According to the fig 4 of the PSMN1R8-40YLC datasheet the max safe Id at 10v should be about 6amps when used as a switch that is always on correct or am I misreading this and that is not what fig 4 is for?

 

Yes, but for continuous operation you need to do thermal design to meet
Tjmax of the part.


Regards, Dana.

 

No, Safe operating area curves show the combination of drain to source voltage and drain (same as source) current that may be safely allowed simultaneously. The point you refer to means you can have 10 volts across the FET at the same time you have 6 amps through it, and that combination can exist continuously ("DC"). It does mean there will be 60 watts of power dissipation (which will melt the thing off the PCB unless great care is taken in thermal management, which is VERY difficult for such a small package). When you use the FET as a switch and it is ON, the drain-source voltage is the product of the current and the ON resistance. The datasheet spec's a maximum ON resistance of 2.1 milliohms with 4.5 volts gate to source, at 25°C die temperature. A rough rule of thumb is to multiply that resistance by about 1.3 for higher temperature unless the datasheet has good spec's for high temp resistance. Often you can go with the maximum current rating, which is 100 amperes (Table 4). Once again care is require with regard to temperature rise. Taking the ON resistance as 2.1 x 1.3 milliohms and multiplying by current squared (same as for power in any resistor), you get about 27 W. That is still quite a bit for such a small surface mount package. At 50 A the dissipation would be about 7 W, which is more manageable. At 10 A the dissipation would be less than 0.3W, which is next to nothing. When the switch is OFF, the current is essentially zero, so the power loss is also zero. See Figure 1 in the datasheet.

Usually for small SMT parts like this it is useful to determine what you can do in terms of heat dissipation first and then use allowable maximum die temperature you or the spec's will allow to determine the maximum allowable current when used as a switch. Almost always that will dictate using the device at well below the spec'd maximum current. If the device is used for high speed switching both conduction losses and switching losses must be considered. The latter can often be greater than the former, though with low supply it is usually pretty small if the switching transition is fast. I generally largely ignore the current rating for FETs and select based on ON resistance.

According to the datasheet you can run that device with 272 W of dissipation with the mounting base temperature at 25°C. To keep the mounting base at that temperature you would have to mount it on a great big pure silver (or diamond) heatsink or run it in liquid nitrogen (OK, a bit of an exaggeration, but in practical terms it is almost impossible with practical methods).

 

An excellent treatment of MOSFET SOA -

https://www.google.com/url?sa=t&rct...3646e9381200&usg=AOvVaw2239I26VQTUdEzNMS_-Qaz


Regards, Dana.

 

No, Safe operating area curves show the combination of drain to source voltage and drain (same as source) current that may be safely allowed simultaneously. The point you refer to means you can have 10 volts across the FET at the same time you have 6 amps through it, and that combination can exist continuously ("DC"). It does mean there will be 60 watts of power dissipation (which will melt the thing off the PCB unless great care is taken in thermal management, which is VERY difficult for such a small package). When you use the FET as a switch and it is ON, the drain-source voltage is the product of the current and the ON resistance. The datasheet spec's a maximum ON resistance of 2.1 milliohms with 4.5 volts gate to source, at 25°C die temperature. A rough rule of thumb is to multiply that resistance by about 1.3 for higher temperature unless the datasheet has good spec's for high temp resistance. Often you can go with the maximum current rating, which is 100 amperes (Table 4). Once again care is require with regard to temperature rise. Taking the ON resistance as 2.1 x 1.3 milliohms and multiplying by current squared (same as for power in any resistor), you get about 27 W. That is still quite a bit for such a small surface mount package. At 50 A the dissipation would be about 7 W, which is more manageable. At 10 A the dissipation would be less than 0.3W, which is next to nothing. When the switch is OFF, the current is essentially zero, so the power loss is also zero. See Figure 1 in the datasheet.

Usually for small SMT parts like this it is useful to determine what you can do in terms of heat dissipation first and then use allowable maximum die temperature you or the spec's will allow to determine the maximum allowable current when used as a switch. Almost always that will dictate using the device at well below the spec'd maximum current. If the device is used for high speed switching both conduction losses and switching losses must be considered. The latter can often be greater than the former, though with low supply it is usually pretty small if the switching transition is fast. I generally largely ignore the current rating for FETs and select based on ON resistance.

According to the datasheet you can run that device with 272 W of dissipation with the mounting base temperature at 25°C. To keep the mounting base at that temperature you would have to mount it on a great big pure silver (or diamond) heatsink or run it in liquid nitrogen (OK, a bit of an exaggeration, but in practical terms it is almost impossible with practical methods).

If I have 10v 6amp load that I am switching is there not that going across the fet at the same time or am I forgetting something like the Vds is only a small portion and the rest is across the load as voltage divides in series?

 

In general you have to make sure SOA is OK. But your last post seems to state
you just want a switch, the source is 10V, the load will draw 6A, so look for a
low Rdson device to do this. Of course if you have high frequency switching
then you need to consider power used in Rdon as well as gate drive power in
your thermal design calculations and drive circuit design, and its calculations.

If a low freq switch, then most times you want source to deliver most of its energy
to load, eg. minimize switch losses. Which in low freq case primarily I^2 x Rdson.

Regards, Dana.

 

Thank you, I figured out where I was getting confused. Not sure why but in my head I completely missed that the voltage is the drain-source voltage and not the supply voltage so now the number are making more sense.

Thanks for the help.