Building a Frequency Counter

    This is a lab frequency counter which I built decades ago, shortly after the discovery of the electron by noted Russian scientist Igor Electronov. At that time I was a young student short on funds and commercial laboratory instruments were totally out of my reach so I had to build my own. Today it would probably not make sense to build this type of instrument as there are commercial units for very reasonable prices and it would be more productive to just work to earn the needed money. But, as I say, things were different then. I designed and built this instrument and it is so badly designed and built that it is a wonder it ever worked and yet decades later it is still working (although I hardly use it as I have a much better HP unit.
    I do not expect that anybody would have any interest in building this design today but I present it here to show the basic workings of a frequency counter and as a matter of personal histerical interest.
    The basic idea behind a frequency counter is quite simple. Suppose we have a signal of 66903 Hz. We feed this signal to a digital counter for one second and then we display the result which will be 66903. Of course, if we display the value in the counter all along the counting process we will just see the numbers flickering so what we do is that we provide a latch and the process is: 1- During one second a gate allows the incoming pulses to reach the counter and then gate is closed and no more pulses reach the counter.
2- A fraction of a second after the counting stops a transfer pulse transfers the value in the counter to a latch where it remains stable.
2- A fraction of a second after the transfer has ocurred a pulse resets the counter to zero so that the process can start all over again.
This is the essence of a frequency counter. All the rest is just implementing this basic concept.
    The period during which counting is disabled and the count transfer and the counter reset take place can be very short when compared to the counting period but it was simpler for me to take a 1 Hz signal and divide it by two so that I had a 0.5 Hz signal which when high would enable counting and when low would disable counting. This meant that the disabled period was the same length as the counting enabled period.
    With this timing period of 1 s and a five digit counter we can count up to 99.999 KHz. If we use 0.1 s we have just multiplied the range by ten up to 999.99 KHz. If we count for 0.01 s we can read up to 9,999.9 KHz and if we count for 1 ms we can read up to 99.999 MHz (assuming the counter can count that fast).
    If we wanted to count even lower frequencies we could have a 10 second counting period which would read and display up to 9,999.9 Hz but this reading would update only every 20 s (three times per minute) so it would be a bit too slow. When measuring low frequencies it is better to measure the period. Instead of measuring how many input signal pulses we have in one internal clock pulse we do the opposite, we count how many internal clock pulses we can fit in one external signal pulse. I had prepared to do this with this instrument as it only takes some simple switching but I never got around to doing it. I guess I never really needed it.
   
   Time Base
    So let us start with the time base which will provide the timing signal(s). Ideally I would have used a 1 MHz crystal to regulate an oscillator and then simply divide by 10 until I had the desired frequency but at that time crystals were *very* expensive. Somehow I got hold of a crystal with a frequency of 5230 Khz so I designed the circuit around this crystal. I designed an oscillator at this frequency and then I designed a digital circuit which divided by 5230 so that I had an output of 1 KHz.
    In binary 5230 is 1 0100 0110 1110 so I needed a 13 bit binary counter. The output of the oscillator was counted by three 7493 four-bit binary counters and one bit counter which was half of a 7473 chip. Some gates below detected when the seven bits were all set and would set the rest flip-flop formed by the two nand gates on the left.
    When gate X8 detected that all the counters had been reset it would reset the flip-flop and the counting would begin again. It was a complex design but, as I said, at that time crystals were very expensive and TTL chips were cheaper compared to crystals so this design made sense at the time.
    The 1KHz signal can be further divided by three cascaded SN7490 decade counters so we have a choice of four frequency ranges as mentioned above. The 1 Hz signal will give a range up to 99.999 KHz while on the other extreme the 1 KHz signal will give a range of 99.999 MHz (with a five digit counter which is what the instrument has).
    Now that I look at the design I realise the divider by 2 after the range switch could have gone before the switch, at point maked "A", and it would make the circuit a bit simpler and it would work just the same. Afetr all, the divider by two is half of a SN7473 chip and the other half is ised right up there. It makes no sense to take the 1 KHz signal to the range switch only to return it to the same ship if it can be avoided. Oh well, I did not realize this at the time. Too late now.
    After the range switch we already have the pulse count enable timing signal. A couple of transistors provide the delays needed to generate the transfer signal and the counter reset signals.
    The output of these three timing signals is like shown in the graphic shown further above.
    Today all this could be done in a much simpler way with integrated components but at that time it required the individual counters, etc.
   
   Counter, latch & display
   Pending
   

   
   
   Input signal conditioner
   Pending
   

   
   
   Photo of the finished unit
frequency counter

   The slide switch above the display switches the input signal between the HF and LF amplifier-conditioners. The black toggle switch is the mains power. The rotary switch has five positions for the different ranges. The metal mini toggle switch was supposed to invert the function and measure ms /cycle of low frequency signals but I never got around to completing that so the switch is not in use.