Rockwell Automation T80020 Application Note Diagnostic Procedures - Manual (Page 35)

Trusted TM
AN-T80020 Diagnostics Procedure
Issue 12 Feb 11 AN-T80020 35
Clearing the MP Non-Volatile RAM (NVRAM) Memory
This procedure is often abused and is only necessary when changing the system.INI module allocation
with processor build 122 (see TN20061) or when a processor suffers from ‘GALPAT’ errors after losing
power on startup (see TN20014). Do not erase the NVRAM for any other purpose because it will delete
all diagnostic evidence in the processor.
GALPAT errors (TN20014)
If a main processor slice has failed due to a GALPAT (galloping pattern) error, entries similar to the
following will appear in the processor Log.
23:53:04 37 Self Test: FCR A DRAM GALPAT test - transient fault in block at address 0xF76CA0
23:53:04 37 Self Test: FCR A DRAM GALPAT test - transient fault in block at address 0xF76CA0
23:53:04 37 Self Test: FCR A DRAM GALPAT test - transient fault in block at address 0xF76CA0
23:53:04 37 Self Test: FCR A DRAM GALPAT test - transient fault in block at address 0xF76CA0
23:53:07 37 Self Test: FCR A DRAM GALPAT test - permanent fault in block at address 0xF76CA0
23:53:07 37 Self Test: FCR A Faulted processor slice removed from operation
23:53:42 24 IMB: Permanent minor fault (MBCU) Chassis 1 Slot 0 FCR A
Here, a discrepancy between slice A (‘FCR’, Fault Containment Region) and the other two slices was
detected in address F76CA0 hex. This was seen five times, which is the count required to declare a
permanent fault. At this point, the slice was disabled. As a result, the IMB later detected a
communications fault. This is much less common after firmware build 115 because errors are
corrected as they are found.
The MP should be swapped to a spare MP, to allow operation to continue. Then the memory can be
cleared to allow a fresh start.
Processor build 122 and System.INI changes (TN20061)
If a System.INI file with a different arrangement of modules is downloaded to a processor with firmware
build 122, and the processor is restarted, the Toolset debugger will be unable to connect. Firmware
build 123 fixes this problem. A workround is to clear the memory.
Clearing the non-volatile RAM
Connect a TC-304 maintenance cable to the faulty processor, and open a communications terminal
program (e.g. TeraTerm, Hyperterminal). Turn the keyswitch to Run. Press Enter and check that a
prompt appears to verify that communications is established.
Remove the faulty processor from the chassis connectors a little way, to turn it off. Re-insert the MP
with the TC-304 cable still connected and the terminal program still active. The MP will now display the
boot-up sequence, similar to that shown below.
P/N 352010 TMR Processor Boot Code - Build 3
(C) Copyright Enea Data AB, 1991-1997
(C) Copyright ICS Triplex, 2001
Cold start
Attempting auto boot, press <ESC> to abort ...
At this point, press the Escape key (ESC) at the terminal program. The MP will report Auto Boot
Aborted and show the boot prompt ‘>’. There is now no application running in the MP; it is only
running the basic low-level boot system.
Type the following:
> envram
This command ‘Erases the Non-Volatile RAM’. This deletes the FAT table that addresses the flash
memory storing the INI and user application, and also deletes the logs and retained variables. The
processor is then unaware of its INI or application.
Restart the MP by cycling power. Load the system.ini file, restart and load the application.