On Mon, 26 Jan 1998 22:06:58 -0500 Andy Kunz <montana@FAST.NET> writes:

>Oh, my.  I have spent the past two weeks working with this one.  I've
>put
>hardware together to fix the problem, basically a daughterboard for
>the PIC
>with a reset chip on it.
>
>This situation, in our instance, is caused by a PIC running at 20MHz
>(that's legal) at 5V most of the time.  The problem is that, during an
>AC
>brownout, it is possible for the 5V rail to drop below 4.5V (very
>bad).
>This will put the chip into a mode where the osc still runs, but the
>outputs are locked in their last set state and no processing is being
>done.
>
>Worse yet, the WDT won't trip either, so we have a totally useless
>device
>out there.
>
>And the brown-out detect circuitry (which may or may not be active,
>depending upon which guy burned them :-( won't trip until the voltage
>hits
>about 3.8V for a long-enough period of time.
>



        This is really too bad.  To me, the whole purpose of the WDT is
to recover from stuff like this...  And you say the WDT dies!  Wonderful!

        On using an external reset generator...  sounds like an external
WDT could be more useful.  Under what other conditions does the WDT get
messed up?
        Back in my 6802 days (I still have a few hundred systems out
there running those), I had a watchdog timer on the processor board.  The
original WDT was level triggered instead of edge triggered.  A WDT reset
consisted of setting a line low, then back high again.  Of course, if the
machine crashed with the line low, the WDT was continuously reset and
never timed out.  I had to drive 100 miles in Kansas to press a reset
button because of this.  We changed to edge triggered WDT.


Harold

_____________________________________________________________________
You don't need to buy Internet access to use free Internet e-mail.
Get completely free e-mail from Juno at http://www.juno.com
Or call Juno at (800) 654-JUNO [654-5866]