I've designed a very simple heating circuit that runs a heater until a certain temperature has been reached (as measured by an NTC thermistor) at which point the heater turns off. The circuit diagram is:

enter image description here

However I am concerned about having a heater that is always-on, in case the thermistor fails, or something along those lines. Would anyone mind verifying the correctness of my circuit, and explain to me how I can use an NTC thermistor to make it so the heater only turns on when the temperature drops below a certain setpoint, rather than having it turn off when the temperature goes above a certain setpoint? Or do I need a PTC thermistor?

  • \$\begingroup\$ Interesting, using a 555 timer as a comparator. Where did you find this circuit? \$\endgroup\$ Jan 4, 2022 at 17:33
  • 1
    \$\begingroup\$ @evildemonic, that's got long history as the 555's fundamentally a pair of comparators driving an SR latch. And a 555 in astable mode is just an RC, a comparator with hysteresis, a latch and a discharge transistor. \$\endgroup\$
    – TonyM
    Jan 4, 2022 at 17:40
  • \$\begingroup\$ Check pin nº 5 (CONT), is it connected to ground or is a mistake? \$\endgroup\$
    – Bravale
    Jan 4, 2022 at 18:07
  • \$\begingroup\$ @Bravale I think you're right, I simulated it and the output was stuck low with it like that. I disconnected pin 5 from ground and with a rising resistance in the bottom leg of the divider the output was high until the trigger voltage reached 2/3 of Vcc and then dropped low. Pin 5 on a 555 usually has a capacitor between it and ground, something like 0.1uF. \$\endgroup\$
    – GodJihyo
    Jan 4, 2022 at 19:04

4 Answers 4


You could try putting the thermistor in the bottom leg of the divider, between pin 6 and VR2 instead of between pin 6 and VR1, then the voltage at pin 6 would go down with rising temperature instead of up.

After doing some simulations I have some information to add:

With the heater connected to the NO relay contacts, the relay needs to be energized to turn the heater on. This will shut the heater off if power to this circuit fails, which is what you would want.

As for a thermistor failure, it would depend on if it fails shorted or open. If the top leg of the divider opens, the relay will shut off. If the bottom leg opens, the relay will stay on. So if it's more likely that the sensor will fail open, put it in the top leg and use a PTC, if it's more likely to fail shorted, put it in the bottom and use an NTC.


You're really saying the same thing:

NTC thermistor to make it so the heater only turns on when the temperature drops below a certain setpoint, rather than having it turn off when the temperature goes above a certain setpoint?

If the wires to the thermistor get broken the heater will stay on though, since high resistance = low temperature. One possible way to deal with it is like this:


simulate this circuit – Schematic created using CircuitLab

You can control, to some extent, the maximum sensor resistance before it shuts down by changing the values of R4, R4' (and R3, if necessary).

In general you should always have some secondary independent means of dealing with failure of the controller though, such that there will be no danger to life or significant other losses. Methods include mechanical thermostats or thermal fuses or secondary limit controllers. Electronics will eventually fail, and not necessarily in the most desirable state (for example, the relay could stick 'on').


It would mean more circuitry, but you could have two comparators and a bit of logic. Set it up so that the NTC has to indicate a temperature between something absurdly low (-100C?) and your set point in order for the thermostat to work.

Use an LM393 dual comparator and you'll (A) still be using an 8-pin package, and (B) it's open-collector, so you can use a wired-OR configuration to implement your window comparator.

That still won't prevent some fault from falsely keeping the heater on -- as mentioned elsewhere, you should have some secondary safety system, such as a mechanical thermostat, or make the heater wimpy enough that it won't catch anything on fire if it's left on, etc.


I love that you are using a 555! Great part.

However, I'm confused. If the thermistor fails in a way that turns on the heater, how will you be able to tell that situation from the situation where the heater is just required all the time? My heater is currently on most of the time.

I suppose if you are worried, you could have two thermistors, and somehow test them both. You could build two identical 555 circuits, and only turn on if both of them want you to turn on. This solves the issue until the second thermistor fails on.

If you really want redundancy, use an arduino. Have the two thermistors be read on separate ports, compare them, and if they don't compare, use a piezo beeper to annoy the heck out of your cat until you fix it. Make sure the piezo buzzer goes off if the arduino dies too... use a watchdog, and a pulldown on the buzzer.


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.