[NRF52810] SPI (without DMA) 's SCK signal isn't continuous

RE: [NRF52810] SPI (without DMA) 's SCK signal isn't continuous

Kenneth — Tue, 23 Aug 2022 07:18:43 GMT

Hello,

Without EasyDMA then after each byte is transmitted, then the SPI interrupt needs to execute, and next byte needs to be written from the SPI interrupt. This will cause a few us delay between each byte. If you need the SPI clock continuation then you need to use SPI with EasyDMA (SPIM_PRESENT), then the entire SPI transfer will occur in one go without any delay between bytes.

Kenneth

RE: [NRF52810] SPI (without DMA) 's SCK signal isn't continuous

Triumph — Tue, 23 Aug 2022 01:46:07 GMT

Hi ~

Thank you for your patience, I read the driver's code and I know the driver write one byte at a time, but I can't understand the clock have so big gap between two bytes. Is the SPI's CLK right in my pasted picture? Can I make the SPI clock continuation?

Regards

Triumph

RE: [NRF52810] SPI (without DMA) 's SCK signal isn't continuous

Kenneth — Fri, 19 Aug 2022 10:15:27 GMT

If you disable DMA, then the driver will write one byte at a time, so this is slower yes.

Kenneth

RE: [NRF52810] SPI (without DMA) 's SCK signal isn't continuous

Triumph — Fri, 19 Aug 2022 01:00:23 GMT

hello,

I delete the defination of SPIM and confirm that use the nrf_drv_spi_xfer() function.

//#define SPIM_PRESENT
//#define SPIM_COUNT 4

RE: [NRF52810] SPI (without DMA) 's SCK signal isn't continuous

Kenneth — Thu, 18 Aug 2022 11:20:02 GMT

Hello,

Is this the \nRF5_SDK_17.1.0_ddde560\examples\peripheral\spi example in nRF5 SDK?

Looking at the sdk_config.h file I can find that EasyDMA is enabled by SPI0_USE_EASY_DMA 1, so I don't see any reason why nrf_drv_spi_transfer() should have any delay between each byte.

Kenneth