SPI Master always sends 16 bits words in non-continuous mode

[trevorhey]

Hi,

Anyone come across such a problem?  I am not using the KSDK libraries, but used them to model my own driver to trim down the code size.

I set up SPI master mode on SPI1. The code below is to show the settings. I am interfacing to a SPANSION SPI memory device, really simple and basic SPI device.

The problem comes in with the setting of the CONT_SCKE bit in MCR.

If I clear this bit (not continuous mode) , the SPI master clocks out 16 bits for each transfer written to PUSHR, if I set it, it clocks out 8 bits?  That should not happen.

All other settings are kept the same for this test, I change only this one bit and see this behaviour.  I use CTAR  0 in all cases.

Why does the SPI module clock out 8 or a6 bits when I change this one setting? Of course I dont want continuous mode.

Thanks for any help

Device = MK24F25612

My SPi init:  Info: SPI1_CTAR_USR = 0

SPI1->MCR |= SPI_MCR_MSTR_MASK;  // Master Mode

//SPI1->MCR |= SPI_MCR_CONT_SCKE_MASK; // clock runs continuosly

SPI1->MCR &= ~SPI_MCR_CONT_SCKE_MASK; // clock doesnt run continuosly

SPI1->MCR |=  1 << SPI_MCR_PCSIS_SHIFT ; // Active low CS

SPI1->MCR &= ~SPI_MCR_DIS_RXF_MASK;

SPI1->MCR &= ~SPI_MCR_DIS_TXF_MASK; // enable FIFOs

// CS pre-delay  ---- must be > 3 ns, SPANSION clock is 9ns min so do at least 1 clock delay is sufficient but Ill do 8

SPI1->CTAR[SPI1_CTAR_USR] &= ~SPI_CTAR_PCSSCK_MASK;

SPI1->CTAR[SPI1_CTAR_USR] |= SPI_CTAR_PCSSCK(1);

SPI1->CTAR[SPI1_CTAR_USR] &= ~SPI_CTAR_CSSCK_MASK;

SPI1->CTAR[SPI1_CTAR_USR] |= SPI_CTAR_CSSCK(8); 

// min CS high time  50ns for SPANSION so at 9ns clock we need @ least 6 clocks, choose 16

SPI1->CTAR[SPI1_CTAR_USR] &= ~SPI_CTAR_PASC_MASK;

SPI1->CTAR[SPI1_CTAR_USR] |= SPI_CTAR_PASC(1); 

SPI1->CTAR[SPI1_CTAR_USR] &= ~SPI_CTAR_ASC_MASK;

SPI1->CTAR[SPI1_CTAR_USR] |= SPI_CTAR_ASC(16);

//baud rate SCK baud rate = (fP/PBR) x [(1+DBR)/BR]

retBaudrate = CalcSPIBaudRate( port,  baudrate, clock);

SPI1->CTAR[SPI1_CTAR_USR] &= ~SPI_CTAR_FMSZ_MASK; // bits per frame = bits-1

SPI1->CTAR[SPI1_CTAR_USR] |= SPI_CTAR_FMSZ(7); //

SPI1->CTAR[SPI1_CTAR_USR] &= ~SPI_CTAR_CPOL_MASK; // Using Mode 0 - CLK inactive state is low CPOL = 0

  SPI1->CTAR[SPI1_CTAR_USR] &= ~SPI_CTAR_CPHA_MASK; // Using Mode 0 - CData is captured on the leading edge of SCK and changed on the following edge CPHA = 0

// SPI1->CTAR[SPI1_CTAR_USR] &= ~SPI_CTAR_CPOL_MASK; // Using Mode 0 - CLK inactive state is low CPOL = 0

//SPI1->CTAR[SPI1_CTAR_USR] |= SPI_CTAR_CPHA_MASK; // Using Mode 0 - CData is captured on the leading edge of SCK and changed on the following edge CPHA = 0

SPI1->CTAR[SPI1_CTAR_USR] &= ~SPI_CTAR_LSBFE_MASK; // bit direction -SPANSION neebs msb first

NVIC_EnableIRQ(SPI1_IRQn);

// Last thing - start it up

SPI1->MCR &= ~SPI_MCR_MDIS_MASK;  // Zero enables SPI

How I loaf the TX fifo:

void fs_SPI_WriteDataMastermode(SPI_Type * port,  dspi_command_config_t * command,   uint16_t data) {

    uint32_t temp = 0;

    /* First, build up the 32-bit word then write it to the PUSHR */

if(command->isChipSelectContinuous) temp |= SPI_PUSHR_CONT_MASK;

temp |= SPI_PUSHR_CTAS(command->whichCtar);

temp |= SPI_PUSHR_PCS(command->whichPcs);

if (command->isEndOfQueue) temp |=  SPI_PUSHR_EOQ_MASK;

if (command->clearTransferCount) temp  |= SPI_PUSHR_CTCNT_MASK;

temp |= SPI_PUSHR_TXDATA(data);

SPI_PUSHR_REG(port) = temp;

}

[kerryzhou]

Hi Trevor Hey,

    Thank you for you interest in kinetis product.

    Please read the K22 reference manual chapter 49.4.4.3 Continuous Selection Format carefully.

    I see your code, when you work in continuous mode:

if(command->isChipSelectContinuous) temp |= SPI_PUSHR_CONT_MASK;

    You set CONT bit,  in the datasheet :

    When the CONT bit = 1, the PCS signal remains asserted for the duration of the two transfers.

    If you send 2 8bit byte, then it will seem that you are send 16 bits, because it isn't insert the tDT in PCSx.

    About details, please read our reference manual carefully.

Wish it helps you!

If you still have question, please contact me!

Have a great day,
Jingjing

-----------------------------------------------------------------------------------------------------------------------
Note: If this post answers your question, please click the Correct Answer button. Thank you!
-----------------------------------------------------------------------------------------------------------------------

[egoodii]

You left out a couple important details:

What is the final baud-rate control contents?

What is a 'representative' content of the 'command' structure in the write-functoin call?  Can we assume you NEVER set 'isChipSelectContinuous' in the continuous-clock mode?