clockcounter
I need to supply 460.8kHz clock to IC (NCR20C12). However, I can only get 1.8432MHz Oscillator. How can generate clock divided by 4 from it? I have a couple of SN74LS393 counters and want to ask, is it possible to divide clock with counter and how to do it? Can't find solution on google, may be because of messing some terminology.
Is the following circuit correct?
simulate this circuit – Schematic created using CircuitLab
Sorry for my bad english, stupid question and thanks in advance.
You are probably correct about why you're always seeing a "1" on the display on power up. As C3 is charging, the Schmitt Trigger U2A will see it as a low voltage until the voltage on C3 exceeds U2A's minimum threshold. With such a large value for R1, it takes a long time to charge up C3. I'm assuming the pendulum is relatively slow if you're using such a low frequency low-pass filter.
An elegant way to prevent that erroneous behavior is to put a similar low-pass filter on the ~CLR pin of U4. If you choose an RC time constant this is slightly larger than \$R1*C3\$, it'll guarantee that U4 will come out of reset only after U2A has stabilized.
The ~RBI feature is actually easier than I think you are thinking it is. If you just hard-wire it to ground, the chip will automatically blank the display for zeros, but illuminate normally for anything besides zero. In other words, you can get rid of U9 completely and just tie the pin to ground. You can see from the datasheet that if ~RBI is driven low AND all of the inputs are low, the display will be blanked. If any of the inputs are not low (a non-zero number), it doesn't matter what ~RBI is, it will actively display the number.
simulate this circuit – Schematic created using CircuitLab
This tends to work better with a Schmitt-trigger NOT gate (one with some hysteresis, or difference between "voltage it changes state on when going down" and "voltage it changes state on going up.") Circuit-lab's logic selection does not appear to have any Schmitt-trigger parts. Incidentally a Schmitt-trigger NAND with the inputs tied together also makes a Schmitt-trigger NOT.
It's a crude solution in any case and the R, C values are highly unlikely to give the precise oscillation you want. A 555-timer or 555 timer into a flip-flop is a somewhat less crude approach.
The frequency is one over the time for one full cycle. A full cycle is 1.5 seconds, so the frequency is 0.666666666666666666666... Hz (0.66666667 in most calculators due to rounding.)
Best Answer
You could do it with your SN74LS393 counter chip by using the correct output but you cannot simply connect the crystal like that. You need to turn the output of the crystal into a digital pulse.
You can do this in any number of ways but the most simple is a "Pierce Oscillator":
U1 is an inverter, something like a 74HC04 or you can find tiny SMD parts in SOT/SC-70 cases with a single inverter. C1 and C2 should be chosen according to the capacitive load your crystal expects, check the datasheet. R1 is a feedback resistor, you could start with a high value and work downwards until you find it is stable.
Another common way to do this is using D flip-flops, which is essentially how the internals of your counter chip will be working. The Wikipedia page on frequency dividers illustrates a divide-by-four counter nicely:
I have previously done this using 74HCT74 ICs.