MOSFET troubles when controlling high current heaters (15A)

[v15va]

Hello!

Firstly, newbie here so apologies in advance if this is in the wrong section / I’ve made some stupid mistakes.

I initially built this circuit to control 1 heating element which draws 3A at 24.4v via a PID loop. I fed the PWM signal from pin 9 of my Arduino Nano to the gate of a FQP30N06L MOSFET and the circuit worked
perfectly (as far as I could tell).

However, I now need to run 4 heating elements from a 24.4v / 15A power supply. When we tried this the MOSFET overheated (to the extent that the solder was almost melting) but curiously the heater appeared to be ‘stuck on’ at all times and the PWM signals from the Arduino seemed to have no effect.

I’ve recently learnt about RDS (on), which at 5v VGS predicts the casing temperature to be 517 degrees Celsius….

Curiously, the module which I’m using suggests connecting the positive terminal of the power supply to the source and the negative to the drain meaning that at the moment the VGS is around 20v and so very close to its maximum, I would have thought swapping the source and drain inputs would mean it’s much closer to being in spec, but would maybe result in more heat being produced?

Also, I’ve toyed with the idea of adding more MOSFETs in parallel, would this reduce the heat significantly? I also know that I might run into issues with natural capacitance in MOSFETS if using them in parallel.

I’ve had a search of the forums and my problems seem very similar to this post by RED_ Heater, MOSFET and PWM [Solved] - Motors, Mechanics, Power and CNC - Arduino Forum but I’m not sure if the same solutions apply at the higher voltage and current values.

Please let me know of any improvements / solutions, I’m very new to using MOSFETs in general as I’m sure you can tell

Any help is much appreciated, many thanks in advance!

PS I've attached a circuit diagram with the temperature sensors omitted.

[slipstick]

What is the module you refer to? Are you using a plain MOSFET or some sort of MOSFET module? Details please. An FQP30N06L is an N-channel MOSFET and normally Source would connect to GND and Drain to the load. You would also have a resistor between pin and gate. If you have connected it the other way round it's probably dead by now.

Steve

[MarkT]

The worst-case on-resistance for the FQP30N06L at Vgs=5V is 45 milliohms, which is a lot
these days (10 milliohms is more usual for a low voltage MOSFET).

Find a MOSFET with 5 milliohms or so on resistance, and give it a small heatsink,
or find one with 10 miiliohms or so and give it a medium sized heatsink,
or find two with 10 milliohms and parallel them.

You either need logic-level MOSFETs, or you'll have to use a MOSFET gate driver chip running from
12V to handle the MOSFET(s). Gate driver chips take logic inputs and drive standard MOSFETs at 12V
with lots of oomph, essential if using high frequency PWM.

Dissipation goes up as the square of current for a MOSFET, so upping the load by a factor of 4
means the MOSFET dissipates 16 times more power - which is what bit you. For very
high current loads its normal to go for large packages with heatsink, you can even get packages
with screw terminals for 100W dissipations or so.

[v15va]

slipstick:
What is the module you refer to? Are you using a plain MOSFET or some sort of MOSFET module? Details please. An FQP30N06L is an N-channel MOSFET and normally Source would connect to GND and Drain to the load. You would also have a resistor between pin and gate. If you have connected it the other way round it's probably dead by now.

Steve

Hi Steve, thanks for the fast reply!

So I'm not sure on the exact model number since it was given to me by a colleague having been salvaged from another circuit (although it is safe to assume that it was working when given to me). I have however attached some pictures of link below it in its original circuit if that is of any help, it appears to have a resistor between the Gate and Source pins, but its markings also suggest attaching GND to the Drain pin and the load to the Source pin which is slightly confusing.

It is probably worth noting that when I connected just the 1 heating element I followed the markings on the module (GND to drain, load to source) and the MOSFET appeared to behave normally, no excessive heating or signs of damage. Do you think this means that the negative marking is in fact routed to the Drain pin? This would mean that the resistor is between the source and gate pins, would that make any sense?

Edit - Photos were too large, so can be found here https://imgur.com/a/ZpkiorV

[v15va]

MarkT:
The worst-case on-resistance for the FQP30N06L at Vgs=5V is 45 milliohms, which is a lot
these days (10 milliohms is more usual for a low voltage MOSFET).

Find a MOSFET with 5 milliohms or so on resistance, and give it a small heatsink,
or find one with 10 miiliohms or so and give it a medium sized heatsink,
or find two with 10 milliohms and parallel them.

You either need logic-level MOSFETs, or you'll have to use a MOSFET gate driver chip running from
12V to handle the MOSFET(s). Gate driver chips take logic inputs and drive standard MOSFETs at 12V
with lots of oomph, essential if using high frequency PWM.

Dissipation goes up as the square of current for a MOSFET, so upping the load by a factor of 4
means the MOSFET dissipates 16 times more power - which is what bit you. For very
high current loads its normal to go for large packages with heatsink, you can even get packages
with screw terminals for 100W dissipations or so.

Hey, thanks for the fast reply!

So I've done a bit of digging and I've found this module: High Power Mosfet Drive Module PWM Control AOD4184A | eBay

(picture attached if the link's dead)

It uses 2x AOD4184A logic level MOSFETS in parallel with a max Rds of 9.5 mOhms (@4.5v). The MOSFETs are flat on the board but I think I should be able to bend them up and attach a heatsink.

So I understand that reducing the PWM frequency would mean that the MOSFET has to 'switch' between its on and off states less per unit time, and it's in this switching state that it produces the most heat. So, if I was able to reduce the PWM frequency, would that further reduce the heat output?