Battery reverse polarity protection circuit

[newtekuser]

I have the following reverse polarity protection circuit for a 12V battery using a P-channel mosfet and schottky diode. I'm using a sample schematic from MPS as a reference to which I've added a fuse, some small caps and a TVS diode for ESD protection. Is there anything else one needs to think about when designing such a system for more robustness/make it more fool-proof? Normally I'd just use a keyed connector to prevent accidental wiring, but in this case my application is using a post terminal.

https://www.monolithicpower.com/learning/resources/designing-a-reverse-polarity-protection-circuit-part-i

[fourfathom]

That TVS diode has a steady-state power rating of about 3W, and with reversed polarity input it will clamp at some unspecified voltage -- probably around 1V (?).  That 3A fuse may not pop quickly enough to save the diode, and this also makes the PFET (P-channel MOSFET) kind of useless.  I like the PFET solution.

You should consider putting the TVS on the output side of that PFET.  Also, you can find PFETs with a +/-20V or so VGS, which will eliminate the need for that Zener diode as long as the power input won't exceed that voltage.

[newtekuser]

That TVS diode has a steady-state power rating of about 3W, and with reversed polarity input it will clamp at some unspecified voltage -- probably around 1V (?).  That 3A fuse may not pop quickly enough to save the diode, and this also makes the PFET (P-channel MOSFET) kind of useless.  I like the PFET solution.

You should consider putting the TVS on the output side of that PFET.  Also, you can find PFETs with a +/-20V or so VGS, which will eliminate the need for that Zener diode as long as the power input won't exceed that voltage.

Thanks much for the feedback!

[Solder_Junkie]

I use an LTC4368 and a pair of back to back MOSFETS in several protection boards that I’ve built. They work perfectly. The circuit is the same as that shown in the datasheet.

Note, the ‘4368 is tiny and the power MOSFETS solder directly to a PCB. I use 100A rated MOSFETS, DMT6002LTS-13 the boards are 2 oz copper to aid heat dissipation.

It might look like overkill, when for basic reverse polarity protection you can use a series diode, but in my case I am protecting equipment worth several thousand and which draws around 25 Amps.

SJ

[xvr]

You should consider putting the TVS on the output side of that PFET.

No. This is will not work out. TVS protected against ESD discharge. If you move them to output of PFET than ESD will kill PFET first.
Better will be to use symmetric TVS version - it will protect against ESD of any polarity, and rest of voltage of wrong polarity will be blocked by ideal diode circuit (implemented by PFET)

[newtekuser]

You should consider putting the TVS on the output side of that PFET.

No. This is will not work out. TVS protected against ESD discharge. If you move them to output of PFET than ESD will kill PFET first.
Better will be to use symmetric TVS version - it will protect against ESD of any polarity, and rest of voltage of wrong polarity will be blocked by ideal diode circuit (implemented by PFET)

By symmetric, do you mean adding the TVS diode before the fuse AND to the drain of the mosfet, or something else?

[Smokey]

Pretty sure that needs to be a bi-directional TVS if you are expecting reverse polarity on the input.
I just use an ideal diode controller when I need this functionality. 

If you are fusing at 3A, and only need less than 2.2A continuous, you can use an integrated solution that also has the second fet for over-voltage and overcurrent protection too:
https://www.ti.com/lit/ds/symlink/lm74202-q1.pdf

[xvr]

> By symmetric, do you mean adding the TVS diode before the fuse AND to the drain of the mosfet, or something else?

I mean bidirectional TVS. Something like P6KE24CA (this is for 24V)

[Terry Bites]

As noted by fourfathom, the current in the Zener is only 1mA or so. That may yield a poorly defined clamp voltage.
Lowering R25 leads to a larger standby current. If the input is from a battery say, in automotive system, that can an issue on standby.
I've had mosfet failures in this kind of scheme from ordinary on-off cycles at the input.
Input ripple and crank load dump transients can be a killer too. Solution: Put a Schottky diode across the mosfet.
Let’s assume there is a large bypass cap on the output. When Vin drops below the voltage on this cap there will be a backflow of current through the body diode.
In the end I used a hotswap controler, an LM74500, with an external MOSFET. See datasheet for application schematics.

Ultimately, a full mosfet controler is the best way out of most of the p-channel pitfalls. Solder_Junkie suggests one solution.
A MAX14672ETB+T (Mouser etc) has all you need for a 4.5A protection scheme in one package. It’s about a buck in small quantities.