[time-nuts] Mechanical 1PPS Oscillator Disciplining
namichie at gmail.com
Sat Jan 10 19:32:34 EST 2015
Back in about 1962 I joined a government research lab. My boss had bought a HP voltmeter (I remember it as being about a yard cubed)
and a HP printer (also a yard cubed) and it printed out on 2 inch wide paper, like a cash register record.
My boss's problem was that he wanted more precision in measuring a time dependent process. My problem was to trigger the voltmeter/printer to sample at exact intervals of time. I found a tuning fork standard (might have been GR) of 400 hertz, so a I built a divider using neon Dekatrons. These had ten electrodes in a circle visible from the end, and subsidiary electrodes would steer the glow to the next electrode when the tube was triggered.
Each of these anodes must have had a separate load resistor, because I found each one had a voltage drop when it was illuminated. I used this with rotary selector switches to select divide integers by using the anode pulse to trigger a monostable (12AX7 ) to reset the Dekatron to zero and start counting again.
We did not have logic circuits in those days, this was known as "Pulse Techniques" as used in military radar, which is where I got much of my early education.
I have forgotten the voltmeter resolution it might have only been 4, 5, or 6 digits, but the data from the paper tape was then hand typed into an IBM card punch and the cards were couriered to head office to be fed into the main frame (probably IBM)
We have come a long way in data acquisition and computing since then.
More information about the time-nuts