555 one shot missing a beat

Question

I built a monostable multivibrator circuit with a 7555 chip.

It has 1MΩ and 4.7μF components for timing, giving a pulse width of 5.17 seconds. It also has an RC on the trigger with 10kΩ and 0.33μF for a time constant of 3.3ms, allowing it to trigger at power on. Power is provided by a 5V signal.

Everything looks happy and as expected, for the most part, except there is a weird situation that I can't seem to get a handle on.

For example, here I'm using a pulse function generator on the input. This input is a single power input for the VCC and trigger/timer RC filters. When I bring the 5V signal high for ~2 seconds, then bring it low for 50ms, it retriggers and starts over, just as I would want it to.

BUT! When I bring that signal high for ~2 seconds and then low for just 10ms, something happens where the timer is not restarted. If I keep that input going, it only functions every other cycle.

If it was the trigger signal not having time to get low enough, that wouldn't trigger every other time. It can't be the timing capacitor not discharging fast enough because we see it is getting right down to 0. Scope traces are trigger signal in purple, threshold/discharge signal in yellow.

Why in the world could this be happening?

Answer 1

You mustn't be switching VCC!

By doing that you're relying on the power-up behaviour of the ICM7555 every time. And that's undefined, far as I read the datasheet. So, what happens isn't that surprising: you're taking away the power supply, and that puts TR, CV, DIS and THR all above the supply voltage. That is outside of what is allowed for operation. In fact, the datasheet even says that's outside of the maximum (damage-free) regions. So, that's a design fault!

Instead, keep VCC at a constant voltage (e.g. 5V, or whatever between 4.7 and 18 V youhave at hand), and only change the voltage at the TR input.

(also, unless you're doing anything else with the same pulse, too: The job of a pulse generator is to make defined-width pulses. Not quite sure you really want your circuit if you have a pulse generator. Also, resistors and large capacitors are not exact at all and thermally stable; a (7)55 is really not an acceptable approach if you need precision in timing.)