ask about this BUG the SPIM: An additional byte is clocked out when RXD.MAXCNT = 1

RE: ask about this BUG the SPIM: An additional byte is clocked out when RXD.MAXCNT = 1

leonhart88 — Sun, 15 Apr 2018 17:20:22 GMT

I just want to add here that if you're just trying to send a single byte out to a peripheral in order to set a register or something, you can just do the following:

nrf_drv_spi_transfer(&m_spi_master_0, spiTransmitBuffer, 1, NULL, 0);

Not sure why this isn't documented well anywhere.

RE: ask about this BUG the SPIM: An additional byte is clocked out when RXD.MAXCNT = 1

Chitrang Talaviya — Wed, 03 Jan 2018 19:42:59 GMT

I am also facing the same issue. I added the workaround code in SDK 14. Iam using nrf52840. I am sending 0x8F to read the WHO_AM_I and I am receiving 0x00 instead of 0x6A.

RE: ask about this BUG the SPIM: An additional byte is clocked out when RXD.MAXCNT = 1

victor_CHINA — Mon, 11 Dec 2017 15:07:20 GMT

hello Sigurd

12 02 33 44 is the slave device ID（which is fixed）

1、now ，I cannot gei the right result yet after add the ppi and gpiote（MY sdk is 13.1，and I add the code according to the spi_pan58_workaround_sdk12.zip，all code between ### Workaround Section Start and ### Workaround Section end in mian.c）

so do you have a project which is spi ok（with soft device is better ）？

I have the 《spi_pan58_workaround_sdk12.zip》

but I cannot build it successfully ，it has some mistake

2、I saw here （when send one byte） nrf_drv_gpiote_in_event_enable(SPI_SCK_PIN, false);// Enable the GPIOTE output pin task. the false is right ？should I change it to be ture？

best regards for you！

RE: ask about this BUG the SPIM: An additional byte is clocked out when RXD.MAXCNT = 1

Sigurd — Sun, 10 Dec 2017 22:50:59 GMT

Ok. Great that you no longer are getting 00 00 00 00 after changing to the correct frequency .

So you are now getting FF 12 02 33 , but it should be 12 02 33 44 ?

What device is the spi-slave? How do you know that you should get 12 02 33 44 instead?

RE: ask about this BUG the SPIM: An additional byte is clocked out when RXD.MAXCNT = 1

victor_CHINA — Sat, 09 Dec 2017 15:40:15 GMT

Hi Sigurd

today I tried

the reason why I just add the spi handle（even is NULL） the result turn to be 00 00 00 00

maybe is because of the frequency

my spi frequency is 500khz before，when I change the frequency to be 1MHZ，even if I add the spi handle ，the result is FF 12 02 33 not 00 00 00 00

RE: ask about this BUG the SPIM: An additional byte is clocked out when RXD.MAXCNT = 1

victor_CHINA — Fri, 08 Dec 2017 15:52:09 GMT

You mean It is because I have not set the spi _xfer_done true maybe?

I will try the example handle without change tomorrow,thank you

RE: ask about this BUG the SPIM: An additional byte is clocked out when RXD.MAXCNT = 1

Sigurd — Fri, 08 Dec 2017 15:39:09 GMT

In the example in the SDK, the spi_event_handler is used to set the spi_xfer_done flag to true.

RE: ask about this BUG the SPIM: An additional byte is clocked out when RXD.MAXCNT = 1

victor_CHINA — Fri, 08 Dec 2017 15:30:54 GMT

Sorry

The SDK I tried spi handle is 13.1

RE: ask about this BUG the SPIM: An additional byte is clocked out when RXD.MAXCNT = 1

victor_CHINA — Fri, 08 Dec 2017 15:29:02 GMT

May be it is because I have not clear some flag in the handle?

RE: ask about this BUG the SPIM: An additional byte is clocked out when RXD.MAXCNT = 1

victor_CHINA — Fri, 08 Dec 2017 15:26:13 GMT

like this

void spi_event_handler(nrf_drv_spi_evt_t const * p_event,void * p_context) { //do nothing }

RE: ask about this BUG the SPIM: An additional byte is clocked out when RXD.MAXCNT = 1

Sigurd — Fri, 08 Dec 2017 15:22:36 GMT

How does your spi_event_handler function looks like ?

RE: ask about this BUG the SPIM: An additional byte is clocked out when RXD.MAXCNT = 1

victor_CHINA — Fri, 08 Dec 2017 15:03:15 GMT

and now my mainly question is :why I just add a spi handle function,the result is turn to be 00 00 00 00

RE: ask about this BUG the SPIM: An additional byte is clocked out when RXD.MAXCNT = 1

victor_CHINA — Fri, 08 Dec 2017 15:00:00 GMT

thanks 1、I know that，I will not put the example to the sdk 14 dirictly

2、this problem is that I tryed in the spi example (sdk14)

I only add the spi handle，with out ppi and gpiote

I think this problem is important，It may be the reason why spi out 00 00 00 00 when I add the ppi and gpiote code

thank you

RE: ask about this BUG the SPIM: An additional byte is clocked out when RXD.MAXCNT = 1

Sigurd — Fri, 08 Dec 2017 14:51:14 GMT

The example project was for SDK 12. I.e. if you just copy the example to SDK 14, it will not compile. So you need to port the code to the SDK 14 SPI example.
Your init function looks correct. You need to add the PPI and GPIOTE code for the fix to work. In the example project with the fix, we disable the PPI and GPIOTE in the spi_event_handler when we get the event NRF_DRV_SPI_EVENT_DONE.

RE: ask about this BUG the SPIM: An additional byte is clocked out when RXD.MAXCNT = 1

victor_CHINA — Thu, 07 Dec 2017 13:56:47 GMT

@Sigurd。 hi,can you hear me? I need your help ,thank you

RE: ask about this BUG the SPIM: An additional byte is clocked out when RXD.MAXCNT = 1

victor_CHINA — Thu, 07 Dec 2017 10:37:40 GMT

thank you for your code

but I have two more questions

1、I haven't see any difference between your code and the spi_pan58_workaround_sdk12.zip（about the spi），so what do you mean spi_pan58_workaround_sdk12.zip is just for sdk12？

what is the difference between the spi_pan58_workaround_sdk12.zip and you code above？

2、today I just try init like this : nrf_drv_spi_init(&spi, &spi_config, spi_event_handler, NULL);

haven't added the PPI and gpiote code yet

only add the spi_event_handler

the result turn to be ：00 00 00 00（right is 12 02 33 44）,even the spi_event_handler is NULL Like

this:void spi_event_handler(nrf_drv_spi_evt_t const * p_event,void * p_context) { //do nothing }

when I remove the spi_event_handler,like this :nrf_drv_spi_init(&spi, &spi_config, NULL, NULL)，the result is turn to be FF 12 02 33

my question is：why the spi_event_handler can influence the result

thank you!

RE: ask about this BUG the SPIM: An additional byte is clocked out when RXD.MAXCNT = 1

Sigurd — Wed, 06 Dec 2017 16:20:39 GMT

It should be shown in the main.c file in the spi_pan58_workaround_sdk12.zip in the post you linked to. Please take a look at that file.

I.e.

Add this with the other defines:

// ### Workaround Section Start ###
#include "nrf_drv_ppi.h"
#include "nrf_drv_gpiote.h"
// ### Workaround Section End   ###

// ### Workaround Section Start ###
static nrf_ppi_channel_t ppi_channel;
//Dummy handler
void in_pin_handler( nrf_drv_gpiote_pin_t pin, nrf_gpiote_polarity_t action)
{
    //Do nothing
}
// ### Workaround Section End   ###

in spi_event_handler, add this at right after NRF_DRV_SPI_EVENT_DONE

 // ### Workaround Section Start ###

// Disable the PPI channel
err_code = nrf_drv_ppi_channel_disable(ppi_channel);
APP_ERROR_CHECK(err_code);

// Disable a GPIOTE output pin task.
nrf_drv_gpiote_in_event_disable(SPI_SCK_PIN);

// ### Workaround Section End   ###

after APP_ERROR_CHECK(nrf_drv_spi_init(&spi, &spi_config, spi_event_handler, NULL)); in main(), you add this:

// ### Workaround Section Start ###
    
     uint32_t err_code;
    
     NRF_SPIM_Type * p_spim = spi.p_registers;
    
    //Initialize GPIOTE and PPI drivers for the workaround
    
    err_code = nrf_drv_gpiote_init();
    APP_ERROR_CHECK(err_code);
    
    // Initializes the GPIOTE channel so that SCK toggles generates events
    nrf_drv_gpiote_in_config_t config = GPIOTE_CONFIG_IN_SENSE_TOGGLE(true);
            err_code = nrf_drv_gpiote_in_init(SPI_SCK_PIN, &config, in_pin_handler);
            APP_ERROR_CHECK(err_code); 
        
    err_code = nrf_drv_ppi_init();
    APP_ERROR_CHECK(err_code);
    
    // Allocate first unused PPI channel 
    err_code = nrf_drv_ppi_channel_alloc(&ppi_channel);
    APP_ERROR_CHECK(err_code);
    
    // Assign the PPI channel to the SCK pin
    err_code = nrf_drv_ppi_channel_assign(ppi_channel,
                                        nrf_drv_gpiote_in_event_addr_get(SPI_SCK_PIN),
                                        (uint32_t)&p_spim->TASKS_STOP);
    APP_ERROR_CHECK(err_code);


// ### Workaround Section End   ###

Then use the function spi_send_recv when you want to send 1 byte data

RE: ask about this BUG the SPIM: An additional byte is clocked out when RXD.MAXCNT = 1

victor_CHINA — Wed, 06 Dec 2017 15:28:57 GMT

should I put the code between the start and end？

and where should I put these all source code？

thank you

RE: ask about this BUG the SPIM: An additional byte is clocked out when RXD.MAXCNT = 1

victor_CHINA — Wed, 06 Dec 2017 15:25:28 GMT

I saw the workaround code in the nordic error document，but I donot know where should I put these code

RE: ask about this BUG the SPIM: An additional byte is clocked out when RXD.MAXCNT = 1

Sigurd — Wed, 06 Dec 2017 14:55:32 GMT

The example with the fix is for SDK 12. Did you add the relevant fix to the SDK 14 SPI example ?

You need to add the ### Workaround Section Start ### and### Workaround Section End ### code to the correct places.

Then use spi_send_recv() function when you want to send single byte.

RE: ask about this BUG the SPIM: An additional byte is clocked out when RXD.MAXCNT = 1

victor_CHINA — Wed, 06 Dec 2017 14:40:29 GMT

hello？ can you answer me？ thank you

RE: ask about this BUG the SPIM: An additional byte is clocked out when RXD.MAXCNT = 1

victor_CHINA — Wed, 06 Dec 2017 09:02:27 GMT

SDK 14.2.0 I get 00 12 02 33 before I added the workaround thank you

RE: ask about this BUG the SPIM: An additional byte is clocked out when RXD.MAXCNT = 1

Sigurd — Wed, 06 Dec 2017 09:00:03 GMT

Hi,

What SDK version are you using ?
What was the result of the read operation before you added the workaround ? Did you get 00 00 00 00 before you added the workaround also? or did you get something else?