MOSETS: Safe operating area - inconsistencies

[supergordon]

When I look on the datasheet for the Alpha&Omega AO4410 N-Channel MOSFET in the safe operating area graph, it tells me the SOA eg.:
at 10Vds only about 150mA continously
at 5Vds about 500mA continuosly
at 1Vds about 2500mA continously
at 0.15Vds about 18A

1.) Why is this MOSFET rated for 18A?
2.) Why can I use it as a load switch with a drain current of more than 5A and the Drain-Source voltage is 12V or 19V with the mosfet (SOP8) only slightly gets warm and is stable - the mosfet doesn't blow up or fail?

I even had an sot23 mosfet from AO that was rated for 5A and had a simular SOA but continously 5A were no problem at 19V. The mosfets do work fine today after many many years of operating.

So what is the SOA in the datasheets exactly telling us?

Sorry for the questions, but I don't get it.

[sigma_xi]

Hi,
I don't see any inconsistencies, but you have to be careful about the actual Vds/Id values present at the same time.

First of all, Note A in the attached data sheet says that the absolute maximum ratings are based on the thermal resistance rating for t<=10s, which is at max 40°C/W. I admit that it is odd to specify the maximum continuous current via this transient thermal resistance, but I assume they have valid reasons for that. Maybe someone else has more insight and can provide an explanation.

1) When the MOSFET is fully on, its Rds is around 6mOhm. Given Id=18A the total power dissipation is around 2W. Compare this to the asymptotic value of the Power rating curve and you will find that their the same. Furthermore, given the Rth=40°C/W, the junction is P*Rth=80°C warmer than the ambient temperature (25°C), so the junction is at roughly 105°C. Hence, in order to stay below a certain junction temperature (in this case around 105°C) the current must not exceed 18A at Ta=25°C, which is why the manufacturer specifies this current as the absolute maximum rating.
 
At the same time, Rds=6mOhm results in a Vds =Id*Rds= 0.11V, which is exactly what you can see in the SOA graph. So no inconsistencies there.

2) When switching a load via a MOSFET, Vds=Vsupply when the FET is off, and Vds is approximately zero when the FET is on. Keep in mind that you want the voltage to drop across your load and not across your FET. Thus, the 19V / 5A are never present at the same time. (If they were, your FET would dissipate around 100W and your SOP8 case will probably explode within a second). Thus, in the off-state you have Vds=19V Id=0A, and in the on-state roughly Vds=0.1V Id=5A, that's why the MOSFET doesn't blow up.

And, to answer your final question: The SOA tells you exactly what its name implies: The maximum safe Vds/Id combinations for different pulse-lengths. The Vds in this case is the actual voltage in the on-state. Therefore, if you really have Vds>15V in the on-state, the maximum safe Id is below 100mA.

I hope this helps in resolving your confusion.

[Wolfram]

The SOA curve concerns Vds and Id applied at the same time. When used as a switch, Id is very small (leakage, a few microamps typically) when the device is turned off, and Vds is very small (Rdson*Iload) when the device is turned on. The SOA curve is mostly of concern during the transition between on and off, and if the device is operated in the linear mode.

[supergordon]

Thanks sigma_xi for your detailed answer. 
I got it now, you helped me understanding where I was confused about. Vd != Vds. Now the SOA graph makes completely sense.

Thanks for explaining all, also thanks to Wolfram. 

I never thought I would get so fast an good answer back and I originally did not know in what forum to post - so I remembered Dave and the  eevblog youtube channel that I love and thought to give a try in this forum. 

[sigma_xi]

You're welcome

[larsdenmark]

Scullcom has a very nice explanation here:


Note that he is making a DC load so the interesting operating region is where the MOSFET works as a resistor (and gets hot). Using the MOSFET as a switch is different.