I haven't used the c65 PIC but have had a similar problem when the C54 is
power up, it can
lockup with out a power on reset chip, an sometime even that does not help.
The one thing
that does seem to help is the watchdog timer.



At 19:19 9/25/96 +-9-30, you wrote:
>Hi everybody!
>
>At the moment I am testing a fast tracked prototype circuit that uses a
P16C65-20 to talk to and read multiple Dallas One-Wire products and then
communicate, using RS232, to a host system.  The problem that I am seeing is
that after a reset from being in an undervoltage condition, the PIC's SCI
UART does not appear to function correctly anymore.  The brownout protection
in this circuit is functioning well within specification (checked with a
digital CRO) and everything else on the PIC seems to function correctly.
>
>Following is some more detailed information :
>1.  The fault in the SCI manifests itself as an inability for it to
transmit a byte of data;  the TXIF bit never gets set again after TXREG is
loaded.
>2.   When measuring VDD it appears that if it doesn't drop below ~0.3 V
then when the power is turned on again this problem occurs.  Strange that it
is this value rather than the specified 0.7V !
>3.   When the PIC has powered up into this faulty state, even providing
subsequent reset pulses on BAR MCLR will NOT get the PIC to reset into a
state where the SCI works again.
>4.   Looking at the TX and RX lines, at the PIC, on a logic analyser, it
seems that after TXREG is loaded the TX line will go low to start
transmitting the start bit of the byte but then NEVER goes high again.
>5.   The code for setting up the SCI is sequenced as listed in the
MicroChip P16C65 manual.
>6.   The problem is perfectly repeatable.
>7.   I have not tested the SCI's ability to receive while in this state yet
but I imagine that it won't be able to.
>
>I am starting to think that it is the baud rate generator component not
resetting correctly after being 'scrambled' at low voltage.  If there is no
way in software to fix this then circuitry to guarantee that VDD goes to 0 V
will need to be added.
>
>If anyone has seen the same or similar behaviour in their own tinkering
please email [or condem me to premature greying :) ]
>
>Thanking you all in advance,
>
>Mark Linsenmeier,
>AWA Transponder,
>Adelaide - South
>
 Thanks.
     __           Micheal Yano
    ----          Internet:     miy@cimtek.on.ca
  -CIMTEK-        IBMMAIL:      i1320285@IBMMAIL.COM
------------      Voice:        (905)847-8811 x22
                    FAX:        (905)847-8822 x22