[sdiy] power supply reverse polarity protection

[Czech Martin]

Some regulators can withstand +37 volts of input to output
voltage (sign intended positive, LM317). It is not so hard
to design a PSU such that this limit will not be hurt.
Except maybe if there is a 110/230V switch in there...
Perhaps a fat zener diode can be used to limit the voltage,
and transients , too.

A thing that is more threatening is wrong polarity.
It may happen during construction , on the work bench
or with self made lab equipment (banana plugs)
that power supplies are swapped to the wrong.

I guess that most regulators will then be forward biased
diodes or transistors, i.e. large currents will flow
perhaps through small current paths. Perhaps the wrong polarity
will only fry the regulator, perhaps an open fault will
appear and then the rest of the costly modulea will fry.

It takes some ms to blow a fuse, and some more to discharge
the electrolytic caps in the PSU and elsewhere.
This is still enough time and current to fry any junction.

A series diode in front of the regulator input will
of course give good protection against wrong polarity.
But this implies also voltage drop and power loss.

Perhaps a parallel diode can limit the wrong polarity
voltage on the regulator input to 0.8 - 1.0 V (lots of current!).
It's hard to predict if the regulator circuit will be 
protected then, and what will happen at the output.

Are there any experiences, measurements?

m.c.