Counter miscounting?

[alec_t]

I would have expected a continuous 256-step ramp output from this sim with a 4024 counter, but there are clear discontinuities at multiples of 16 clock pulses. Any idea why this is, or how to overcome apparent miscounting?

 

[alec_t]

In answer to my own question, the discontinuities arise because I failed to take into account the internal resistance of the Q outputs of the counter. It appears LTS models this as ~500 Ohm. By subtracting 500 Ohm from the value of each weighting resistor to compensate, the ramp output becomes monotonic. Something to bear in mind if you want to use a counter (or shift register) plus weighting resistors for synthesising a waveform.

 

[crutschow]

So here's a good example of a simulation discovering a somewhat subtle problem before the circuit is actually built.

 

[alec_t]

Yes. I've become a recent convert to simulation. I'm sure it will prove an invaluable tool for designing circuits and save a fortune in blown fuses, popped transistors etc. Plus, you don't need to keep a huge inventory of components for experimenting. It's important, though, to realise its limitations.

 

[crutschow]

Certainly you must be aware of the limitation of simulations as you should for all tools. I made note of your success with it because there seems to be a lot of overly negative comments about circuit simulations on some of these forums, which I think may discourage people from using a very useful tool. I find them very handy to verify the basic operation of an initial design where it can catch a lot of bugs before you build the circuit. You still need to breadboard it, of course, to validate the final design.

 

[alec_t]

Agreed. Theory and practice rarely coincide exactly, particularly in the realm of electronics. A case in point is oscillator circuits. What works as a sim sometimes doesn't in reality; and vice versa. The component models don't seem to take random stimuli and noise fully into account, and it's these which can initiate oscillation.

 

[crutschow]

Sometimes you can initiate oscillation in a simulation by momentarily inputting a small AC signal (using a voltage controlled switch for example) into the appropriate node of the oscillator to kick-start the circuit.

As you said, in the real circuit noise and other random factors will start the oscillation.