GE PACSystems RX7i Reference Manual page 135

Chapter 3. CPU Configuration
Fault Parameters
Recoverable Local Redundancy CPUs only. (Fault group 38) Determines whether a single-bit ECC error
Memory Error causes the CPU to stop or allows it to continue running.
Choices: Diagnostic, Fatal.
Default: Diagnostic.
Note:
CPU Over (Fault group 24, error code 1.) When the operating temperature of the CPU exceeds the
Temperature normal operating temperature, system variable #OVR_TMP (%SA8) turns ON.
(To turn it OFF, clear the Controller Fault Table or reset the PLC.)
Default: Diagnostic.
Controller Fault (Read-only.) The maximum number of entries in the Controller Fault Table.
Table Size Value set to 64.
I/O Fault Table (Read-only.) The maximum number of entries in the I/O Fault Table.
Size Value set to 64.
Configuring the CPU to Stop Upon the Loss of a Critical Module
In some cases, you may want to override the RUN Mode behavior of the System Configuration Mismatch
fault. A given module may be critical to the PLC's ability to properly control a process. In this case, if the
module fails then it may be better to have the CPU go to STOP Mode, especially if the CPU is acting as a
backup unit in a redundant system.
One way to cause the CPU to stop is to set the configured action for a Loss-of-Module fault to Fatal so
that the CPU stops if a module failure causes a loss-of-module fault. The correct loss-of-module fault
must be chosen for the critical module of interest: I/O controller, I/O module, and Option module. The
Ethernet communications module is an example of an Option module.
This approach has a couple of disadvantages. First, it applies to all modules of that category, which may
include modules that are not critical to the process. Second, it relies on the content of the fault table. If
the table is cleared via program logic or user action, the CPU will not stop.
In systems that use Ethernet Network Interface Units (ENIUs) for remote I/O, a critical module of interest
may be the Ethernet module that provides the network connection to the ENIU. Other techniques can
be used to provide a more selective response to an Ethernet module failure than the Loss-of-Option
module fault. One technique is to use application logic to monitor the Ethernet Interface Status bits,
which are described in Monitoring the Ethernet Interface Status Bits in the PACSystems RX7i, RX3i and
RSTi-EP TCP/IP Ethernet Communications User Manual, GFK-2224. If the logic determined that a critical
Ethernet module was malfunctioning, it could execute SVC_REQ #13 to stop the CPU.
Since the ENIU uses Ethernet Global Data to communicate with the PACSystems CPU, another selective
technique is to monitor the Exchange Status Words to determine the health of individual EGD exchanges.
For details on this status word, refer to Exchange Status Word Error Codes in PACSystems RX7i, RX3i and
RSTi-EP TCP/IP Ethernet Communications User Manual, GFK-2224. Because the types of errors indicated
by the exchange status word may be temporary in nature, stopping the CPU may not be an appropriate
response for these errors. Nevertheless, the status could be used to tailor the response of the application
to changing conditions in the EGD network.
In some cases, the critical module may reside in an expansion rack. In that case, in addition to the loss-
of-module fault, it is recommended to set the Loss-of-Rack fault to Fatal. Then if the rack fails or loses
power, the CPU will go to STOP Mode.
124
When a multiple-bit ECC error occurs, a Fatal Local Memory Error fault (error
code 169) is logged in the CPU Hardware Fault Group (group number 13).
PACSystems* RX7i, RX3i and RSTi-EP CPU Reference Manual GFK-2222AD