1
\$\begingroup\$

I have a 5V DC input to a fixed 5V to 3.3V LDO (LM3940) that I want to add reverse polarity protection in front of via a P-channel MOSFET.

P-FET Reverse Polarity Protection Circuit

What I am struggling to determine is how to analyze/choose a P-FET based on its parameters such as \$R_{DS(on)}\$.

I understand the concept of connecting the P-FET "in reverse" to allow the body diode to conduct and then fully turn on the FET via the \$V_{GS}\$ differential. I therefore narrowed my search down to P-FETs with a \$V_{GS(th)}\$ well above -4.3V (my 5V input minus the .7V body diode drop).

I understand that once past \$V_{GS(th)}\$, the forward voltage drop across the FET is a function of \$R_{DS(on)}\$, which is itself a function of \$V_{GS}\$ and \$I_{d}\$.

For example, my LDO has a minimum input voltage of 4.5V to reliably product 3.3V, so the voltage drop across the FET has to be .3V or less at 5V and 1A (my circuit's safe maximum current with buffer).

So, how can I choose a P-FET that has sufficiently low \$R_{DS(on)}\$ at 5V \$V_{d}\$, between my operational range of 100mA and 1A, to always have less than a \$V_{DS}\$ drop of less than .3V?

\$\endgroup\$
4
  • 1
    \$\begingroup\$ I think you mean well below not well above. \$\endgroup\$
    – Andy aka
    Sep 21 at 17:32
  • 3
    \$\begingroup\$ as V_GS is <0V, I think by "above" he means closer to zero (which seems correct to me). But with negative numbers, above and below often get confusing (is it in absolute value or not?) \$\endgroup\$
    – Sandro
    Sep 21 at 17:37
  • \$\begingroup\$ Haha not going to lie guys I literally had to sit back in my chair and think about it for about 30 seconds when I wrote it originally \$\endgroup\$ Sep 21 at 21:02
  • \$\begingroup\$ Remember in this configuration your gate-source voltage rating also has to exceed your power supply voltage! If it's a 5 V supply that probably won't be an issue, but doing reverse-polarity protection on a 12 V or 24 V supply requires care, and you probably need to use a different circuit entirely if the supply voltage is 40 V or higher. \$\endgroup\$
    – Hearth
    Sep 21 at 23:48

1 Answer 1

3
\$\begingroup\$

If you ask this question, I suppose this is either a hobby project or a very small series/prototype. This means that priority is to get something simple and reliable, not to reduce cost by a few cents as in mass production.

So I would recommend the following procedure:

  1. go to any decent electronics vendor with filtering features (ie not ebay, Amazon or aliexpress, but something like Digikey, RS, Mouser, ...)
  2. select P-channel MOSFETs
  3. check filters that you want composents that are in stock, active, and with datasheet (you can also exclude marketplace components if relevant, and set filters on quantities if provided)
  4. select through hole (or surface mounted if you prefer)
  5. select VGS_th bellow some low value (between 2 and 3 V): by having a decent margin, you increase the likelihood that the MOSFET is fully on at 5 V
  6. select a saturation well above your needs. You say 1 A max, so filter for >10 A: this way, you are unlikely to run into thermal issues (often, near maximum, you are limited by duty cycle or need to add a heatsink)
  7. select the lower end of the Rds_on values (you need <300 mΩ if you want less than 0.3 V drop, so filter for bellow 60 mΩ)

With the above filters, I expect that most remaining MOSFETs will do (on Digikey, I end up with 10 results).

Pick one which size and price suits you, and check the datasheet in details to see if it fits your needs :

  • with Vgs = 5 V, at Ids=1 A, Vds is bellow 0.1 V
  • Rds_on(@Vgs=5V) * Ids^2 << max power

NB: this procedure is very conservative, so you will end up with some oversized MOSFETs: but excepted if space is at premium, there is no real downside to it (excepted a few cents more, but I suppose it's better to spend 1$ more than to have to read 20 datasheets until you find a suitable MOSFET). If you go for mass production, where 1$ per PCB is far more important than the cost of the time you spend selecting the ideal part, then my procedure is not adapted.

EDIT : procedure if you start from a random P-MOS datasheet :

  1. check that V_GS_th is smaller than 4.3V (with some margin)
  2. check that your mosfet supports your max voltages (ie VGS_max>5V and Vds_max>5V) : for such low voltage, you will have a hard time finding mosfets that do not satisfy this constraint. For a 24V application, this becomes non trivial.
  3. check the Vds vs Ids curve for V_GS =5V, to check that your voltage drop (Vds) is bellow 0.3V @ Ids=1A (for V_GS=5V)
  4. check that the dissipated power (P=Vds*Ids, taken at VGS=5V from the Vds vs Ids curve (at Ids=1A)) is bellow the maximum you can dissipate (some datasheet tel you this directly, for others you have to use thermal resistance and maximal junction temperature). Be careful about the exact condition of data (sometimes, it is only valid at 10% duty-cycle, or with a huge copper plane on your PCB, ... : read the fine print)
  5. check the footprint is of a reasonable size/type for your application, and that the price is OK

If I haven't forgotten anything, that's all for your application

\$\endgroup\$
2
  • \$\begingroup\$ I sincerely appreciate the suggestion, and ultimately this may be what I do since you are absolutely correct, this is a low volume hobby dev board application. But I would even more so appreciate a suggestion on how to solve this problem from a data(sheet) position, based on the fundamental concepts which underpin MOSFET operation. \$\endgroup\$ Sep 21 at 21:06
  • 1
    \$\begingroup\$ I added a second procedure starting from the datasheet to check if a mosfet is OK \$\endgroup\$
    – Sandro
    Sep 21 at 21:25

Your Answer

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

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