PNP for soft switch and reverse battery protection

It's a long thread, maybe this was stated. Is there a reason why a Schottky diode can't be used for reverse polarity protection? That seems simpler than adding two transistors. Maybe not. Two of the same line item can be easier in fabrication than two line items.

Did you miss...

Message-ID: ...Jim Thompson

Sometimes you simply can't spare 400mV, such as in single-cell applications.

A MOSFET that will pass (say) 10amps is a lot smaller and cheaper than a Schottky that will.

Clifford Heath.

Yes I did. Thanks for asking. And for the circuit suggestion.

But I think if I have to use two devices, then I would rather just use a P-channel mosfet for the polarity protection. It wouldn't need any additional parts since its gate would be tied directly to ground.

Then the switch transistor could either be another mosfet, or just a simple PNP switch. With either choice, I wouldn't need the diodes or the extra resistor.

I need very little current, so a Schottky would be cheaper. But it's mainly the voltage drop. Anything that produces a voltage drop just shortens the life of the battery - assuming the circuit will only operate properly down to a certain supply voltage.

Also, Schottkys aren't known for having all that great reverse leakage values, which might make them problematical for protecting against reverse polarity.

From everything I've read, the P-channel mosfet, installed backwards, is the school-approved solution for reverse polarity protection. They are more expensive, but for my needs, the TP2104 should work fine, and it's 60 cents at Digikey, versus maybe 40 cents for a Schottky.

I have some IRLML6302 that I bought for soft power-switches and reverse protection in single-LiPo applications. $0.37 in singles from Digikey.

Yep, like this...

I used this scheme about 20 years ago for a LiIon battery charge chip design I did for California Micro Devices.

I don't know if I was the first to do so or not. I actually used only one FET and switched the body connection... the chip was not only a charger, but also a discharge controller. ...Jim Thompson

if you don't need the gnd and battery negative to be the same you can use nfets on the low side instead

Why not use a series schottky diode?

ok

nah. backwards its just like a worse PNP. VCE max may be 5V, Hfe takes a hit by a factor of 3 to 30

it does does not protect. current will flow from the base to the collector, turning the reversed PNP on.

through 100K the microcontroller should be safe, but the PNP collector will be trying to pull current out of the voltage regulator.

pnp -9 --+----. .------[VREG]-- | \ / | | ------- | | | ---+--- -->Z----| 5V | 100K | | | \| NPN |-+-100K-+-uC /| | / +--|

If there is very little current, the voltage drop on the Schottky is less, possibly as low as 200 mV.

Your circuit uses less current than the reverse leakage current of a Schottky diode? Wow! That's low.

How many are you making?

So, use a PNP with higher base-emitter reverse voltage. This one is good to 25V:

Yes it will, but -

Detroy the PNP and all the 9V unregulated circuitry outside the regulator and controller.

I was hoping that with the base protected by the diode, or even open, no current would flow from collector to emitter, thus protecting everything from reverse battery connection. And that does work at 5V or 6V, but not at 9V. In my tests, current began to flow at 8V, so it wouldn't protect against a backwards 9V battery.

I've seen a lot of descriptions of the Zener effect when you backward bias the B-E junction of a PNP, but what I'm seeing is the same Zener effect from C to E, although at a higher voltage, no matter what's happening with the base - it can be open, or tied high, and C-E current still flows. After a couple hours of Googling and reading, I've yet to find any discussion of this phenomenon or why that happens.

It's a shame the reverse protection doesn't work at 9V (or actually 9.6V with a new battery), because then the single PNP could both switch the power on and protect against reversal, which is what I wanted to do.

Yes, someone else also recommended that one. It's not clear to me that the Vebo rating necessarily matters in this case. I mean, you certainly would think it would matter, but I'll have to order one at some point and see what happens. At this point, though, I've been using through-hole parts from Digikey, so the surace mount item from Mouser is a bit off the trail. But I do need to test it.

But you know, the world is not full of PNPs with that kind of Vebo rating, or even something as modest as 10V, which would be enough for me.

The c-b junction is forward biased, in series with the b-e zener. So c-e zeners about 0.6 volts higher than b-e.

Some transistors magically become 6.2 volt "reference zeners", super low TC zener diodes, used backwards c-e with base open. But longterm, zenering a b-e junction damages the part.

Why not use a series schottky diode, instead of complicated circuits?

I just wanted to avoid even that voltage drop. Any drop just means the battery has to be replaced sooner. Also, a P-channel mosfet requires no additional parts, has essentially no voltage drop, and in case it even matters, only costs a little more that a Schottky diode.

One other possible problem is that Schottky diodes tend to have larger reverse leakage currents than regular diodes, and I'm just not sure how much reverse current my circuit can withstand. This may be a non-problem. I just don't know.

if preserving the battery is important, why are you using "9V", which is one of the most expensive forms? or is that 6 AA cells in series?

Part of the circuit needs unregulated 9-ish volts, and the

one more question....

for the reverse polarity protection function, why is 100uA leakage a problem?

Reverse polarity protection is just to keep the magic smoke in if someone momentary connects the battery backwards... ...no?