Cinterion
®
Java Terminals Hardware Interface Description
8.2 Restart Conditions
69
EHSxT_BGS5T_HID_v02
2014-05-23
Confidential / Preliminary
Page 55 of 70
8.1.1
Reset stages
There are up to three possible escalation stages during a module reset:
•
First stage (regular fast shutdown): The watchdog shuts down the module via an internal
fast shutdown signal. The fast shutdown procedure will then still finish any data activities on
the Java module's flash file system, thus ensuring data integrity, but will no longer deregis-
ter gracefully from the network, thus saving the time required for network deregistration.
Afterwards, i.e. after an internal V180 signal has gone low, the module is regularly re-
started.
•
Second stage (emergency restart): The watchdog resets the module via an internal
EMERG_RST signal. The emergency restart procedure causes the loss of all information
stored in the Java module‘s volatile memory.
•
Third stage (power off): The watchdog switches the module off.
After the first and second stage the watchdog waits for up to three seconds for the internal V180
signal to go LOW. If the V180 signal does not change, the watchdog escalates to the next
stage, until it finally ends up switching off the module. The watchdog can be configured to au-
tomatically switch on resp. power up the module after a shutdown (always-on mode).
8.1.2
Reset Delay
The watchdog implements a protection mechanism to prevent too frequent module resets.
When the delayed reset mechanism is enabled, the watchdog will start its activity only after the
specified amount of time, MIN_START_TIME. A reasonable value for this timeout is 30 min-
utes. After the watchdog startup, after a module reset and also after enabling the watchdog, no
reset of the module is performed before the timeout of MIN_START_TIME.
When the watchdog is enabled, resets can be prevented once for a certain amount of time. This
timeout, TRG_DEL, can be configured via the I
2
C interface. It can be particularly useful when
a software update shall be performed. Using the TRG_DEL timeout will prevent the watchdog
from resetting the module during the running TRG_DEL timeout, so that the update can be per-
formed safely. An upcoming reset event will be shifted and catch up after the TRG_DEL time-
out.
8.2
Restart Conditions
When the watchdog is enabled, it will observe the modules on/off state. When it is configured
to keep the module "always on", it will restart the module after the specified amount of time after
it has discovered that the module has turned off. This important feature is useful in rough envi-
ronments with often power losses and out-of-temperature conditions where it secures a safe
module operation. The timeout condition for the restart feature is called ALWAYS_ON.
