On Wed, 12 Jun 2002 at 12:24:19 -0400, someone wrote: > From: personal > > Two of my PIC stopped function, today. I lost them. > > All can be programmed and verified OK, one can even > write to EEPROM. > > But, did not osci and dead fish. > > how to check if a PIC is broken? seems PIC > is weaker than normal mosfet........ (Warning - all the following is concerned with PIC'84, although much of it will apply to other PICs too.) I wouldn't go along with the original message's last statement - in my experience PICs are pretty tough little ICs! We get a lot which are put in wrong-way-round and the like, and most of them survive. Simply getting the PIC to program isn't anything like good enough (although it is obviously very important!). Programming only exercises three pins. Generally, what seems to break PICs is similar to what breaks many digital ICs, namely contention on output pins. That is, an output pin tries to go HI when it is connected either to GND or to another pin which is trying to drive LO. (Or it's trying to drive LO and is connected to VDD or to another pin which is trying to drive HI.) PICs can source/sink a fair bit of current, and the resultant dissipation in a contention situation is enough to blow the driver, rendering that pin dead (sometimes the pin can be used for input, but generally the bondwire goes too, making it open circuit). After a lot of messing around, I ended up producing a little unit with an 18-pin ZIF socket, some LEDs, a dc-dc converter, an analogue switch and a second PIC. The latter attempts to program the device under test (DUT) - if programming fails, this is indicated on a LED and the control PIC stops. The DUT is programmed to have a 4-bit twisted ring counter pattern (0000 > 0001 > 0011 > 0111 > 1111 > 1110 etc) on sets of four output pins at a time, with the pins connected A0 to A4 to B0 to B4 (etc). The control PIC monitors the 4-bit set, looking for the right patterns, and gives a fault indication if the expected patterns don't appear. Once all four patterns have been detected (ie each subset of pins has worked OK as an output), the control PIC gives an "OK" indication. Source is just over 900 lines of heavily commented assembler - if anyone's interested it's at . HTH, -- Tim Forcer Department of Electronics & Computer Science The University of Southampton, UK -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.