Original question:
[FAQ] [H] Monostable Multivibrators - Top Questions Answered
I want to generate a pulse that always falls between 20ns and 50ns over the entire temperature range of an industrial device. Is 20ns a feasible lower bound for this device? I have to extrapolate from the graphs in the datasheet to reach that pulse length, but nowhere is a minimum either specified or suggested.
Hi Ian,
Looking at the provided timing data:
I doubt that would be possible. The table is logarithmic and there is some nonlinearity at the minimum timing resistance and capacitance below 100pF. Most likely you'd be seeing a minimum pulse width between 10 and 100ns, and it wouldn't be consistent across voltage/temperature variations.
I don't know of an easy way to get that level of accuracy with standard logic parts. A precision comparator + RC would be my first pick of how to get a relatively consistent pulse width, but getting a capacitor with that level of accuracy across temp could be difficult / expensive.
You could get close using logic gates. For example, the SN74LVC244A has a delay between 1.5 and 5.7ns per gate across process/voltage/temperature variations, so putting all 8 channels in series would give you between 12ns and 45.6ns total.
