I have a relay that I want to have a 3 second delay switch off from when a pressure switch (normally open) returns to the open state. When the pressure is below a certain level the pressure switch closes, once the pressure has reached the set level the switch opens.

I've built this circuit on my breadboard tonight running at 12 V and it gives me a 3 second delay, driving an automotive relay 100 ohm coil resistance.

Does anyone have any suggestions on improvements?

enter image description here

I've updated the design with the great feedback received. The LED was a white LED at the time, but I'm intending to use a green one in the final design so altered the resistor. I've made the changes as suggested, such as a current limiting resistor for SW1. With regards to the back EMF diode I saw this post (Why don't relays incorporate flyback diodes?). Sand used design 4. It said the Zener voltage should be 2*vcc so I put a 24v zener in. Does this sound right? Also should I remove D1 now?

enter image description here

  • \$\begingroup\$ If it works, why would you look for improvements? \$\endgroup\$ Jul 22, 2021 at 21:34
  • \$\begingroup\$ The switch shorting out the capacitor is not good for the switch. Put a low Ohm resistor in series with the switch to limit the short circuit current. A 555 ic might be another choice rather than an opamp acting as a comparator. \$\endgroup\$
    – Kartman
    Jul 22, 2021 at 21:47
  • \$\begingroup\$ I have added an edit. \$\endgroup\$
    – user173271
    Jul 25, 2021 at 17:37

2 Answers 2


If that's an ordinary red LED and the op amp saturates at say 10V then there is about 20mA through the LED. Although this would be just about within the LED's spec' it would usually be an unnecessarily high LED current when the LED is just being used as an indicator. I would consider increasing R4 to reduce the LED's current to below 10mA which should still give the LED sufficient brightness.

.... also, when an LED is being used as a "RELAY ON" indicator it (and its series resistor) would usually be seen in parallel with the relay coil so that the transistor is sinking the LED's current rather than the op amp sourcing it.

You could turn the comparator into a Schmitt trigger by adding some hysteresis to the comparator. This would be done by adding say a 1M resistor between pins 1 & 3. The idea is to create two switching levels, one capacitor charging and two capacitor discharging. This should then remove any possible relay chatter problems caused by a slowly charging capacitor with any noise about. Hysteresis will slightly alter the timing of the circuit as it currently stands.

If you decide not to utilise hysteresis then at the very least there should be a capacitor (say 47uF) across R2 to filter out any supply ripple/noise getting to the threshold level but you can't add this capacitor if you decide to add positive feedback (hysteresis).


The zener, when used in that position, should have a value of between Vcc and the maximum Vce specification of the transistor (perhaps not too close to Vce(max) to give a safety margin). Using a zener, rather than an ordinary diode, has the advantage that the energy in the relay's coil is removed faster thereby switching off the relay quicker but this is only necessary if a faster switch off is required and in the vast majority of applications an ordinary diode will suffice. The higher the zener voltage, the faster the switch off but keep the zener value less than Vce(max). You require either an ordinary diode or a zener but not both. If you were to use both then the generated back EMF at transistor switch off time would be limited by the ordinary diode to 0.7V above Vcc and the zener, which would never come into service, would be redundant.

  • \$\begingroup\$ Thank you both Tony and James for all your help. I wish I could accept both your answers (I have up voted both). I think I have alot to learn. \$\endgroup\$
    – Trev
    Jul 26, 2021 at 21:09

A couple minor concerns.

  • If the switch is flash gold-plated (<=2A type) reduce current to 10mA with series R otherwise, leave as is to purge oxide.
  • the relay contacts for long life on discharge also need a reverse diode to gnd. from pump flyback arc.
  • the design margin, if it were for production, needs improvement in several areas to eliminate e-cap, add 50% hysteresis, using 10M charge R and 0.2uF ceramic with the larger ramp threshold using a low input current active low comparator.
  • if there are quality standards or cost reductions expected, other considerations might apply, but for DIY, it's good enough for gov't work.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.