I did a search and there are lots of other threads/discussions about mosfets not turning off. I did not find my answer though.
I am using an arduino to control a bank of heating coils in an automotive application. I searched for an appropriate mosfet and found the IRFZ44n with a threshold voltage of 2 - 4v so thought it was appropriate (the arduino outputs are 5v). I have a 10k pull-down resistor on each gate so no floating outputs. I works perfect on the bench with a small load but once installed and suppling 4 amp loads (14.5vdc) some of the mosfets will not turn off. I am just using the bang bang method of switching, not PWM so no high speed switching.
Do I have the wrong type of mosfet? My understanding was that if the gate supply was not high enough, the mosfet would not turn on rather than not turn off. Any idea of what I am missing?

Thanks!

 

found the IRFZ44n with a threshold voltage of 2 - 4v so thought it was appropriate (the arduino outputs are 5v). I have a 10k pull-down resistor on each gate so no floating outputs. I works perfect on the bench with a small load but once installed and suppling 4 amp loads (14.5vdc) some of the mosfets will not turn off.

5V is not enough to fully turn on a MOSFET with a Vth of 2 - 4V.
You need to look at the Vgs used to specify the ON resistance, which is 10V for standard MOsFETs (see below), so likely the MOSFETs overheated and shorted.



You need a logic-level MOSFET with a max Vth of 2V or so.

 

Replace your IRFZ44 with an IRLZ44.

Bob

 

I did a search and there are lots of other threads/discussions about mosfets not turning off. I did not find my answer though.
I am using an arduino to control a bank of heating coils in an automotive application. I searched for an appropriate mosfet and found the IRFZ44n with a threshold voltage of 2 - 4v so thought it was appropriate (the arduino outputs are 5v). I have a 10k pull-down resistor on each gate so no floating outputs. I works perfect on the bench with a small load but once installed and suppling 4 amp loads (14.5vdc) some of the mosfets will not turn off. I am just using the bang bang method of switching, not PWM so no high speed switching.
Do I have the wrong type of mosfet? My understanding was that if the gate supply was not high enough, the mosfet would not turn on rather than not turn off. Any idea of what I am missing?

Thanks!

With static outputs; it should be easy to measure how low the output takes the gate. Some logic devices don't quite cut it when interfaced to a MOSFET gate.

You could re arrange the software to load the output registers with inverse levels. Drive a bipolar transistor that shunts the MOSFET gate, the collector will require a pull up resistor - if you take that to a 12V rail, there's no need for logic level MOSFETs.

 

5V is not enough to fully turn on a MOSFET with a Vth of 2 - 4V.
You need to look at the Vgs used to specify the ON resistance, which is 10V for standard MOsFETs (see below), so likely the MOSFETs overheated and shorted.

View attachment 140107

You need a logic-level MOSFET with a max Vth of 2V or so.

I understand now that to reliably turn ON the mosfets for full current I need the logic level version. But that does not answer the question of why they are not turning OFF with the higher current. Once they are on (conducting), if I pull the gate to ground, should they not turn Off no matter what voltage they were turned On with?

 

With static outputs; it should be easy to measure how low the output takes the gate. Some logic devices don't quite cut it when interfaced to a MOSFET gate.

You could re arrange the software to load the output registers with inverse levels. Drive a bipolar transistor that shunts the MOSFET gate, the collector will require a pull up resistor - if you take that to a 12V rail, there's no need for logic level MOSFETs.

So what you are saying is the arduino can not sink enough current to pull the gate to ground to turn it OFF?

 

I pull the gate to ground, should they not turn Off no matter what voltage they were turned On with?

Not if they are zapped due to overheating.
If they are not fully on, the high Vds times the ON current can generate enough heat to damage the MOSFETs.
Or they make get so hot, they go into thermal runaway where they will stay on independent of the gate voltage.
Do those MOSFETs work okay again after this happens, with a smaller load?

 

Not if they are zapped due to overheating.
If they are not fully on, the high Vds times the ON current can generate enough heat to damage the MOSFETs.
Or they make get so hot, they go into thermal runaway where they will stay on independent of the gate voltage.
Do those MOSFETs work okay again after this happens, with a smaller load?

Yes, they work fine when I put it back on the bench and test with smaller load. On and Off with no problem. So the mosfets are not damaged.

 

Yes, they work fine when I put it back on the bench and test with smaller load. On and Off with no problem. So the mosfets are not damaged.

Then it may be thermal runaway.
Have you touched them to see how hot they are when carrying the 4A load?

 

I understand now that to reliably turn ON the mosfets for full current I need the logic level version. But that does not answer the question of why they are not turning OFF with the higher current. Once they are on (conducting), if I pull the gate to ground, should they not turn Off no matter what voltage they were turned On with?

You probably need to analyse the symptoms a bit more.

If the MOSFETs are failing under full load - that should'nt be difficult to misinterpret.

If V-OUTlow isn't low enough to turn the MOSFET off, it will still be carrying current but probably also dropping some of the supply voltage - that should warm up or even get hot.

As someone else pointed out - if the current is too much and overheats the MOSFET, it'll carry on conducting when you turn off the gate voltage.