>I think its all a matter of what you expect your circuit to see. Adding >external SOT23 Schottky diodes may be no better than relying on the >internal "diodes", the current still must be limited by some means. The >SOT23 blowing up is no better than the PIC blowing up. The SOT23 has >current limit X, the internal diodes have limit Y. The internal "diodes" >are sufficient if I can guarentee a current limit significantly less than Y >(keeping in mind that the Abs Max Voltage spec. must also be met). Yes, a series input current limiting resistor is needed, BUT No, the situations are quite different. When you use internal protection diodes they cannot start "protecting" anything until they are injecting current into the IC itself and causing changes in the voltages at points internal to the IC. Disaster is threatening if you want it to also to chew gum while it is doing this. The series resistor will limit the current for a given applied voltage but any fault current drawn will go where it may do harm. If you want the circuit to run properly when the Vin >> Vcc is applied then you must limit Vinmax at the PIC pin to ~= Vcc (and Vinlow to ~= 0v If you use external diodes you can design the circuit to start clamping the input voltages before the IC's internal diodes START to pass current into the IC. The reason to use Schottky diodes is that they have a lower "on" (forward conduction) voltage than the internal silicm diodes so they will conduct vefore the internal diodes do and shunt the "fault" currents to Vcc or ground or where-ever the design allows them to. In the first case the IC (PIC) is definitely violating the manufacturers "normal operating conditions" specification. In the latter case it may not be (it's up to the designer). In practice if Schottky diodes are taken from input to Vcc and Ground (appropriate polarity to prevent conduction during normal operation then technically the input spec will be violated by the time the Schottky's are conducting (about +/- 0.3 volts ouitside rails) but in practice this will work with utterly minimal current injected into the protection diodes. The following simple cct nominally allows the formal input Vinhi spec to be met. ZZZ is a zener at say 4.4 volt (hard to buy :-))(4v3 and 4v7 are standard values) Diode D conducts when Vin is > Vzzz + 0.6 A single zener can be used for many inputs but one diode is needed per input. the soft knee of a zener diode makes this cct less sharp than ideal. It has the advantage of NOT pumping up the Vcc for large Vins. A mirror image circuit referenced to Vcc would protect against low Vin-low's. . D | PIC |-------|<|---- .---- | In Z | | Z | Z |---------------RRR----- Vin | Gnd-.------------------------- Russell McMahon _____________________________ >From another world - www.easttimor.com What can one man* do? Help the hungry at no cost to yourself! at http://www.thehungersite.com/ (* - or woman, child or internet enabled intelligent entity :-)) -----Original Message----- From: Jim Hartmann To: PICLIST@MITVMA.MIT.EDU Date: Wednesday, 13 October 1999 22:03 Subject: Re: PICs: Good Design versus what (usually) works - Tutorial / was ... >I think its all a matter of what you expect your circuit to see. Adding >external SOT23 Schottky diodes may be no better than relying on the >internal "diodes", the current still must be limited by some means. The >SOT23 blowing up is no better than the PIC blowing up. The SOT23 has >current limit X, the internal diodes have limit Y. The internal "diodes" >are sufficient if I can guarentee a current limit significantly less than Y >(keeping in mind that the Abs Max Voltage spec. must also be met). > >Jim Hartmann > > > > > >David Bengtson @MITVMA.MIT.EDU> on 10/11/1999 08:37:30 >PM > >Please respond to pic microcontroller discussion list > > >Sent by: pic microcontroller discussion list > > >To: PICLIST@MITVMA.MIT.EDU >cc: >Subject: Re: PICs: Good Design versus what (usually) works - Tutorial / > was ... > > >On Sun, 10 Oct 1999 17:15:10 +1300, in you wrote: > > >>Agree fully. But this problem may occur if resistors alone are used. >> >>> The very very minute cost of either diodes, transistors, or buffers >>> ICs is just not worth destroying your project. >> >> >>Yes. As someone noted, such costs may be not insignificant in volume >>production but must still be considered. >> > >Just a quick comment here. Even though it may seem silly to have $1.00 >in protection components protecting a $0.50 microprocessor, you should >also keep in mind the cost of failure. If that $1.00 in protection >stops a field service call at $500, then perhaps it is worth it. Even >more important, if your products develop a reputation for being >unreliable, your in really deep doo-doo. > >Dave Bengtson >