My lovely RaspberryPi is dead...

I'd go, maybe, 2200 uF w/ da e-lytic.  Your best protection against ESD indoors is humidity.  I have an off-the-shelf feedback controller running a dumb humidifier through an SSR.  Attached to the front-end of the controller is a linear humidity sensor I got for $20.  Whole deal just < $200.  Took me half-an-hour to setup.

After reading this post for no other reason other than a notificaiton on my droid I thought I would go mess with my RPi since I was having some trouble with RASPXBMC which turned out was because I needed a power supply with higher current. I resolved that with a 12$ mini hub I bought on a impulse.  Here is where it gets kinda odd.

My 12$ hub is plugged into wall power

The pi power is plugged into the Hub.  Great now it powers up.

For giggles I plugged the hubs USB imput cable into the PI thinking this can't work.  I am creating a ground loop.  It works perfectly.

So then I thought I wonder if I can plug in the wifi donge, keyboard and mouse to the hub.  It works PERFECTLY!

Then I thought.  I wonder if it will also power my ARDUINO!.  It works perfectly and I can program it through INO over SSH through the Pi.

Then I thought.  I wonder if I can put the PiFace daughter board on.  It works!

Then I thought.  I wonder if I can put the LCD screen on the Arduino.  It STILL WORKS!!

I stood back and looked at it and discovered that I accidently created a whole system of mashed together platforms that I could build just about anything with all because I was reading this thread about power issues.  Its been running for 2 days now with out a glitch.  Next step will be to build a battery pack that can power the whole lot of it.

I know I didn't help any but I thought it was funny.

An improvement to the superZener clipper I'd illustrated,would be to substitute a TL431 for the Zener.  We'd need a few resistors, but we could then trim the clipping point.

>>> created by Don Hersey in Raspberry Pi

>

Another clever idea for a protection circuit:

Simple reverse-polarity-protection circuit has no voltage drop | EDN

The only downside is that reverse power is connected until the relay operates.  I know it is for a very short time but it only takes a short time to do damage.

David

Well, let's give that one a think. . .

Now, it may defeat the original purpose, (which is a dubious one in my mind, I don't lose sleep over these little losses in prototype systems) somewhat, but an inrush surge limiter in series w/ da power could clear that up, maybe, at the cost of always having to draw 200mA.  Be a little brutish, but we could throw a backward Shottky diode across our load and hope our relay contacts don't weld before the relay actuates, a fairly safe bet, IMHO, if we are talking wall-wart.  Seems more bulk capacitance across the supply would help by allowing more time in the reverse-case. . .

Frankly, if it is a little CMOS controller, I often just use a linear regulator to drop to my logic potential.  I just tell my self that the intervening diode(s) are helping my drop-out situation.

You could add another relay powered when the supply is correct way round to provide power to load so zero voltage drop but takes a little more current.

David

One could use a bistable relay. . .

This seems to get sorta rococo.  This event of getting skunked by the reverse-polarity wall-wart-from-perdition is a rare, although terrible, event.  I'm thinkin' if Qdiss is that bigga deal, how about a series picofuse and a shunt backwards Shottky diode?  We can tape a spare up under the box.

Even simpler  ==  Just swap the NO and NC contacts, and turn around all the diodes and LEDs…. DOH

So when power is applied the relay coil energizes only if polarity is correct, then power is passed to load after a delay of a few milliseconds.

No idea what the author was smoking, it’s an old trick and he still managed to get it backwards.  The way he does it, that few milliseconds of reverse polarity is an awful long time in the world of electrons.