COP8ACC Datasheet, PDF (29/43 Pages) National Semiconductor (TI) – 8-Bit CMOS ROM Based and OTP Microcontrollers with 4k or 16k Memory and High Resolution A/D

WATCHDOG Operation (Continued)
• The initial WATCHDOG service must match the key data
value in the WATCHDOG Service register WDSVR in
order to avoid a WATCHDOG error.
• Subsequent WATCHDOG services must match all three
data fields in WDSVR in order to avoid WATCHDOG
errors.
• The correct key data value cannot be read from the
WATCHDOG Service register WDSVR. Any attempt to
read this key data value of 01100 from WDSVR will read
as key data value of all 0’s.
• The WATCHDOG detector circuit is inhibited during both
the HALT and IDLE modes.
• The CLOCK MONITOR detector circuit is active during
both the HALT and IDLE modes. Consequently, the de-
vices inadvertently entering the HALT mode will be de-
tected as a CLOCK MONITOR error (provided that the
CLOCK MONITOR enable option has been selected by
the program).
• With the single-pin R/C oscillator mask option selected
and the CLKDLY bit reset, the WATCHDOG service win-
dow will resume following HALT mode from where it left
off before entering the HALT mode.
• With the crystal oscillator mask option selected, or with
the single-pin R/C oscillator mask option selected and the
CLKDLY bit set, the WATCHDOG service window will be
set to its selected value from WDSVR following HALT.
Consequently, the WATCHDOG should not be serviced
for at least 2048 instruction cycles following HALT, but
must be serviced within the selected window to avoid a
WATCHDOG error.
• The IDLE timer T0 is not initialized with RESET.
• The user can sync in to the IDLE counter cycle with an
IDLE counter (T0) interrupt or by monitoring the T0PND
flag. The T0PND flag is set whenever the thirteenth bit of
the IDLE counter toggles (every 4096 instruction cycles).
The user is responsible for resetting the T0PND flag.
• A hardware WATCHDOG service occurs just as the de-
vices exit the IDLE mode. Consequently, the WATCH-
DOG should not be serviced for at least 2048 instruction
cycles following IDLE, but must be serviced within the
selected window to avoid a WATCHDOG error.
• Following RESET, the initial WATCHDOG service (where
the service window and the CLOCK MONITOR enable/
disable must be selected) may be programmed any-
where within the maximum service window (65,536 in-
struction cycles) initialized by RESET. Note that this initial
WATCHDOG service may be programmed within the ini-
tial 2048 instruction cycles without causing a WATCH-
DOG error.
TABLE 8. WATCHDOG Service Actions
Key Data
Match
Don’t Care
Mismatch
Don’t Care
Window Data
Match
Mismatch
Don’t Care
Don’t Care
Clock Monitor
Match
Don’t Care
Don’t Care
Mismatch
Action
Valid Service: Restart Service Window
Error: Generate WATCHDOG Output
Error: Generate WATCHDOG Output
Error: Generate WATCHDOG Output
Detection of Illegal Conditions
The devices can detect various illegal conditions resulting
from coding errors, transient noise, power supply voltage
drops, runaway programs, etc.
Reading of undefined ROM gets zeros. The opcode for
software interrupt is 00. If the program fetches instructions
from undefined ROM, this will force a software interrupt, thus
signaling that an illegal condition has occurred.
The subroutine stack grows down for each call (jump to
subroutine), interrupt, or PUSH, and grows up for each
return or POP. The stack pointer is initialized to RAM location
06F Hex during reset. Consequently, if there are more re-
turns than calls, the stack pointer will point to addresses 070
and 071 Hex (which are undefined RAM). Undefined RAM
from addresses 070 to 07F (Segment 0), and all other seg-
ments (i.e., Segments 4... etc.) is read as all 1’s, which in
turn will cause the program to return to address 7FFF Hex.
This is an undefined ROM location and the instruction
fetched (all 0’s) from this location will generate a software
interrupt signaling an illegal condition.
Thus, the chip can detect the following illegal conditions:
1. Executing from undefined ROM
2. Over “POP”ing the stack by having more returns than
calls.
When the software interrupt occurs, the user can re-initialize
the stack pointer and do a recovery procedure before restart-
ing (this recovery program is probably similar to that follow-
ing reset, but might not contain the same program initializa-
tion procedures). The recovery program should reset the
software interrupt pending bit using the RPND instruction.
MICROWIRE/PLUS
MICROWIRE/PLUS is a serial synchronous communications
interface. The MICROWIRE/PLUS capability enables the
devices to interface with any of National Semiconductor’s
MICROWIRE peripherals (i.e. A/D converters, display driv-
ers, E2PROMs etc.) and with other microcontrollers which
support the MICROWIRE interface. It consists of an 8-bit
serial shift register (SIO) with serial data input (SI), serial
data output (SO) and serial shift clock (SK). Figure 17 shows
a block diagram of the MICROWIRE/PLUS logic.
The shift clock can be selected from either an internal source
or an external source. Operating the MICROWIRE/PLUS
arrangement with the internal clock source is called the
Master mode of operation. Similarly, operating the
MICROWIRE/PLUS arrangement with an external shift clock
is called the Slave mode of operation.
The CNTRL register is used to configure and control the
MICROWIRE/PLUS mode. To use the MICROWIRE/PLUS,
the MSEL bit in the CNTRL register is set to one. In the
master mode, the SK clock rate is selected by the two bits,
SL0 and SL1, in the CNTRL register. Table 9 details the
different clock rates that may be selected.
29
www.national.com