SPI clock issue

RE: SPI clock issue

Arthur — Tue, 31 Jan 2017 12:41:23 GMT

I succeed to get the answer from my MPU with your advices !! Thank you

RE: SPI clock issue

bjorn-spockeli — Tue, 31 Jan 2017 09:33:05 GMT

@Arthur: If your using a MPUxxxx accelerometer then I suggest that you take a look at this GitHub repo made by a colleauge, it contains SPI and TWI examples for MPUxxxx devices

RE: SPI clock issue

artis — Mon, 30 Jan 2017 19:13:40 GMT

I ran into the same problem. I think that the used interface is misleading (it suggests that rx lengths and tx lengths are independent). The best thing you can do is have your tx length and rx length the same. So if your command is 1 byte long and you are reading 2 bytes, make your tx and rx array 3 long. The command should be at tx[0]; the response will be available at rx[1] and rx[2].

RE: SPI clock issue

Arthur — Mon, 30 Jan 2017 16:41:06 GMT

I am using nRF52 DK with SPI_USE_EASY_DMA set to 1

RE: SPI clock issue

Arthur — Mon, 30 Jan 2017 16:40:33 GMT

I replaced this :

static uint8_t       m_tx_buf[1] = {0xF5};
nrf_drv_spi_transfer(&spi, m_tx_buf		, 1, m_rx_data_spi	 , 1);

with this :

static uint8_t       m_tx_buf[2] = {0xF5, 0x00};
nrf_drv_spi_transfer(&spi, m_tx_buf		, 2, m_rx_data_spi	 , 1);

and know I get the correct signals in the oscilloscope.

But m_rx_data_spi remain always empty.

RE: SPI clock issue

Wojtek — Mon, 30 Jan 2017 14:23:55 GMT

would be nice if you state what device You are using. If it is nRF52, this may be something with SPI DMA 1 byte transfer bug in the chip? You could check the errata to confirm that.