ELS61T with J2ME app suddenly restarts | Thales IoT Developer Community
March 24, 2022 - 1:20pm, 1366 views
Hi, guys
I'm just pushing on on my journey to a stable application platform on ELS61T terminals.
Now the pressing issue is frequent sudden module restarts. In 24/7 operation, there can be intervals when modules restart every couple of minutes, then all is well for some hours. In general, at least 3-5 restarts happen daily (30+ devices).
Scanning the forum posts I found https://iot-developer.thalesgroup.com/threads/bgs5t-restart-after-few-i2c-readwrite-commands talking about I2C bus conditions (pull-ups) as a source of restarts. As we are using I2C comms in this project, this is a candidate for causing these events.
As I'm not very proficient in electronics, can someone explain in layman's terms about what are the requirements for a stable i2c communication with ELS61T. We have 2 Arduino Nano Every microcontroller connected to it.
Thanks in advance for any insight!
Regards,
Jure
Hello,
In fact there is not much information in the referenced topic. Additionally the module is different - BGS5. So for now I don't think that we can bind these two issues. You'd have to do some testing without i2c to verify if the problem disappears.
As for the pull-up's I'm also not HW specialist. I found this in the HW spec: "The value of the pull-up depends on the capactive load of the whole system (I2C Slave + lines). The maximum sink current of I2CDAT and I2CCLK is 4mA."
BR,
Bartłomiej
Hello,
yes, I did some test without I2c and the situation does not improve...
But there is another pattern I've noticed. Quite some modules experience a series of restarts every day between 2 and 3 o'clock in the morning? There is nothing in our application specific to this time interval. Here is an example: SIM provider log of active data connections:
This show a restart every minute every day between 2 and 3 o'clock in the morning.
What could be the cause of this weird behavior? The only thing vaguely related to this is the DST interval, but that is only relevant at two specific days in a year (today incidentaly)...
I'd appreciate and comment to help me dig on...
Regards,
jure
Hello,
Looks really wired. So we can exclude i2c. Maybe it is related to the power supply. What else is the app doing? If the app and the module is not doing anything special at that time, maybe it's the network. Maybe you could for instance test if this depends on the RAT (4G/3G/2G) by limiting the technology with AT^SXRAT command.
Please check the firmware version with ATI1 - we'll see if it's the latest.
BR,
Bartłomiej
Hello,
I've done many tests and experiments, but still no improvements...
I think it's the latest?
It's such a shock for me because the same app has been running for years on EHS5/EHS6 modules, without a single restart for months at a time.
Next is a dummy app just sitting there, logging loop steps and doing nothing but waiting for a restart...
Cheers,
Jure
Hello,
Yes, it is possible that the mdule can reboot and OpenWrt system not as these are in fact separate systems.
What do you mean by a dummy app? Is is a MIDlet running on the module (what is the name?) or some app in OpenWrt system?
By "no mobile data connection" you mean that the module is not registered to the network or there is just no data traffic? Have you tested with SXRAT command as I suggested?
Does the problem exist on more than one device?
BR,
Bartłomiej
Hi,
I intend to strip down my production J2ME app to nothing but the core loop doing some logging, to confirm that a simple do nothing app running continuously does not have these random restart events. Using no mobile data connection means registered (at+cops=0) but not opening any TCP/UDP connections... Yes, as I've reported, different SXRAT settings do not seem to affect this behavior.
Half of my fleet (meaning 30+ devices) exhibits this problem, but the restart frequency varies, from once every couple of days (which would be acceptable) up to once every hour (at certain intervals even more often).
Drives me crazy...
Jure
So with an app running with no i2c, no data exchange but module registered to the network it still happens.
With a dummy app which does almost nothing (network registration?) it still happens?
You have not observed any relation to RAT. For instance setting 2G only does not improve anything.
Is there anything special printed by the module whe it resets? Maybe you have observed any relation to device location?
Maybe you have played with the hardware watchdog? It makes it possible to program the module reboots under a certain condition or periodic reboots.
If we have no other clues I think that it might be hard to diagnose further without module tracing. But for that you'd have to report it to Thales support via SalesForce (if tou have access) or via your distributor.
What you could also try before that is testing with the latest firmware. Please let me know.
BR,
Bartłomiej