Reverse polarity protection for a DC power input

A board with a barrel jack, a screw terminal or a battery lead will eventually be wired backwards by somebody. The first time is usually in the field or on a customer’s bench, and the result is a dead regulator or a microcontroller that never boots again. Reverse polarity protection is a few cents of parts, but each of the common choices has a failure mode that the schematic does not show.

What you are actually protecting against

Name the event before you pick a circuit. There are three that get confused:

• Reverse polarity: the supply is connected with plus and minus swapped. Current flows the wrong way through everything with a diode to ground, including every IC’s protection structures.
• Reverse current: the input is correctly wired but the source collapses or is switched off while something downstream (a battery, a large capacitor bank, a second supply) pushes current back out.
• Overvoltage: the right polarity at too high a voltage.

A reverse polarity circuit does not necessarily handle the other two. Texas Instruments’ application report on ideal diodes keeps reverse polarity protection and reverse current blocking in separate sections for that reason, and also covers input short or supply interruption and ORing of multiple supplies. A single P-channel MOSFET, as TI’s reverse-polarity application note points out, does not protect against overvoltage. If your product can see a hot-plugged supply with ringing, or a long cable on a vehicle, that is a separate design item.

The three common circuits

Series diode

One diode in the supply line. It cannot conduct backwards, so a reversed supply does nothing. It is two pins, easy to buy and easy to second-source.

The cost is the forward drop, paid continuously and dissipated as heat in the diode. An Electronic Design article on automotive designs notes that a conventional diode’s drop hurts when the source is weak, such as a cold-crank battery, and that a Schottky lowers the drop at higher cost. The same effect applies to any product running from a small cell or a long, thin cable: the regulator downstream sees less headroom than the nominal input.

Use it when the load is small, the input has headroom, and heat is not a concern. Check the diode’s reverse voltage rating against the worst case, not the nominal supply.

P-channel MOSFET in the positive rail

A P-channel MOSFET with its gate tied low (through a resistor) turns on when the supply is correct and stays off when it is reversed. There is no driver circuit, and the drop is the MOSFET’s on-resistance times the current, which can be far lower than a diode at moderate load.

Details that bite:

1. Orientation. The body diode must point in the direction of normal current flow, source toward load. Put the part in backwards, at the schematic or on the footprint, and the body diode conducts the reversed supply straight through. This is a schematic review item and a pin-1 check at assembly.
2. Power-up. The body diode conducts briefly at power-up until the channel turns on, per the Electronic Design article. Do not assume the first microseconds look like the steady state.
3. Gate voltage limit. With a supply above the part’s gate-to-source rating, tying the gate to ground can overstress the gate oxide. TI’s application note describes using a zener and series resistor to limit the gate-to-source voltage, and states that MOSFETs have a specified maximum gate-source voltage. Read that number in the datasheet’s absolute maximum table for the exact part. Do not copy a zener value from a forum post, because the right value depends on the threshold of your part and the supply range of your product.
4. Threshold. The gate must be driven far enough below the source to fully enhance the channel at your lowest input voltage. A part rated at a low on-resistance at a gate voltage you never reach will behave like a resistor that gets hot.

Use it for a single supply in a moderate voltage range, where you have checked the gate limit and the minimum-input enhancement.

Ideal-diode controller with an N-channel MOSFET

A controller IC drives an external MOSFET to behave like a diode with a tiny drop. TI’s ideal-diode application report covers this family and uses it for reverse polarity, reverse current blocking and ORing two supplies. It is the right choice when you need to block reverse current as well, or when the drop of a diode or the input range of a single P-FET does not work.

It costs more parts, a slightly more involved layout, and a controller whose own supply and quiescent current need to fit the power budget. If the product has a sleep-current target, read the controller’s quiescent current in the datasheet and in the context of the whole board. We have seen the same trap discussed in why sleep current beats the datasheet.

Checks before you commit to one

Run these on the schematic, then again on the first board:

1. Worst-case input. Write down the highest and lowest supply the product will ever see, and the longest cable. Check each part’s absolute maximum ratings against the highest, and the MOSFET’s enhancement against the lowest.
2. Dropout budget. Subtract the protection drop at maximum current from the lowest input. Does the regulator still have its required headroom?
3. Fuse or current limit. Protection against a wrong plug is not protection against a short. Decide where the fuse or limiter goes relative to the protection element.
4. Bulk capacitance placement. Electrolytic and ceramic bulk capacitors belong after the protection element, or the reversed supply charges them backwards.
5. Heat at the real current. Calculate dissipation from the worst-case load current, not the typical one.
6. A reversed-plug test. Wire it backwards on a current-limited bench supply with the limit set low, and confirm that nothing downstream sees a negative voltage. This belongs next to the checks in our first power-up note. Do it with a supply you are willing to damage, and with the load disconnected the first time.

Where it goes wrong in production

A circuit that works on the bench can still fail on a line. A footprint with unmarked body-diode direction invites a flipped MOSFET, and a placement error there silently turns protection into a short path. Put polarity marks on the silkscreen and in the assembly drawing, and make the design review checklist include the orientation of every protection part. A test fixture that powers the board is also a place where a backwards connector gets plugged in dozens of times, so give the fixture its own protection rather than relying on the product’s.

Last, record the reason you chose the circuit. Three years from now somebody will look at a P-FET with a zener and a resistor and wonder if they can delete it to save a line on the bill of materials.

How we approach it

We review the input stage against the real supply range, the real cable and the real load, pick the simplest circuit that survives it, and write the reasoning down next to the schematic. If you send us the schematic, the intended supply and any field failures, we can tell you which of the three fits and what to measure on the first board.

Primary sources used in this note: TI, Basics of Ideal Diodes; TI reverse-polarity application note SNOAA23; Electronic Design, Reverse-Polarity Protection in Automotive Design.