John Dammeyer wrote:

>The circuit using R45 and R43 limit the current through the diode but a
>spike might still trash the processor.
>
My experience is that circuit consisting of the input through a resistor 
(I use a 330 typically, although 1K or 10K isn't unheard of) and then to 
a pair of schottky diodes - one to vcc and one to vss tend to provide 
more than enough protection from about anything you can throw at it.   
The schottky diodes have a lower Vf  (~340mv vs ~1V @100ma) and switch 
faster - primarily only

You also need to make sure that you have enough capacitance onboard such 
that any "spikey" current which gets coupled into Vcc gets absorbed by 
the cap before it can raise the voltage.  This should be a cap across 
the supply rail.  I'd probably use a low-ESR type just because I like my 
caps to act like caps, and not caps with resistors in them.   Remember 
that there still needs to be somewhere for the current (wattage) 
generated by the wrong-voltage to go.  If your power supply can't sink 
(i.e. it only sources) power, this needs to be your circuit.  12V into 
330 is something like 36ma or so.  So for a VCC+12V wrong input you 
would have to have a minimum load of 36ma (assuming the 330 I use, not 
the higher value you used).

-forrest
-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist