My nrf9160 which has the v1.0.0 modem and using v1.0.0 sdk has suddenly started continuouly resetting .

[Julia]

My nrf9160 which has the v1.0.0 modem and using v1.0.0 sdk has suddenly started continuouly resetting when i try to run asset tracker application.  I have tried at-command sample also and that gives an assertion Fail with bsdlib irrecoverable error: 3. I seem unable to program it with anything that works. I have programmed with programmer tool and from the command line using west. Any sugestions?

[Håkon Alseth]

Hi,

 

We recently had a corner-case issue in nrfjprog v10.1.1 with nRF9160, where it did not properly run "--sectorerase", thus you needed to do a "nrfjprog -e" first.

What version of nrfjprog are you using?

 

Kind regards,

Håkon

[Julia]

nRfConnect v3.0.0

[Håkon Alseth]

Are you certain that you have updated the modem successfully?

Here's a guide: https://devzone.nordicsemi.com/f/nordic-q-a/49264/unable-to-update-firmware-on-nrf9160/196300#196300

Do you have more than one DK behaving this way?

 

Try issuing "nrfjprog -e", then reprogram the application firmware.

[Julia]

I had definitely upgraded the modem correctly as I followed that guide when I did it. It was also working fine for a few days with the upgraded modem and firmware.

Have done  nrfjprog -e quite a few times. All applications fail as the problem is in the bit where the SPM tries to jump to the application code.

I believe the problem  is in the IPC communication somewhere when bsd_init is being called.

I can happily run a hello world Zephr example .

[Håkon Alseth]

Could you see if you can get modem traces (or do you get this far at all?)?

Here's a blog that explains how: https://devzone.nordicsemi.com/nordic/nordic-blog/b/blog/posts/how-to-get-modem-trace-using-trace-collector-in-nrf-connect

 

If you disable "CONFIG_ASSERT" in the at_client prj.conf (to ensure that you can trace even if assert occurs), then add the trace enable, it should ignore the error and continue running at_client. Here's the settings needed:

CONFIG_ASSERT=n
CONFIG_BSD_LIBRARY_TRACE_ENABLED=y

Then send this command in at_client, before doing anything else:

AT%XMODEMTRACE=1,2

 

Could you see if the trace collector is able to get any data, and upload the .bin file to me?

 

Kind regards,

Håkon

[Julia]

Link monitor shows error when I send the command :-
***** Booting Zephyr OS v1.14.99-ncs2 
*****

The AT host sample started

AT%XMODEMTRACE=1,2

ERROR

[csteed]

addition:

we asked the customer tp read out addresses 0x00FF8010 to 0x00FF8020 in UICR?

Using the command "nrfjprog --memrd 0x00FF8010 --n 16"

Reply from customer:

Both boards say:

0x00FF8010: FFFFFFFF FFFFFFFF FFFFFFFF FFFFFFFE

Have now asked the customer:

Can you  program UICR as follows?

nrfjprog --memwr 0x00FF8014 --val 0xFFFFFFC8
nrfjprog --memwr 0x00FF8018 --val 0x00000000
nrfjprog --memwr 0x00FF8020 --val 0xFFFFFF20

 

Then perform a reset with nrfjprog -r or on-board reset button.

[Julia]

The first two writes were fine the last one nrfjprog --memwr 0x00FF8020 --val 0xFFFFFF20

gave

Parsing parameters.
ERROR: The area to write is not erased.

My board still continually resets although I'm not seeing the BSD error

[tesc]

Hi,

Can you check the contents of 0x00FF8020, with the command "nrfjprog --memrd 0x00FF8020 --n 4"?

Regards,
Terje

[Julia]

0x00FF8020: FFFFFF20 | ...|

[tesc]

Hi,

That should be the correct value. (I.e. no reason to write it with the memwr command.)

If you go on to the next step, "reset with nrfjprog -r or on-board reset button", do you get things to work then?

If not, can you try again the steps for getting trace as described by Håkon Alseth?

Regards,
Terje

[Julia]

After reset the board is still just constantly rebooting for asset tracker

For at command with modified config At commands still give errors.

here is a trace file but suspect it hasn't dtrace-2019-07-23T12-58-27.835Z.binone anything