Trivial topic indeed: why one would ever need to protect a circuit from a rolled-over power supply? There are countless potentials for mistakes that can result in an electronic developer pulling out his hairs and staring at his design worth of many hours or days of work - the one that he bricked just a moment ago.
To name just a few:
- Universal power supplies with their rather ambiguous plugs that can be used with both "+" on inner or outer connector.
- A wall-plug adapter: they all use "+" on the inner connector, don't they? Wrong!
- Any type of a connector without hard keying, for ex. computer "jumper"-like junkies that are very cheap and very practical for quick mock-ups.
- Scenarios like that: on one day all wires remained at your reach were either black or blue; few days later you were sure that "-" was blue; day after it proved to be black (you planned on using black/red pair) - oops.
- Just a "bad day" would be enough, no special reason to swap two humble wires or plug it holding a board upside down - and it's all gone...
Still there are people out there who would throw in my face something like: "I would never be so stupid to reverse the polarity of a power supply!" Let them develop few more novel electronic designs on their own and gain some experience first.
For the rest of us - we do know from the experience that shit happens. I could easily recall few such disasters in my career. Once I had to re-solder 25 chips out of 27 that the design consisted of. Luckily it was all made of the old good DIP cases at that time...
Since then I routinely put a rectifier diode near each DC power inlet on almost every experimental board I build.
Well, this topic applies not only to experimental designs: recently I saw a gigantic laser cutter machine that was handicapped by one very unfortunate technician who accidentally reversed power wires feeding the cutter's head vertical stabilizer/sensor. Surprisingly that particular circuit survived (it had a diode shortening the reversed PS). But lots of amplifiers and logic that were following it - got fried instantly. That was one of the very expensive PS roll-over, and to fix that - Russian-Italian language barrier was not making things any simpler...
The diode in series
Diodes could be used in series to block the reversed polarity peacefully.
This option is simple, safe, and has only one disadvantage - the voltage drop across the diode: something as from 0.2V on a Schottky and up to 0.7...1V on usual rectifier diodes. This loss can become quite annoying especially if we use batteries or a stabilized power supply. Power loss on the diode will also become a burden if we need to deliver relatively high currents.
The diode in parallel
Diodes could short-circuit incorrect polarity at the PS inlet.
There is no loss associated with this solution provided everything is plugged as it should. In case of the failure the power supply is at risk. Depending on the power supply capabilities - the shorting diode can be fried and then the circuit it was trying to protect will go to hell too. Anyway I used this kind of protection while working with stabilized supplies where I knew for sure the over-current protection was good enough. Still once I got burning marks on my fingers after touching such a power supply that was fighting against a mighty Schottky rectifier.
p-channel MOSFET - obvious but expensive
The straight solution would be to use a power MOSFET to protect the circuit. Shall we had a cheap source of quality p-channel MOSFETs - we could put it in a positive supply rail.
You may want to check
n-channel MOSFET - the best
|Power enhancement mode MOSFET + 7.2-15V Zener + R of few dozens of KOhm = SAFETY
Provided that a reasonably big power MOSFET has been chosen for a particular design - there will be a neglectable voltage drop across the protection device if everything is operating under normal conditions. If you dig through your stock of free MOSFET's (see below) I am sure you'll find a few with Rds(ON) that equals just few milli-Ohms. Even if the device consumes few Amperes of the current - there still will be just few milli-Volts of a voltage drop across the protection MOSFET's channel.
In case of the reversed PS polarity accident - there is no current at all. Both the target circuit and the (wrongly connected) power supply are in no danger.
The only imaginable disadvantage of this circuit is virtual (thus it is in our imagination only): the fact that it brakes what is commonly perceived as a "ground" connection. But mind you - should you build a final device with permanent and possibly inter-connected power supplies (with the ground "stars" etc.) - you would never need polarity protection, right? And if one would still opt for security - there are better practices of supplying power to the sensitive designs, than using industry standard 3-pin regulators. But this topic merits few big posts, if not a book. Let's leave it open for now.
Static Electricity Warning
Everyone would warn you about MOSFETs being extremely sensitive to the static electricity - and for the good reason. MOSFET's are indeed very fragile and could die instantly and forever. But... not much to worry about our little power beasts. Some of them already have a built-in protection, but none of them would die because of your cat passing by: power FET devices unfortunately have large input capacitance (the one between the gate and the rest of the crystal). This capacitance effectively protects power MOSFET' gates from attaining high and dangerous voltages in the event of a human body discharge. Therefore following basic ESD protection practices will ensure the device's good health, even when it's coming from the oven (oops, that possibly was a hint about direction to the free MOSFET's ;).
I am not alone
What am I describing here is no doubts a well known practice. Oh well, if only those military developers were blogging their schematics! Look what I found on this topic on one of the forums:
> > I believe it is pretty well standard practice to use an N-channel
> > MOSFET in the return lead of military power supplies (28V input).
> > Drain to supply negative, source to the negative of the PSU and
> > the gate driven by a protected derivative of the positive supply.
FREE source of power MOSFETs
coming soon 😉
A very simplistic 100KHz square wave generator is protected:
|Power MOSFET + 9.1V Zener + 15KOhm resistor = SAFETY 🙂
[Click on images for close-up]
Sine wave and saw-tooth generators are safe on this board:
|Power MOSFET (IR LU7843)
|9.1V Zener + 10KOhm SMD resistor