1
\$\begingroup\$

I'm trying to get a 37 AWG, 0.6mm diameter nichrome wire to reach 300*C so it can heat up inside of an enclosure and according to its datasheet that requires 5A, I cut out enough nichrome wire to offer 2.4ohm. The source provides 12V and 5A max, the MOSFET used is STFH18N60M2(datasheet) and it acts as a switch and I'm sending a PWM signal(%100 duty cycle at the moment) to its gate from Arduino.

Now the nichrome wire does heat up, I don't know if it reaches 300*C but what went wrong is that the MOSFET apparently got too hot and melted the plastic of the breadboard where it was connected, I didn't think I would need a heatsink and I'm wondering if something is wrong with my circuit:

enter image description here

\$\endgroup\$
7
  • 2
    \$\begingroup\$ I don't know what voltage your Arduino outputs, but that MOSFET has and RdsON of 0.25 ohms even with 10V to the gate. If you're actually running 5A through it, P=I^2R=25*0.25=6.25W, so even with 10V to the gate you'd expect it to get hot. More likely you're putting 3.3 or 5V to the gate and your ON resistance is substantially higher. You might want to use a MOSFET with better RdsON and add a MOSFET driver if required. \$\endgroup\$ – K H Feb 18 at 11:00
  • 1
    \$\begingroup\$ Often, when your MOSFET gets too hot it means your MOSFET wasn't fully turned on because the gate voltage was too low. \$\endgroup\$ – user253751 Feb 18 at 11:06
  • 1
    \$\begingroup\$ Is the NMOS really turning on and off? Or just constantly conducting via the body diode? \$\endgroup\$ – vangelo Feb 18 at 11:09
  • 2
    \$\begingroup\$ On your schematic transistor is updown. Internal diode conducts, the load can not be controlled. And choose MOSFET with lower voltage, it is usually smaller resistance. \$\endgroup\$ – user263983 Feb 18 at 12:09
  • 2
    \$\begingroup\$ 5A in a breadboard? Are you sure it wasn't the overheating breadboard that heated the MOSFET? This needs to be soldered (and with adequate heatsinking on the MOSFET). \$\endgroup\$ – user16324 Feb 18 at 13:28
2
\$\begingroup\$

Your heater at 2.4 ohms with 12 V will draw a current of I = V/R = 12/2.4 = 5 A.

While that is within the capability of the MOSFET's 13 A rating, you are failing to account for the power dissipated in the MOSFET itself.

At 5 A and 0.280 ohms RdsON you will get P = I^2*R = 5^2 * 0.280 = 25 * 0.280 = 7 watts. The datasheet shows a Rthj-amb of 62.5 C/W.

So 7 watts will give you 7 * 62.5 = 437 C! That's WAY too hot. The max operating temperature is 150 C.

You need a heatsink capable of keeping the device below 150 C or you need a different MOSFET that will not dissipate so much heat.

\$\endgroup\$
4
  • \$\begingroup\$ Right, thanks... I didn't know anything about the Rthj-amb, guess I had to learn that the hard way. I think I will be better off with a different MOSFET with better suited parameters, I've seen most MOSFETs used in this application are connected to the Arduino(5V) through a BJT circuit, would that be something you'd advise I add to my circuit as well ? \$\endgroup\$ – BigRedMachine Feb 18 at 11:17
  • 1
    \$\begingroup\$ As noted in one of the comments, to minimize power dissipation in a MOSFET you need to ensure that it's 100% on. The Arduino output may not be able to do that. Using a BJT allows you to buffer the output and make sure it fully switches on the MOSFET. But with a 5A load you will likely find that you always need a heatsink of some sort. BTW - this is "engineering". You don't just hook a bunch of parts together and hope for the best. You read specs, you design, you calculate, you test, then you have a working system. \$\endgroup\$ – jwh20 Feb 18 at 11:29
  • 1
    \$\begingroup\$ I would definitely advise just getting a better mosfet. There are many options that have far better characteristics than this one, for example, (infineon.com/dgdl/…) will be like >20x better than what you have now \$\endgroup\$ – BeB00 Feb 18 at 11:31
  • 1
    \$\begingroup\$ If you spend a little bit more, you can get something like this (rocelec.widen.net/view/pdf/alq61corkr/…) which should allow you to squeak by without a heatsink, at <100mW power dissipation \$\endgroup\$ – BeB00 Feb 18 at 11:35
1
\$\begingroup\$

Besides all the correct comments regarding a better MOSFET model, here is a circuit which takes a [0 to 5V] input (low is on, high is off) with 12V load. It also correctly polarizes a NMOS (the 'source' pin is connected to 0V and not inverted like in your schematics), guaranteeing a higher Vgs (you should also check this when you choose another MOSFET):

enter image description here

enter image description here

\$\endgroup\$

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.