5-6 Series 90™-70 Enhanced Hot Standby CPU Redundancy User's Guide May 2000 GFK-1527A
5
Faulting RCMs, Losing Links, a nd Terminating Communications
There are distinct differen ces between losing a redundant comm unications link and faulting an
RCM.

Faulting the Redundancy Communications Modu le

Faulting the Redundancy Communications Module occurs only when a hardware-related failure
such as a parity error or VME bus error exists.
The following actions are taken when a Redundancy Communications Module is faulted:
1. Loss of Module fault is logged in the PLC Fault Table.
2. All LEDs on the Redundancy Commun ications Module are turned OFF. Th e LED s on th e
other Redundancy Communications Module continue to be updated as long as that RCM
is OK.
3. The module fault contact is set. If the failed Redundancy Communications Module is in the
local main rack, then the SLOT_0X fault contact is set (X is the slot number for the
Redundancy Communications Module). If the failed RCM is in the other unit's main rack, then
the SLOT_71 fault contact is set.
4. The corresponding communi ca tions link is no longer used. If the other lin k is still operating,
that link is used for all further data transfer, and the units can remain in synchronization.
5. If no other communications link is available, the unit functions as a standalone unit when in
RUN mode.
After replacement of the faulted Redundancy Communications Module, power must be cycled to
restore the RCM to service.

Losing a Link

Losing a Link occurs when a link timeout occur s ( that is, no data received in the expect ed t ime
period). Since the system is not certain that a lost link is due to a hardware failure, the Redundancy
Communications Module is not faulted. Some possible causes for a link timeout are:
1. Remote unit has failed and is unable to communicate.
2. Configured fail-wait timeout is too short and a long sweep or communications window has
resulted in a link timeout. Normally the other link will continue to function in this case and the
PLCs remain synchronized. If the condition continues, the remaining communications link
will timeout in a subsequent sweep.
3. A hardware problem is present that prevents data from being transferred but is not detectable
by error checking mechanism s su ch as parity errors (there are no known pr obl em s in this
category).