Solutions for current balancing between MOSFETs in parallel

Question

I am designing a programmable DC load using 120W MOSFETs available at my local store. Due to its limited power rating, I am planning to have 2-3 of them connected in parallel as shown in schematics below to achieve a minimum of 300 watts.

In the LTSpice schematics attached, I simply hooked them up in parallel to the input ports (24DC) and the simulation results look okay however after doing some research online and from some great comments to my past queries on this forum I've come to realise that MOSFET gate threshold differes even between the same models due to varying temperatures. I looked up for current balancing solution to ensure Ids on all MOSFETS is always the same but the only solution I've been able to find by far are some exotic current balancing modules from nexperia. I was wondering if I can achieve the same using basic passive/active components?

I am thinking of attached a termistor directly to chassis of each MOSFET and use Arduino to calibrate Vref for each MOSFET but I'm afraid the thermal feedback will be too slow and MOSFETs will smoke well before Arduino has change to adjust the Vref.

Edit 1:

Additional information :

1. The input reference V1 and V5 are analog inputs coming from same analog output pin of an arduino to set the current drain. Theoratically and even in the simulation, the single reference should keep all three MOSFETs balanced however other variables like junction temperature of each mosfet will change the gate voltage threshold leading to large current flowing from MOSFET/s running cooler then the rest.

2. Although the schematic only shows 2 MOSFETs, I am looking for 3 MOSFETS to run in parallel with 5A max each making total Amp rating 15A. I am trying to understand how we can ensure the current through each one of these MOSFETs is equal at all time irrespective of their junction temperatures.

MOSFET Datasheet - https://www.farnell.com/datasheets/67691.pdf

Answer 1

When you operate MOSFETs in their linear region you have no choice but to control each one individually. Then, it boils down to analysing each MOSFET and its ability to deal with a peak of 5 amps at some undisclosed supply voltage for some undisclosed length of time. But the data sheet can tell certain things even though the MOSFET is clearly not designed for linear applications. For instance, the safe operating area from the data sheet: -

The above is the safe operating area for the MOSFET and note that it only gives you information for "on-times" up to 10 ms so, if you want to handle 5 amps for longer periods you have to either (a) choose a MOSFET that is capable of operating in linear applications (there are a few) or (b), make a few guesses like what I've added in colour: -

• At 5 amps for up to 100 ms, the maximum voltage across the MOSFET is limited to about 14 volts
• At 5 amps for up to 1 second, the maximum voltage across the MOSFET is limited to about 5 volts
• For up to 10 seconds you might be able to run with 1.5 volts across the MOSFET

And, if you do the power calculations for each scenario you get 70 watts, 25 watts and 7.5 watts i.e. even though the MOSFET is rated for a maximum power handling of 140 watts, you get nothing like that in the linear current-control region.

So, does the MOSFET do what you want? Only you can say.

The reason why it can't handle linear region power is because the MOSFET will suffer from extreme thermal runaway. Here's a modified version of figure 3 in the data sheet: -

• If \$V_{GS}\$ is below 5.2 volts (the zero temperature coefficient point), as the MOSFET warms it draws more drain current. You should be able to imagine where this ends and, believe me, it's rapid (hundreds of micro-seconds to a few milliseconds for some MOSFET constructions)
• If \$V_{GS}\$ is above 5.2 volts, the MOSFET will shut-down as it warms and, protect itself.

Ask yourself what likely value of \$V_{GS}\$ you will have for 5 amps of drain current: -