Now and then, you need a repeating one-shot, such as a multi-shot. For example, when an input logic signal or alarm condition occurs, you may need an immediate pulse to provoke a sequence that eventually resets the condition. But, if the resetting fails, you need to keep generating pulses until the message gets through. If you give up sending pulses, something as important as an entire space mission could have to wait for an automatic reset.

Although you can generate multiple reset pulses many ways, most methods are fraught with annoyances. The first pulse can be twice as long as the follow-on pulses. Or, as for the retriggerable one-shot, keeping the pulse generator on prolongs the pulse instead of making new ones. When turned off, the pulse generator sometimes cuts the last output pulse off, hatching short glitches.

The circuits in Fig 1a and 1b are gated Schmitt oscillators. The oscillators respond to a negative input by producing a pulse of fixed length, regardless of how long the input stays on. The oscillators refuse to operate for a certain period. If the input is still there, the devices produce another pulse of the same fixed length.

In its quiescent state, the signal at the junctions is high (Fig 1a). C₁ couples the high-going transient to the diodes when the output toggles. The diodes clip the signal to 0.6V above the positive rail. A relaxation cycle begins as C₁'s charge drains away, and its voltage heads for the Schmitt trigger's lower threshold.

When the circuit receives a continuous input signal, each subsequent relaxation cycle starts with the diodes driven high. The diodes start high, even through C₁'s voltage decays to the upper Schmitt level during the output pulse's on period. So, the circuit generates almost exactly the same pulse length every time, including the first time.

Fig 1b's circuit, the simpler of the two versions, lacks clamping diodes and a transistor. The circuit's first pulse is slightly longer because the transition starts at a higher voltage. Fig 1b's circuit is a trifle less temperature-stable than Fig 1a's. (DI #1636)