0
\$\begingroup\$

I am working on a circuit design where a heater (10 Watts) load is switched ON and OFF using a relay. The relay is in turn switched ON and OFF from N-MOSFET. MOSFET is switched ON and OFF using a Micro-controller and an Opto-Isolator. Our concern is that in case of MOSFET failing short/closed the relay will always be ON and the heater will stay ON. Is there any way in the switching circuit design to cut off the power in case of MOSFET failing short? Or is there a way that MOSFET always fails OPEN?

Thanks in advance for your time and help.

schematic

simulate this circuit – Schematic created using CircuitLab

\$\endgroup\$
7
  • \$\begingroup\$ A thermistor, pot and FET switch with another driver for hysteresis is generally adequate, but a secondary switch to latch OTP power off, if needed would need a secondary switch with memory. \$\endgroup\$ Sep 5 '17 at 19:19
  • \$\begingroup\$ The mosfet won't last long with no flyback diode across that relay... \$\endgroup\$
    – Trevor_G
    Sep 5 '17 at 19:58
  • \$\begingroup\$ And it's not just the MOSFET... what happens if D1 fails short, or the relay fuses itself closed... \$\endgroup\$
    – Trevor_G
    Sep 5 '17 at 20:05
  • \$\begingroup\$ @Trevor if D1 has a zener voltage smaller than the breakdown of the mosfet, that should be fine. Yes, still I would put a flyback diode instead of D1. \$\endgroup\$
    – next-hack
    Sep 5 '17 at 20:15
  • \$\begingroup\$ @Almeida If you're worried about that FET failing in short, then you could add another in series... Yes, this might fail too, but what's the probability then? \$\endgroup\$
    – next-hack
    Sep 5 '17 at 20:21
1
\$\begingroup\$

In high-reliability (like automotive) applications the Central Processing Unit usually employs checks on critical circuitry, which requires additional GPIOs, to watch if the command to a sensor or some action circuit gets executed properly. In this case if M1 fails with short, and CPU commands it to be OFF, the GPIO input will remain low, indicating fault condition. Then the CPU can turn corresponding power rail off (for which it should have another power gate to control). Likewise, if the M1 fails with OPEN, the same watch mechanism will detect the "open" fault.

\$\endgroup\$
1
  • \$\begingroup\$ Thanks Ali Chen! We have few spare GPIO to sense the fault condition and to control an additional switch for power rail. \$\endgroup\$
    – Almeida
    Sep 6 '17 at 21:40

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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