SPI freezing in initialization, while SCK connected to GND

[exet]

Hello, dear developers!

Tell me the solution to the problem, how to properly initialize the SPI peripherals so that the program does not hang while waiting for the end of data transfer, if the SCK and GND lines are short-circuited.

This is the initialization SPI

    nrf_spi_mngr_transaction_t internal_transaction = {
        .begin_callback      = spi_transaction_begin_cb,
        .end_callback        = spi_transaction_end_cb,
        .p_user_data         = (void *)&cb_data,
        .p_transfers         = p_transfers,
        .number_of_transfers = number_of_transfers,
        .p_required_spi_cfg  = p_config
    };

The interrupt callback at the beginning of the data transfer works, but the callback at the end of the data transfer does not work if the SCK line is shorted to ground. How can I fix this so that at least an error is generated or a timeout is triggered?

[Kenneth]

The clock signal is routed back internally to clock data on miso/mosi, so if you ground clock pin, then this will not occur and you will not get a callback at the end (since no data have been clocked). There is no direct fix to this, since it's hardware implementation, other than to ensure that the clock line is not grounded.

Kenneth

[exet]

Thanks for the answer. And why does the SDK not provide for handling such an event as a timeout so that the program does not stop in an endless loop waiting for the end of the transaction? After all, the rest of the program should continue to work ...

[Kenneth]

I assume it's a trade-off between code size, resources and still be somewhat pragmatic. It's something that I can see happen in a prototype development situation, I don't see a timeout would help in an end product, it would only happen if there is an electrical failure, in which case a timeout wouldn't really solve anything.

Kenneth