MOSFET power rating

[Techman_7]

Hello,
For SiC MOSFET NVH4L020N120SC1, below is the datahsheet snap shot for maximum ratings.
Can anybody let me know how is this possible, 500W loss at 101A current?

Also, HOW Tj = Tc = 25 degree C ?

 

[KlausST]

Hi,

500W = 100A × 5V or 50A × 10A or 5A × 100V or 1A × 500V..... there are infinite possible configurations.

Maybe you are fixed on R_ds_on ... which btw. is not shown in the snippet above.
But no one says the MOSFET is 100% of the time FULLY ON.

There is loss at linear operation,
There are switching losses (maybe covered by the above)
There is conduction loss
There is body diode loss

Klaus

 

[Techman_7]

Hi,

500W = 100A × 5V or 50A × 10A or 5A × 100V or 1A × 500V..... there are infinite possible configurations.

Maybe you are fixed on R_ds_on ... which btw. is not shown in the snippet above.
But no one says the MOSFET is 100% of the time FULLY ON.

There is loss at linear operation,
There are switching losses (maybe covered by the above)
There is conduction loss
There is body diode loss

Klaus

Its mentioned 101A continuous current, so there is no swithcing loss, only conduction loss = 500W.
So, if RdsON = 20 mohm, we can not get 500W.

Also how Tj = Tc = 25 deg C ?

 

[danadakk]

Also how Tj = Tc = 25 deg C ?

Contact vendor and let them know table has errors.

Not sure what this statement means :

Clue here :

Regards, Dana.

 

[Techman_7]

How can we prove 500W conduction loss by : Id^2 x RdsON ?

 

[danadakk]

So when its hot 101A^2 x .05 = 510W......?


Regards, Dana.

 

[FvM]

Its mentioned 101A continuous current, so there is no swithcing loss, only conduction loss = 500W.
So, if RdsON = 20 mohm, we can not get 500W.

Do you see somewhere in the datasheet that Pd=500 W describes fully on state? It could be e.g. constant current at Vds = 5V. You are jumping into conclusion.

 

[danadakk]

Do you see somewhere in the datasheet that Pd=500 W describes fully on state? It could be e.g. constant current at Vds = 5V. You are jumping into conclusion.

We have the following :

And we have :

Not at 101 A, but we have :

and :

And we have Rdson varies with Id. .3 C/W x 500 W = 150 C rise over 25C = 175C

Looking at Imax allowed in Drain, and worst case Rdson = 50 mohm

So the computation fits ....? I think problem here is if they give us 500W in datasheet,
but then futz around with specs at 60A, makes conclusion hard(er).


Regards.....Dana.

 

[KlausST]

Hi,

You are mixing two independent specifications:
* current rating 100A
* power dissipation of 500W
Again: they are independent.

No where is mentioned that the power rating is for 20mOhms. No where it says it is at full V_GS.

Let's say you have a V_DS of 100V, then you may regulate V_GS to get an R_DS of 20 Ohms...resulting in a current of 5A and thus resulting in a power dissipation of 5A x 100V = 500W. Steady state = DC= 100% duty cycle = no oscillation.
As already written there are infinite conditions of V and I to get 500W of power dissipation.

Or 100A, 50mOhms = 5V = 500W.

Klaus

 

[barry]

MOSFETS aren't only operated as switches...

 

[danadakk]

Kalus that data sheet is a mess.

No where is mentioned that the power rating is for 20mOhms. No where it says it is at full V_GS.

Its "sorta" implied here :

I was trying to back into how the 500W came about, which I think OP was looking for.
The calc using 101 A, and worst case Rdson of ~ 50 m ohms.....but I have to confess I
think I broke one of my cardinal rules, dont use "sorta", I got out over my skis a little
too far on this one.

 

[barry]

Kalus that data sheet is a mess.



Its "sorta" implied here :

View attachment 184987

I was trying to back into how the 500W came about, which I think OP was looking for.
The calc using 101 A, and worst case Rdson of ~ 50 m ohms.....but I have to confess I
think I broke one of my cardinal rules, dont use "sorta", I got out over my skis a little
too far on this one.

There’s no implication there. That simply states the recommended range of Vgs. There’s no way to infer that this means Rds is minimum, not even sorta.

 

[danadakk]

Like I previously said

but I have to confess I
think I broke one of my cardinal rules, dont use "sorta", I got out over my skis a little
too far on this one.

But its interesting the Vgs = 20V conditions discussed here :

The data sheet is poorly done, consistency, . I repeat.

 

[barry]

Im afraid I don’t see any inconsistency here. They specify Rds at a specific Vgs, not ALL Vgs. They specify a recommended range of Vgs. They specify a threshold voltage for a specific drain current. Where do you see inconsistency?

 

[KlausST]

The data sheet is poorly done, consistency, . I repeat.

I can't remember any Mosfet datasheet that states V_GS or R_ds_on for power dissipation rating.
I can not see that this datasheet is different in this.

***
I don't deny that there are Mosfets where max_power_dissipation limits the max_current even when the Mosfet is fully ON, but there are many other Mosfets where this is not the case.

*****
For Mosfets with low R_ds_on it's not uncommon that I x I x R_ds_on does not match max allowed power dissipation.
This is, because the chip is big and when fully ON the power is not dissipated equally over the whole chip (or R_th_jc is not equal over the whole chip). There will be hot spots.
This is why we have an SOA graph where the rising line (R_ds_on_limit) does not meet the falling part (power_dissipation_limit), there is a flat horizontal part inbetween.

The datasheet is consistent in my eyes.

Klaus

 

[FvM]

The basic misunderstanding is about meaning of maximum ratings. The requirement is that each limit has to be kept absolutely. There's no correlation between individual values implied, except that no limit should be exceeded. Regarding discussed Pd,max spec, there's no Id, Vds or Vgs operation point implied, just steady state. SOA limits have to be regarded in addition.

 

[danadakk]

Interesting discussion on hot spots, Spirito effect, and region of concern on SOA :

https://www.analog.com/en/technical-articles/mosfet-safe-operating-area-and-hot-swap-circuits.html

 

[Easy peasy]

500W loss at 101A is entirely possible - however your device may not last too long under these conditions - even if you can keep the tap to 25 degC (which is the condition for 500W dissipation )

At 101 amps and 30m-ohm say, with the device hard on, the dissipation is 306W - you would still need considerable cooling . . .

thermal design is not trivial.

--- Updated Sep 20, 2023 ---

note from the data sheet, at 100 deg C case ( i.e decent heatsink ) the device can carry 71.4 A DC

and / or dissipate 250W max