the MOSFET won't receive a PWM signal. It will just be used as a ON/OFF switch. I still will put the recommended decoupling capacitors in the circuit.

PWM or not, every time the heater is switched ON and OFF, huge spikes are put on the power supply lines.

 

PWM or not, every time the heater is switched ON and OFF, huge spikes are put on the power supply lines.

So I also need a decoupling capacitor between the gate and source to prevent that?

 

So I also need a decoupling capacitor between the gate and source to prevent that?

Yes and no. Switching on and off large currents will always have the potential to disturb power rails. What can be done to reduce this effect is to slow down the rise and fall ramps of the switching circuit. A capacitor on the gate will do this by creating a low pass filter (which is also an integrator).

I would leave this out for this application at this time.

 

Yes and no. Switching on and off large currents will always have the potential to disturb power rails. What can be done to reduce this effect is to slow down the rise and fall ramps of the switching circuit. A capacitor on the gate will do this by creating a low pass filter (which is also an integrator).

I would leave this out for this application at this time.

R4 (the gate resistor) also helps with that right?

 

The gate capacitance of AOD4148A is 1500 pF.
With R4 = 100 Ω, the RC time-constant is 150 ns.
10 nF capacitance would give a time-constant of 1 μs.
Or you can increase R4 to 1 kΩ.

 

What seems like a good idea (and I have done this before) layout the circuit with bypass capacitors and then don't load the actual component. If everything looks fine, just leave them off and if you want, on the next revision of the board, remove them. It is a very low cost way to cover you posterior.

 

What seems like a good idea (and I have done this before) layout the circuit with bypass capacitors and then don't load the actual component. If everything looks fine, just leave them off and if you want, on the next revision of the board, remove them. It is a very low cost way to cover you posterior.

It's not so much the cost that matters. However, I have very limited space, so an electrolytic capacitor might not fit. But the approach you describe seems like a good idea, thanks

 

The gate capacitance of AOD4148A is 1500 pF.
With R4 = 100 Ω, the RC time-constant is 150 ns.
10 nF capacitance would give a time-constant of 1 μs.
Or you can increase R4 to 1 kΩ.

Wouldn't that cause additional heat generation in the MOSFET because it stays in the 'not fully on' state for longer? Or does it not matter because it doesn't switch often, as it is used as an on/off switch?

 

Yes, heat is generated when the transistor is in the linear region.
And you are also correct. It depends on how frequently the transistor is switched.

 

Yes, heat is generated when the transistor is in the linear region.
And you are also correct. It depends on how frequently the transistor is switched.

Okay, so if I understand correctly, I need to make the following adjustments to my circuit to improve its reliability:

• Add a 100nF ceramic capacitor and a 10µF electrolytic capacitor between VCC and GND of the 555 IC.
  
  
• Increase R4 from 100Ω to 1000Ω to extend the MOSFET's on-time witch will reduce voltage spikes during switching.

 

“to improve its reliability”

Make a note that the reasons for doing this is for the improved reliability of other circuits in the whole system and not necessarily for this part of the circuit.

 

“to improve its reliability”

Make a note that the reasons for doing this is for the improved reliability of other circuits in the whole system and not necessarily for this part of the circuit.

Do you mean other components on this PCB (which are not present) or other electronics in the car?

 

What I mean is that you can ignore everything said on this thread and you will not notice any difference for your situation.
The problem surfaces when you try to design a circuit like this very same one in a different situation, i.e. where there are other components that are sensitive to noise on the supply line. So yes, other electronics in the car.

I have been there, done that, seen it.

I once designed a circuit that used an NE555 timer. Another part of the circuit used a monostable multivibrator which was required to generate a pulse with precise pulse-width. It took me awhile to diagnose why the pulse-width was always shorter than expected. The problem was resolved by adding proper decoupling capacitors. It was further improved by changing to a CMOS 555-timer, LMC555.

Another example was when a sensitive circuit randomly misfired. In this case, the culprit was the soldering iron station which periodically switched on and off for temperature regulation. The problem went away (though not resolved) when the soldering station was turned off.

 

What I mean is that you can ignore everything said on this thread and you will not notice any difference for your situation.
The problem surfaces when you try to design a circuit like this very same one in a different situation, i.e. where there are other components that are sensitive to noise on the supply line. So yes, other electronics in the car.

I have been there, done that, seen it.

I once designed a circuit that used an NE555 timer. Another part of the circuit used a monostable multivibrator which was required to generate a pulse with precise pulse-width. It took me awhile to diagnose why the pulse-width was always shorter than expected. The problem was resolved by adding proper decoupling capacitors. It was further improved by changing to a CMOS 555-timer, LMC555.

Another example was when a sensitive circuit randomly misfired. In this case, the culprit was the soldering iron station which periodically switched on and off for temperature regulation. The problem went away (though not resolved) when the soldering station was turned off.

Clear. Thanks a lot for your input. Since I might want to add feedback to the circuit in the future (temperature sensor and a variable PWM signal), I'll go ahead and use decoupling capacitors just to be sure.

 

Clear. Thanks a lot for your input. Since I might want to add feedback to the circuit in the future (temperature sensor and a variable PWM signal), I'll go ahead and use decoupling capacitors just to be sure.

It is generally never a choice. You go ahead and do it always. SMD components will reduce the footprint and are always better than thru-hole capacitors for reducing noise.

 

I have a small issue with the circuit discussed in this thread. When the circuit is powered on, the 555 timer turns on by itself (false trigger). Is there something wrong with my circuit causing this, or is this normal? Also, I'm curious about the best solution to prevent this issue.

 

I have a small issue with the circuit discussed in this thread. When the circuit is powered on, the 555 timer turns on by itself (false trigger). Is there something wrong with my circuit causing this, or is this normal? Also, I'm curious about the best solution to prevent this issue.

That's quite common with 555s. If it causes a problem, then a capacitor to 0V on RESET and a pull-up resistor should stop it.

 

That's quite common with 555s. If it causes a problem, then a capacitor to 0V on RESET and a pull-up resistor should stop it.

Thanks for your reaction, what values du you suggest for these components?

 

Thanks for your reaction, what values du you suggest for these components?

It will depend on how quickly your supply voltage rises, but try 100nF 47k.

 

It will depend on how quickly your supply voltage rises, but try 100nF 47k.

It gets powerd by a big 12v lead acid battery (through a relay). I will try the values that you suggested