[SOLVED] Simple Mosfet switch gets hot

[zedman2]

I would like to replace the PWM switch (which is failed) in a DeWalt screwdriver with a simple on/off switch using a MOSFET to switch the high Apms.

I made this simple circuit, it does what it should but when runs in 1-2 seconds the FET gets really hot.
It runs from a 20V Li-Ion battery pack, the motor draws peak 20 A when starts and ~3 A when runs, FET rated at 20A (continous).
Ugs is around 17V so I think FET should be open.
What am I missing here?

 

[Easy peasy]

Rds-on of the fet, power = I^2R, R goes up with temp ....

 

[zedman2]

Rds-on of the fet, power = I^2R, R goes up with temp ....

Rds on is 0,077 ohm, 3x3x0,077=0,693W should it be this hot? In 2 secs i can't touch it...

 

[andre_luis]

Although the motor is called 'DC', the current flowing through it is far from being just continuous. The brushes generate a significant amount of spikes that must be accounted for; I would consider adding ceramic capacitors in parallel with the motor, in addition to snubber in parallel with the MOSFET. But to be honest, in applications like this I would start with a relay.

 

[KlausST]

Hi,

a datasheet is full of data.

You mention the 20A rating, but did you also see the 100°C condition? It does not mean the MOSFET gets 100°C hot, instead it means that you - as the designer - are responsible to keep the temperature below 100°C. Usually with the use of a heatsink.

Klaus

 

[zedman2]

Thanks,

I saw the 100 C rating but as I understand if I keep the temp under 100 C than 20 A could continously go through, and below 25 C 28 A.

So I would not expect this amount of heat at 3A.

 

[KlausST]

Hi,

how did you measure the 3A?

0.7W will make a transistor - without heatsink - hot enough that you can´t touch it.
With an rth-ja of 62K/W you get about 43K temperature rise. So if you start at 25°C ambient temperture you may expect 68°C.
But indeed I don´t expect it to become that hot within 2s.
Maybe it´s caused by the 20A starting current. Which causes about 45 times the heat as 3A

Usually the DC current os measured with a DC Ampere meter. A Dc Amperemeter usually shows the average value.
But for power dissipation calculation one needs to use the RMS current.

For clean DC the average value and the RMS value is identical. But as soon as there are variation, spikes, pulses... then the RMS value will be higher than the average value. A motor usually causes some amount of pulses.

Klaus

 

[danadakk]

You double checked your gate drive to be close to 16V when on ?

What is the Vds when its hot ?

Is it a genuine Vishay, ST, or IRF part, not a Chinese knockoff ?

Thetaj-a is 62 C /W so its around 42C, 107F, should be fine.



Regards, Dana.

 

[Easy peasy]

You do not really know the currents flowing in the motor - this is the basic problem here.

Once the startup peak currents heat the fet - it gets hot and its Rds-on is much higher straight away - this affects the dissipation with just the supposed 3A, i.e. can easily be 1.5 times higher ...

a TO-220 with no heatsink is good for only about 0.5 watt before thermal runaway gets it...

so your 0.7 watt AFTER startup - will likely cause the temp rise you are seeing - i.e the somewhat undefined " really hot "

put some sort of heatsink on it.

 

[KlausST]

Hi,

put some sort of heatsink on it.

or use a MOSFET with less R_ds_on

Klaus

 

[betwixt]

I concur with Klaus, it doesn't really matter that it gets hot, it always will and you can't cheat Ohms Law. The IRF540 has a high Rds and that in turn means high power dissipation. Use a MOSFET with much lower Rds and you should be OK. IRFZ44 comes to mind, Rds(on) is much lower and they are readily available.

Brian.

 

[zedman2]

Thanks guys,

I tried to put on 100nF + 100 ohm snubber, nothing changed.

That is true I can't measure the exact current, I used a multimeter.
I examined the Voltage on the motor on scope, it goes from 15 to 18V as load changes (motor runs) so the current does the same I think, so the 3A is just a raw average.

I found a STP60NF06 in the drawer which I salvaged from another screwdriver I tought I'll give it a go.
It's 60A / 240A pulsed, with Rds 0,016 Ohm.

TADA: it works like charm.

I noticed the motor runs faster instantly, it does get a bit warm after 10-20 sec but way much better.
Since it primarily used for screw in and stop like movements, it will do.

Could the Rds 0,016 vs 0,044 (IRF540) count this much?

 

[KlausST]

Hi,

Could the Rds 0,016 vs 0,044 (IRF540) count this much?

about three times the resistance = 3 times higher dissipated power = 3 timer higher temperature rise.

25°C + 43°C = 68°C you can´t stand that heat with your finger.
25°C +14°C = 39°C is about your body temperature.

It´s simple physics and simple math. I´m sure you could have answered the question on your own. ;-)

Klaus

 

[zedman2]

Sometimes it's really hard to accept what math / physics say is in reality, in the real world, is a matter of touch or not to touch...

I think I'll go for IRFB7440PBF with 2,5 mOhm.

thanks everyone,
Zedman