Hi Olin, RA4 doesn't have an input protect diode to Vdd so the spec for 
it differs from all the other pins.  I am using the A/D, thanks for the 
warning about readings being affected.  Fortunately it is a 
non-precision application so it will probably be fine, but I'll 
double-check.

As for additional components ... it isn't the $ cost as much as board 
space, more components to fail or be mixed up, more points of possible 
failure during assembly, yadda yadda.  Am trying to make it all as 
simple as possible... and this particular design tradeoff I didn't know 
enough about to do it intelligently.  I've taken the cautious approach 
in the past but wanted to dig a little deeper this time.

J


Olin Lathrop wrote:

> Jesse Lackey wrote:
> 
>> Now reading the absolute max ratings, of course, you can't do either of
>> these things.  Any pin except A4/MCLR is Vdd+0.3V max; A4 is 8.5V max.
> 
> 
> So what's the question on RA4?  Do you think "absolute maximum" is 
> ambiguous
> or somehow doesn't apply to you?
> 
>> There is another spec that says input clamp current (IIK (VI < 0 or VI >
>> VDD)) is +-20mA.  (This is the 18F1220/1320 datasheet FWIW.)  I presume
>> this doesn't mean the PIC clamps the current to +-20mA... what does it
>> mean, exactly?
> 
> 
> I take it to mean that if you put more than 20mA thru the protection diodes
> they can go poof.
> 
>> So, if I current limit to a milliamp or less, maybe even .1mA, am I ok
>> here?  I don't see the harm in forward-biasing the input protect diodes
>> with a small current, but then again I don't know much about chip design
>> norms...
> 
> 
> A small current is OK in some cases.  Even small currents can upset A/D
> readings.  Also this current must not be on when the chip is powered up, or
> it could go into SCR mode.  100uA applied after the chip is fully 
> powered on
> and if you're not using the A/D is probably OK.
> 
>> I can add transistor switches etc. if need be, but if the above is
>> workable it is definitely preferred.
>>
>> This is going to be a commercial product, so unfortunately solutions
>> that "cheat" and will "normally work at least for awhile" aren't going
>> to cut it...
> 
> 
> What's the cost of a failure?  Compare that to the $.05 for the extra
> transistor and resistors.  Unless this is a really high volume product 
> doing
> it right sounds like the clear choice.
> 
> 
> ******************************************************************
> Embed Inc, Littleton Massachusetts, (978) 742-9014.  #1 PIC
> consultant in 2004 program year.  http://www.embedinc.com/products
-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist