troubleshooting of produced prototype: nrf9160

RE: troubleshooting of produced prototype: nrf9160

Andreas — Fri, 11 Dec 2020 13:40:06 GMT

Hi,

Rev 2 of the SiP has this reverted back to 4.7µF, so for future designs I would use 4.7µF, not 47µF. Rev1, in case you end up using that, will in most cases work just fine with 4.7µF also.

Best regards,
Andreas

RE: troubleshooting of produced prototype: nrf9160

Pertti Kasanen — Fri, 11 Dec 2020 08:21:45 GMT

I’m curious about this 4,7uF -> 47uF recommendation. We consider adding it to our new version (one that we want to work well outdoors in the Scandinavian winter climaate).

This has not been included in the new reference desing files or in the Hardware Integration Guide 1.2. Is this 47uF still a valid recommendation (with the new Rev 2 B1A SoC)?

RE: troubleshooting of produced prototype: nrf9160

Andreas — Fri, 10 Jan 2020 15:10:01 GMT

Hi,

By the way, your application does use the modem, right? If not, could you try running a sample that uses the modem? (the modem FW has to enable the power supply for the SIM)

Best regards,

Andreas

RE: troubleshooting of produced prototype: nrf9160

David — Fri, 10 Jan 2020 12:11:10 GMT

Hi,

I understand.

I just came over these posts,

https://devzone.nordicsemi.com/f/nordic-q-a/55241/not-getting-1-8-volt-on-the-sim_1v8-pin-on-custom-board

maybe this is firmware related?, anything I should know about before I do further investigations on this?

Regards, David

RE: troubleshooting of produced prototype: nrf9160

Andreas — Fri, 10 Jan 2020 12:07:47 GMT

Hi,

OK, thanks for correcting.

I am a 'HW guy' so not really the right person to judge your FW or the modem FW, but as Didrik judged it we believe this to be HW related.

Best regards,

Andreas

RE: troubleshooting of produced prototype: nrf9160

David — Fri, 10 Jan 2020 12:03:15 GMT

Hello again,

I must correct myself:

David said:

I measure 0 Ohm between sim traces and gnd, when board is power off. Is this normal?

J7 was set to ground.

Thank you for replying so fast, It seems that the main issue now, is the missing 1V8 voltage from the SIM vdd.

I would appreciate it, if your support team could answer me on, what possily can cause this, maybe firmware related?

Regards, David

RE: troubleshooting of produced prototype: nrf9160

Andreas — Fri, 10 Jan 2020 11:55:47 GMT

Hi,

[quote user="Awia"]If I meassure the voltage on some other boards of my prototype, I meassure a stable 2.2V on DEC0, on the one prototype that I am having the issue with the sim-interface voltage, I meassure here between 2.1 and 2.2 VDC, oscillation. [/quote]

OK, nice to know although I am not sure this answers anything at this point.

[quote user="Awia"]I am still waiting for the correct capacitor components, for DEC0 -> GND. Therefore, DEC0 is floating at the moment.[/quote]

Not sure if this is ideal, 4.7µF is better than nothing I would assume. It is not necessarily incorrect, but with 47µF the clock signals will be more resilient across PVT variation and is recommended for mass production devices.

[quote user="Awia"]I measure 0 Ohm between sim traces and gnd, when board is power off. Is this normal?[/quote]

Not sure, I have forwarded it internally.

[quote user="Awia"]You are correct, I am using an external strap between (pin 2 of J7 -> C24 pad 1 (capacitor below analoge switch)) [/quote]

Thanks for the confirmation, this is all good then.

Best regards,

Andreas

RE: troubleshooting of produced prototype: nrf9160

David — Fri, 10 Jan 2020 11:19:26 GMT

I measure 0 Ohm between sim traces and gnd, when board is power off. Is this normal?

Regards, David

RE: troubleshooting of produced prototype: nrf9160

David — Fri, 10 Jan 2020 10:07:05 GMT

You are correct, I am using an external strap between (pin 2 of J7 -> C24 pad 1 (capacitor below analoge switch))

RE: troubleshooting of produced prototype: nrf9160

David — Fri, 10 Jan 2020 10:02:47 GMT

Hello,

If I meassure the voltage on some other boards of my prototype, I meassure a stable 2.2V on DEC0, on the one prototype that I am having the issue with the sim-interface voltage, I meassure here between 2.1 and 2.2 VDC, oscillation.

As I mentioned here first, I have 3 other boards of the same prototype, which does not have the fluctuation on the SIM 1V8 VDD, but on these I meassure 0VDC on SIM-interface VDD.

I am still waiting for the correct capacitor components, for DEC0 -> GND. Therefore, DEC0 is floating at the moment.

I have not tested "shorting the SIM traces past U8", do you mean SIM_IO, SIM_CLK, SIM_RST ?, should I test shorting all these together?

Regards, David

RE: troubleshooting of produced prototype: nrf9160

Andreas — Fri, 10 Jan 2020 10:01:57 GMT

Hi,

Another thing, looking at the pictures, you have cut SIM_1V8 between J7/U8 and nRF9160, but from the second picture you are measuring SIM_1V8 at pin 2 of J7. Fair enough. But looking at the layout I am not seeing how this pad/net is connected to the SIM_1V8 from the SiP, that is only fed into the VCC pin:

Is there any external strap, internal layers or anything that I am missing?

Best regards,

Andreas

RE: troubleshooting of produced prototype: nrf9160

Andreas — Fri, 10 Jan 2020 09:32:09 GMT

Hi David,

It might seem like this is in some way related to your HW, either the PCB design or SiP internal. Have you been able to get it working if shorting the SIM traces past U8?

Just to be on the safe side, I am assuming your prototype is behaving normally otherwise? What voltage are you seeing on DEC0 (should be 2.2V)?

I will be forwarding internally to get someone who knows the SIM module to comment.

Best regards,

Andreas

RE: troubleshooting of produced prototype: nrf9160

David — Thu, 09 Jan 2020 13:09:06 GMT

Hello,

I've managed to get the prototype working with my application, when following this tutorial

https://devzone.nordicsemi.com/nordic/cellular-iot-guides/b/getting-started-cellular/posts/nrf-connect-sdk-tutorial.

Therefore, I now have to solve the sim-interface issue on my prototype.

I only had 10uF ceramics laying around, so I added 4 for IO_supplyu/DEC0.

If I cut the 1V8 sim vdd trace, which goes to the analoge switch, I still can measure the same fluctuation on this voltage, when measuring between "gnd / analoge switch pin 14".

I can even measure it on the analoge switch VCC input "pin 14", even if it is not connedted to the 1V8 sim vdd from nrf9160 SiP.

I have checked the close-by traces for this analoge switch, and these are (SIM_IO, SIM_CLK, SIM_RST).

See attached pictures below, for visual.

What can be causing this ?, I think it can have something to do with the SIM_IO, SIM_CLK, SIM_RST lines?

At what circumstances does the sim-interface behave this way?

Regards, David

RE: troubleshooting of produced prototype: nrf9160

Didrik Rokhaug — Thu, 09 Jan 2020 09:03:00 GMT

Could you give me some more information about how your application fails?

Are you not able to program it, or does it fail at run-time?

What error messages do you get?

[quote user="Awia"][/quote]

I've read up on the SPM, and for everything to work, I have to first upload the SPM firmware without "ns", then I have to upload my application firmware, as "ns".

Also, I have to set in proj.conf, CONFIG_SPM=n for this to be available as .elf, for me to build. If not, this wont work.

While it is true that you had to flash the two applications separately before, that is no longer necessary. If you build for the non-secure version of a board (with "ns" at the end of the board name), it should configure the application correctly, build the Secure Partition Manager and merge the hex files into a single merged.hex.

What happens if you remove the CONFIG_SPM and CONFIG_TRUSTED_NONSECURE lines from your prj.conf, and add

include($ENV{ZEPHYR_BASE}/../nrf/cmake/boilerplate.cmake)

right above the other include macro in your CMakeLists.txt file?

RE: troubleshooting of produced prototype: nrf9160

David — Wed, 08 Jan 2020 14:29:16 GMT

Hey Didrik,

Thank you for the reply. I will have a closer look at what you commented, about the hardware.

First I have to solve the application FW/SW for the prototype, maybe you have help me answer me on some troubles I have been having?

So, Im using the PCA10090 for reference.

I've set-up the Segger embedded environment correctly, and installed west with the latest repositories, v1.1.0.

I have managed to create a FW which I can upload over SWD to reference board, this works. This works both with .elf and merged.hex (non-secure) and (secure).

If I try to upload this to my prototype, it will not work with merged.hex, it only works when I build without "ns", as .elf.

-----

I've read up on the SPM, and for everything to work, I have to first upload the SPM firmware without "ns", then I have to upload my application firmware, as "ns".

Also, I have to set in proj.conf, CONFIG_SPM=n for this to be available as .elf, for me to build. If not, this wont work.

-----

Ideally, I would like to use the refernece board for prototyping, so that this can be 100% compatible with my prototype, as-far as code compatibility, as multiple image file (merged.hex).

How can I achieve this?

Im not very experience with Zephyr yet, and trying to learn.

I'd appreciate a fast response.

In advance, thank you.

below is project files.

Regards, David

CONFIG_BSD_LIBRARY=y
CONFIG_GPIO=y
CONFIG_SERIAL=n
CONFIG_STDOUT_CONSOLE=y
CONFIG_UART_INTERRUPT_DRIVEN=y
CONFIG_TEST_RANDOM_GENERATOR=y
CONFIG_NETWORKING=y
CONFIG_NET_BUF_USER_DATA_SIZE=1
CONFIG_NET_SOCKETS_OFFLOAD=y
CONFIG_NET_SOCKETS=y
CONFIG_NET_SOCKETS_POSIX_NAMES=y
CONFIG_NET_RAW_MODE=y
CONFIG_TRUSTED_EXECUTION_NONSECURE=y
CONFIG_LOG=n
CONFIG_LOG_DEFAULT_LEVEL=4
CONFIG_HEAP_MEM_POOL_SIZE=1024
# LTE link control
CONFIG_LTE_LINK_CONTROL=y
CONFIG_ADC=y
CONFIG_ADC_0=y
CONFIG_ADC_NRFX_SAADC=y
CONFIG_MAIN_STACK_SIZE=4096
CONFIG_SPM=n

devzone.nordicsemi.com/.../main.c devzone.nordicsemi.com/.../CMakeLists.txt

RE: troubleshooting of produced prototype: nrf9160

Didrik Rokhaug — Wed, 08 Jan 2020 14:17:08 GMT

Hi, and sorry for the long wait.

I have talked to one of our hardware experts, and there are some differences between your design and our reference designs:

You are missing a 47uF capacitor on IO_supplyu/DEC0.

The recommended value for the DEC0 decoupling capacitor has been increased to 47 uF. This is done to improve low-frequency clock accuracy which improves protocol robustness in lower limit temperature conditions. No issues have been observed in practical use on boards with lower values, but 47 uF is expected to perform better over conditions and production variation.

When it comes to your SIM interface, the differences between your SIM interface and ours is that C17 = 1uF instead of 100nF and C16 = 1uF instead of 10nF.

Do these changes solve your problem?

Best regards,

Didrik

RE: troubleshooting of produced prototype: nrf9160

Didrik Rokhaug — Fri, 20 Dec 2019 12:11:23 GMT

Could you also share the rest of your schematics?

It is very limited what we can do with what you provided.

RE: troubleshooting of produced prototype: nrf9160

David — Fri, 20 Dec 2019 08:01:47 GMT

RE: troubleshooting of produced prototype: nrf9160

Didrik Rokhaug — Thu, 19 Dec 2019 15:06:12 GMT

Hi.

As the problem only occurs on some of your boards, I suspect there is a HW problem.

If you can share some schematics, I will try to find someone that can look at them to see if there is anything wrong.

Best regards,

Didrik