[nRF5 SDK v17.0.2] SPI with MAX31865 not receiving

[kmcn]

Hi all,

I'm trying to use the nRF52840-DK board to interface with the MAX31865 RTD-to-Digital Converter. I've adapted the MAX31865 C++ library to C and replaced the Arduino SPI functions with the Nordic functions. Currently I am unable to receive back any value from the sensor even though the SPI transfer is completed.

The initialisation of SPI and pins is as follow:

/* Defines - SPI -----------------------------------------------------------*/
#define SPI_INSTANCE  0 /**< SPI instance index. */
static const nrf_drv_spi_t spi_temp1 = NRF_DRV_SPI_INSTANCE(SPI_INSTANCE);  /**< SPI instance. */
static volatile bool spi_xfer_done;  /**< Flag used to indicate that SPI instance completed the transfer. */

void max31865_init(void) {

    nrf_drv_spi_config_t spi_config = NRF_DRV_SPI_DEFAULT_CONFIG;
    spi_config.ss_pin   = SPI_SS_PIN;
    spi_config.miso_pin = SPI_MISO_PIN;
    spi_config.mosi_pin = SPI_MOSI_PIN;
    spi_config.sck_pin  = SPI_SCK_PIN;
    spi_config.bit_order = NRF_DRV_SPI_BIT_ORDER_MSB_FIRST;
    spi_config.frequency = NRF_DRV_SPI_FREQ_2M;
    spi_config.mode = NRF_DRV_SPI_MODE_3;

    APP_ERROR_CHECK(nrf_drv_spi_init(&spi_temp1, &spi_config, spi_event_handler, NULL));

    NRF_LOG_INFO("SPI example started.");

}

SPI Transfer:

nrf_drv_spi_xfer_desc_t xfer_desc;
xfer_desc.p_tx_buffer = &address;
xfer_desc.tx_length = sizeof(address);
xfer_desc.p_rx_buffer = rxBuffer;
xfer_desc.rx_length = n;

APP_ERROR_CHECK(nrf_drv_spi_xfer(&spi_temp1, &xfer_desc, 0));

Is there anything significant missing in the initialisation step or something to look as to why nRF doesn't receive anything back?

Thanks

[haakonsh]

Currently I am unable to receive back any value from the sensor even though the SPI transfer is completed.

 What do you mean? Is the RX buffer empty?

Do you have a scope of the SPI communication?

[kmcn]

Hi,

Yes I've looked at the SPI comms using an oscilloscope and there is no output from the MOSI/MISO pins and the clock waveform is not square. The RX buffer is empty and it appears as if the pins are not being initiated. 

[haakonsh]

Can you show me a scope?

[kmcn]

Hi,

This is the best signal I get from the oscilloscope currently.

MOSI:

CLK:

[haakonsh]

The SPI Lines are way too capacitively loaded, are you breadboarding or using long jumper wires? I think you need to improve the grounding significantly. 

I also suggest that you enable high drive strength for the SPI pins. See nrf_gpio_cfg and nrf_gpio_pin_drive_t.

Use NRF_GPIO_PIN_H0H1. 

[kmcn]

Hi again,

I've been able to get a hold of a logic analyser to look the SPI in better detail. The below pictures compare the difference between the SPI comms from the Arduino and the SPI from the Nordic. Both SPI have been set with the same clock frequency(1M) and mode (CPOL 0, CPHA 1). I've tried enabling the MOSI and MISO pins to a high drive strength with NRF_GPIO_PIN_H0H1 and am still not able to receive back data on the MOSI line for the Nordic.

The MOSI lines are different between the devices but I have not been able to determine why. 

Arduino:

Nordic:

Thanks

[haakonsh]

We need to verify the HW before we start digging into the SW (which seems to be correctly initialized).

What is your HW platform?
Are you using the nRF52840DK with jumper wires to the Maxim device?

[kmcn]

Hi,

Yes I'm using the nRF52840DK with jumper wires to the Maxim device on a breadboard. Nothing else is connected to the breadboard.

[kmcn]

Hi here is a picture of the hardware set up. 

Red - 5V / Power

Black - GND 

Grey - SDI/MOSI - 29

Purple - SDO/MISO - 30

Blue - CS - 31

White - CLK - 26

The RDY pin on MAX31865 is NC as it is NC on the working Arduino set up. The MAX31865 breakout board is 5V compliant and has a voltage regulator on board to drop the power supply to 3.3V.

(The other wires on the breakout board are the Logic analyser).

[KM123]

Hi kmcn

Is your problem solved. I am facing similar type of issue with max31865.
if you have solved then please help me.

thanks

[kmcn]

Hi KM123

I was able to fix my issue with the MAX31865 development board! I ended up having to hardcode in pulling the CS GPIO pin high or low throughout the code where ever a SPI write or read occurred. I am currently waiting on a PCB delivery to test this code with the MAX31865 chip itself. 

I hope this helps.