Home Page . . . . . . . Email: . . . . . . . . neelandan-at-gmãil·çöm
This circuit was designed to monitor a seconds pulse
generator for a nixie clock which was suspected of irregular
running. Was it missing pulses? Was it picking up noise and
producing extra pulses? The best way to find out was to compare
the test output with a known good one.
Sixteen light emitting diodes are arranged in a circle to light up in sequence. A four bit binary up/down counter has its outputs decoded to drive sixteen leds. The reference signal is fed to the "up" input. The test signal is fed to the "down" input. If the two frequencies are exactly same, the same led remains lit. If they are (slightly) different, the spot of light travels arount the ring at the difference in frequency. Thus this circuit is useful for comparing two signals that are nearly, but not quite, equal in frequency.
For convenience, the circuit was built on two boards:
one contains the counter, decoder and leds. The other contains
the power supply and analogue circuitry.
The counter is a vintage 74193 pulled off a junked circuit board. It is, according to the data sheet, a synchronous up/down binary counter with separate count up and count down inputs. Three outputs of this counter drive a 7442 BCD to Decimal decoder, here used as a three bit binary to eight decoder. Each output connects to two leds, the anodes of which are commoned and driven by transistors controlled by the fourth output of the counter. The counter inputs are fed from two monostables contained in a 74123 dual monostable multivibrator circuit.
The other board has a 5 V power supply regulated by a
7805 and two two-transistor amplifiers to process the two signals
whose frequencies are to be compared. It was made in two parts
mainly to aid in troubleshooting because the discrete amplifier
stages will be a bit difficult to set up to be both sensitive and
responsive to high frequencies. The present arrangement seems to
work, at least at the low frequencies at which I tested it. A
front end using a F.E.T would be more appropriate for use at
frequencies above a few megahertz.
The circuit diagram shows only the second channel of the two inputs, the first channel is identical. The source of power is a 9V "wall wart" type of mains adapter. A more permanent solution would be to incorporate a bridge rectifier and a 9V transformer, and to build it into an insulated plastic cabinet.
It would almost always be necessary to modify the circuit if you were to attempt to build it. This was built as an attempt at using up some old components. The sixteen way decoder, for example, could be a 74LS154 - a 24-pin integrated circuit. The other two circuits could be replaced by their modern equivalents. If HCMOS circuits are used, all unused inputs will need to be tied either high or low. In this case, however, inputs at 'high' are just left open circuit.
If you could arrange to feed the necessary frequencies successively to the reference input, this could form the basis for a solid state guitar tuner - you can tell by the direction of rotation whether the string is sharp or flat, and by the speed just how much.
Home Page . . . . . . . Email: . . . . . . . . neelandan-at-gmãil·çöm