NFC code not working

[Gordon]

Hi,

I'm trying to get NFC working on a PCA10040 dev board. Using the nfc/record_url hex file it works fine, so the hardware is obviously ok.

However, when I use the exact same code as in the record_url example from my software, it doesn't do anything. All functions execute without errors, but then log_uart_printf gets called with:

[NFC_HAL]: initialize
[NFC_HAL]: start
// ------------- put NFC device near 
[NFC_HAL]: fielddetected
[NFC_HAL]: reinitialize
[NFC_HAL]: fielddetected
[NFC_HAL]: fielddetected
[NFC_HAL]: reinitialize
[NFC_HAL]: fielddetected
[NFC_HAL]: reinitialize
[NFC_HAL]: fielddetected
[NFC_HAL]: reinitialize
[NFC_HAL]: fielddetected
[NFC_HAL]: fielddetected
[NFC_HAL]: fielddetected
[NFC_HAL]: fielddetected
[NFC_HAL]: reinitialize
[NFC_HAL]: fielddetected
[NFC_HAL]: reinitialize

Also, will the NFC code ever be distributed as something that's not a binary blob? I'm using a language interpreter (Espruino), and it expects functions to be compiled without hardware floating point (the call style is different). Currently the only way I can use NFC is to recompile with VFP, which means I'm then unable to pass floating point values into any functions!

[Gordon]

It wasn't defined, but I just added it - and still have the problem. Thanks for looking into this - compiling should be fine, just get the latest code from GitHub and NFC=1 NRF52832DK=1 make flash under Linux. You might also want to add return true to the start of jshSleep in targets/nrf5x/jshardware.c to stop it sleeping (but that doesn't make any difference either). If you have a word with Michael Dietz, he's done a lot of work on Espruino and could build it easily. Once running, connect to the UART at 9600 baud and type NRF.nfcURL("espruino.com") - that should start up NFC (I had experimented with using the NFC example code directly, and at boot - but there was no difference so I kept my original implementation).

[Hung Bui]

@Gordon: I would prefer to keep it simple, like try to keep it as close to the nrf_url_record as possible. We can strip down the makefile from github, compare it with the makefile from the SDK and check what make the difference.

I'm a little bit short on time for today and tomorrow, but if you have the setup ready I don't think it would take too much time to compare and figure out what make the difference as I can see the makefile for record_url is pretty simple.

[Gordon]

Ok, check out github.com/.../nordic_stripped_nfc and just run make. It'll build a binary+softdevice you can copy over (LED2 shows its alive, the rest is std NFC example in targets/nrf5x/main.c). I stripped a bunch out of the Makefile so it should be a lot easier to look through now.

[Hung Bui]

Hi Gordon,

Thanks for making the repo, I had some issue with the dependencies and path I need to fix but I will try to compile it here.

In the mean time could you try to help debugging the code by checking if the NFC peripheral is actually activated with the call NRF_NFCT->TASKS_ACTIVATE = 1; in nrf_nfct_clock_event_handler(). This only executed if the HFCLK is started (in function nrf_nfct_field_event_handler()).

From your log, we can spot that the SELECTED event supposed to come after "fielddetected" but it didn't come. We suspect for some reason it's not activated. If the NRF_NFCT->TASKS_ACTIVATE = 1; is called but "selected" is not arrived, you can try to turn on interrupt on READY event and AUTOCOLRESST and COLLISION event. These are the events supposed to come before we get the SELECTED event.

SELECTED event tells that the tag has been selected by the reader and the communication between them are fine.

[Gordon]

I just tried turning on an LED when NRF_NFCT->TASKS_ACTIVATE is called, but it looks like it never gets there - if that helps?