4
\$\begingroup\$

I am picking the low and high side MOSFET switches for a 2 amp output buck converter, simplified schematic shown here. Vin = 20 V and Vout = 15 V.

enter image description here

The IC manufacturer has a spreadsheet for calculating various components, and for my design specs the spreadsheet recommends MOSFETs with continuous current >=2 A (low side) and >=3 A (high side), and min power dissipation 0.1 W. I found a dual package mosfet with Id max at 25 C = 6.1 A (DMT3020LFDB) and 0.7 W power dissipation, and otherwise looks like a nice choice for this design.

I noticed the SOA graph in the datasheet, I am not familiar with this graph, so I am asking here to find out if it is relevant for a switching application vs linear mode operation. The high side switch is Vds = 20 - 5 = 5 V, and the low side 15 - 0 = 15 V.

How does one use this graph? Do I look at Vds = 15V, draw a vertical line up to the DC diagonal intercept and then follow to the left where it intercept Id ~0.03 A (max current at DC, which would conflict with other parts of datasheet). Or do I find Id = 2 amp, follow to the right where it intercepts the DC diagonal, and then down to Vds = 0.3 V, and then determine the minimum gate voltage (Vg) so Vds < 0.3V? In my case Vg = 5, channel is open, Vds < 0.3 V.

enter image description here

DMT3020LFDBQ MOSFET datasheet, source of SOA graph: https://www.diodes.com/assets/Datasheets/DMT3020LFDBQ.pdf

\$\endgroup\$
1
  • \$\begingroup\$ SeriouslyStan - Hi, To comply with the site rule on referencing, please edit the question & add the name and link of the PDF / video / webpage (whatever it was) which was the source of any image you didn't create completely on your own. (Or, if you copied it from a book, please add a full citation - see the rule for details.) Also please remember it's your responsibility to follow that rule in future too. (Please see the tour & help center as site rules here differ from typical forums.) TY \$\endgroup\$
    – SamGibson
    Commented Mar 6 at 22:01

2 Answers 2

1
\$\begingroup\$

The high side switch is Vds = 20-5=5V, and the low side 15-0=15V.

That is not correct. Each MOSFET will experience the same voltage across D-S (drain-source) when in the OFF state, and that voltage will be Vin, which is 20V in this case, not 15V.

In regards to the Safe Operating Area curve: the duration of the switching transition when both voltage and current are not zero (or very close to zero) should be very short, much less than 1us and certainly much less than 100us, which is the curve with the shortest time on the chart. If you want to be conservative, you can use the 100us curve for your application. If you look at the 20V point on the horizontal and draw a vertical line up, it hits the line at a current of 5A. So if the maximum inductor current is kept below 5A you will not exceed the SOA for an on-pulse duration of 100us.

However, the MOSFET rated just 20V will not be suitable for this application, you should select a MOSFET with at least 30V rating.

\$\endgroup\$
0
\$\begingroup\$

The SOA curve is for (linear mode operation - wrong terminology) Saturation region. If we think about it, look at the graph axis's, and choose a point. Let's say 100 V and 5 A, bellow some said line (for 10 us for instance). I'm completely making these numbers up btw, I am just illustrating the point. In a switching application, when the mosfet is off, there is almost zero current through the channel but has volts across it, and when the mosfet is on there is a very small voltage across it with current going through it. Neither of those states describes any point within the SOA curve. So, SOA curves could be used for diode( mosfet) OR'ing, or inrush limiting, etc. Hopefully that helps you out.

EDIT

Andy in the comments correctly pointed out that technically you can use the SOA for all regions of a FET's operational zones.

\$\endgroup\$
7
  • \$\begingroup\$ How is it that the SOA curve shows the line "limited by Rds(on)" if it's for linear operation? \$\endgroup\$
    – Andy aka
    Commented Mar 6 at 23:19
  • \$\begingroup\$ Andy, I used the wrong terminology, I meant saturation region. Was working with BJTs that day..... I corrected my answer \$\endgroup\$ Commented Mar 7 at 19:44
  • \$\begingroup\$ No, that's not true either; it's for all regions. \$\endgroup\$
    – Andy aka
    Commented Mar 7 at 20:35
  • \$\begingroup\$ I believe you are arguing /pointing out technicalities and definitions, which fair enough. But, I believe the majority of use cases are for the saturation region when using an SOA graph. Vast majority wouldn't even look at it for anything to do with SMPS or relatively fast switching cases \$\endgroup\$ Commented Mar 7 at 22:00
  • \$\begingroup\$ Of course you need to look at the SOA for switching applications; it's vital. \$\endgroup\$
    – Andy aka
    Commented Mar 7 at 22:38

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.