Making sense of Tiny SMD mosfets with high current ratings

Hi. Trying to make sense of this.
How can mosfets such as this one BUK9M10-30EX - Nexperia USA Inc.
be so small yet can handle some 10, 20, or even (in this case) 54 AMPS CONTINUOUS! These packages as well as the leads are absolutely TINY!
Moreover, there is not enough space to make traces thick/wide enough to allow that kind of current.
Heck, even if you add up all of the SOURCE or DRAIN pads, the PCB trace width (assuming typical 2oz or 1oz thickness) will not even come close to being able to handle that amount. In this case if you make a trace as wide as all of the pads you MAYBE will be able to safely get 2-3A max.

Can someone please shed some light here.


Notice the under side of the MOSFET.

It has a very large DRAIN surface where heat is transferred out from and is connected to a 'heat sink pad' for example.

Also the ON resistance is .010 ohms, therefore there is little voltage dropped across it.

W = V*V /R = V * I

I am not question the capability of the MOSFET itself regarding heat dissipation. Rather the cumulative trace width of the SOURCE. Or even the entire DRAIN is not wide enough. If there would be no MOSFET, just a plain PCB the traces of this thickness itself would be underrated for this type of current.
A trace width calculator will tell you that for even 10A @ 2oz you would need at least 0.142" of trace width. THAT IS WIDER THAN THE ENTIRE MOSFET!

IPC-2221 Standard

Your worrying about the PCB design not the MOSFET ?

When you design the PCB for this chip you would make sure the wires connected to the DRAIN and SOURCE were very close to the device leads, perhaps within 5mm.

There are 3 SOURCE leads, lets say they a #24AWG and lets say they have a total length of 4 mm.
This would be the same as ~24AWG at 1mm (0.00328084 feet).
24AWG copper is 25 ohms per 1000feet and we have 0.00328084 feet . . .
You can do the current carrying capacity of the 3 source leads.

90% likely you would be using these in a circuit < 10 amps anyway.

How can they rate a component that is practically impossible to rig to live up to its spec

They do it all the time. :wink:

This is an SMD meaning there should be a way to have traces wide enough to handle more than just 2-3 amps

Drawing something up

The most you would be able to get from this 54A MOSFET is 5.525A!
Now, does that make sense?
I see the same issue with almost every MOSFET of these tiny sizes.
Am I getting something wrong?

Yeah.. I know. I do this all the time with through hole stuff. Very difficult to thicken unmasked traces next to SMDs.

I guess lesson learnt.?. specs do not necessarily reflect reality. ... ok
Just wanted to make sure that I am not missing anything.. Am I?

It is quite common that a spec does not necessarily reflect a realistic usage situation and some maximum ratings are intended as a guide only.
Take AMS1117, a simple linear voltage regulator. The spec says max power 1Amp. The 3.3 version says max voltage 15volts.
Does that mean you draw 1 Amp through it with a 15v input?
Possibly, if you have a big enough heat sink to stop package melting while dissipating 12 watts.

Also consider that all the traces do not have to be on the same board layer.

If you look at the chart in the datasheet, 54A continuous is only at slightly over 1V. At 10V you're below 3A continuous. At 30V you're at about 0.5A continuous.

On the other hand, if you don't need a continuous current, you can do over 200A for 10 microseconds (3-11V)

The overall thermal considerations are far more important than just current and trace widths. Nexperia has a lot of useful and necessary info, you just have to dig a bit to find it.

Understanding power mosfet datasheet parameters

LFPAK MOSFET thermal design guide

Well they do reflect reality but you have to read all of them to get the full picture. Just like in the news papers if you just read the headline you will get a distorted view of the story. Conditions can be contrived to maximise any one parameter, when you dig int to the full story you will see that this parameter is rarely going to be the limiting factor, there are other factors that come into play that define the limits of any one design. So the headlines are true but just like politicians they lie by using the truth.

1 Like

You are assuming a 2 oz PCB I don't think a 50 amp design would be designed at 2 oz.

Getting back to the apparently unrealistic specifications. In general you will find the specifications provide a maximum allowable value. In this case, ignoring the PCB, the limit will be getting rid of the heat, not the 50A max current.

As stated above, the specifications of nearly all semiconductor based devices "headline" specifications at the very limit of capability. Now since all semiconductor companies do the same thing, it makes a rough comparison relatively easy.

To be honest I raised my eyebrow at the spec's of this part. I finally considered the max currents to be useful for transient situations that are much longer than the thermal time constants of the device. Perhaps a motor startup current under load.

But this is next level misinformation. It says 54A continuous.
Also for get about pcb traces. The size of the source leads combined are not thick/large enough for 54A continuous.

Power MOSFETs are vertical current flow, so that the whole width of the chip is available to carry current.

Silicon when doped is capable of low resistivities, for instance doped at 10^20 atoms per cubic cm, silicon has a resistivity of about 10^-5 ohm-m, only 60 times that of pure copper.

And then the current is flowing though a chip that's perhaps 3mm x 3mm and 100µm thick, so theoretically you could have a resistance as low as 100µohms in that chip if made of
heavily doped silicon - its not hard to see that a resistance of a few milliohms in a real device
is entirely plausible.

And of course the bond-wires have to carry large currents too, but they just use lots of
thick bondwires for high currents, each heavily heatsinked (by virtue of being short and welded to the heat-tab).

The device in question has an on-resistance of 8mohm, so 54A dissipates about 23W, which is only
possibly with appropriate heatsinking - the datasheet is only talking about the raw device,
not the combination of device and PCB. For most uses of the device you'd not be expecting
anything remotely like 54A (5A might be a typical value if you want to use the device without a
large copper block heatsink!).

Any then the other issue to consider is that manufacturers want their device to look as good as
possible. No good engineer would be taken in by these values - in practice this device is not
usable at the max continuous rating, its simply far too expensive to get the heat out of the chip
when its this small, you'd choose a much better package if you needed to lose 23W continuously.

In short just ignore current specs for a MOSFET, the on-resistance is what matters in all practical uses!