory (PROM) device with some funky input, output, and searching apparatus integrated with it. I knew nothing of ROMs and PROMs, even in principle, at the time I designed my gizmos. I just sort of independently stumbled upon those solutions. Sometimes I look back on those efforts with amazement. I was only 13 years old at the time. How could I have come up with something like that? I don't feel quite that mentally alert these days, that's for sure! Considering the time and energy I poured into solving these problems, I could have just memorized my multiplication tables and been done with it, but building the machines was a lot more fun. It certainly gave me a much greater sense of achievement.