active rectifier for reverse voltage protection?

[Michael Prados]

Hi All,

I've got a device which is powered by a car battery, and much to my consternation, when people hook it up in the field, they sometimes find a way to attach the battery terminals backwards.  In normal operation, it can use up to 30 Amps, so although it is fused in several places, the fuses are burly enough that other components fail first in reverse voltage.  This is also a large enough current that the voltage drop of a Schottky would be problematic; I can't really afford to lose more than 100-200mV vs the worst case state of charge for the battery, and even Schottky's seem to be above 500mV in this current range.

I've been considering an active rectifier circuit at the main power connection to the device, which would consist of a low on resistance power FET on the high side (I'm not counting on chassis ground, but I can't count on isolating it either,) with some appropriate circuit to make it active only when the battery is hooked up correctly.

Looking around, I found this:

http://hackaday.com/2011/12/06/reverse-voltage-protection-with-a-p-fet/

Which lays out some useful advice for using a P-channel FET like this.  I am considering using this FET-

http://www.digikey.com/product-detail/en/NP100P04PLG-E1-AY/NP100P04PLG-E1-AYCT-ND/2772845

But, I wonder if anyone else has ever tried this, and has some advice, or an alternative approach.  Thanks,

-Mike

[Windell H. Oskay]

On Sep 24, 2012, at 12:37 PM, Michael Prados <mpr...@gmail.com> wrote:

> Hi All,
>
> I've got a device which is powered by a car battery, and much to my consternation, when people hook it up in the field, they sometimes find a way to attach the battery terminals backwards. In normal operation, it can use up to 30 Amps, so although it is fused in several places, the fuses are burly enough that other components fail first in reverse voltage. This is also a large enough current that the voltage drop of a Schottky would be problematic; I can't really afford to lose more than 100-200mV vs the worst case state of charge for the battery, and even Schottky's seem to be above 500mV in this current range.
>
> I've been considering an active rectifier circuit at the main power connection to the device, which would consist of a low on resistance power FET on the high side (I'm not counting on chassis ground, but I can't count on isolating it either,) with some appropriate circuit to make it active only when the battery is hooked up correctly.

Yes, that's a fine approach. There are also dedicated ICs just for this purpose-- driving high-side FETs, while offering reverse polarity and/or under/over volt protection. See TI's line of "hot swap controllers" with external FET, for example.






Windell H. Oskay, Ph.D.
Co-Founder and Chief Scientist
Evil Mad Science LLC
175 San Lazaro Ave, STE 150
Sunnyvale CA 94086
http://evilmadscience.com/

[Jon Wagner]

Linear makes a bunch of parts that do what you need (you'll need external FETs, and some passives).

This one does the whole banana: UV, OV, RV, OC  (undervoltage, overvoltage, reverse voltage, overcurrent).

http://www.linear.com/product/LTC4364

You especially want that if it might go on a car battery while the car is running, because a car can have an inductive spike to as high as 100V in common failure cases.  Best example is when the alternator is charging full-bore and the connection to the 12V battery rattles loose.  The voltage can spike very high in that case.

This one is much simpler, but doesn't do OV/UV/OC

http://www.linear.com/product/LTC4359

Jon

[Michael Prados]

Thanks, that's quite a device, and around $20 of parts in my volume, but probably worth it for my application.  I'll be spending some quality time with that data sheet, and will have to sweat the FET selection a little.  I like their semi-tangential digression on transient stress in MOSFETs; looks like one of those hidden gems where a piece of wisdom that should have been in a text book gets passed down by an engineer eager to save others from sharing her pain.  

-Mike

--

You received this message because you are subscribed to the Google Groups "SFanalogcircuits" group.
To post to this group, send email to sfanalog...@googlegroups.com.
To unsubscribe from this group, send email to sfanalogcircui...@googlegroups.com.
Visit this group at http://groups.google.com/group/sfanalogcircuits?hl=en.