Reading from slave chip (ADS1293) through SPI

RE: Reading from slave chip (ADS1293) through SPI

Simonr — Thu, 04 May 2023 12:37:21 GMT

This case is close to 4 years old at this point, so the chances you'll get a reply from anyone active in this ticket are not likely to reply. If you need technical support I'd recommend creating a separate ticket where you describe your problem in detail and it will be assigned to the correct person. You can link to this case if you think it's relevant as well.

Best regards,

Simon

RE: Reading from slave chip (ADS1293) through SPI

NeMo93 — Thu, 04 May 2023 10:00:21 GMT

Hi, i am facing a similar problem while trying to set up the SPI connection between ads1293 and nrf52840. Did you figure out how to solve this issue?

RE: Reading from slave chip (ADS1293) through SPI

Konrad213 — Fri, 12 Jul 2019 19:47:07 GMT

Thanks again! I changed to DC and now I see that everything is the same for the Arduino and Nordic chip except for the SDO on the slave chip, nothing gets sent back from the ads for some reason when using the Nordic chip.

RE: Reading from slave chip (ADS1293) through SPI

hmolesworth — Fri, 12 Jul 2019 18:15:42 GMT

You have the 'scope incorrectly set to the AC setting on both probes; change to DC ..for the slower clock speed also slow down the timebase to see the whole waveform. 50mSecs seems already very slow, what speed are the SPI clocks?

The Arduino is doing 2 separate SPI transactions of 1 byte each; you can do the same on the nRF by sending 2 messages of 1 byte each; the first would be the command and the second would clock out the response. Inefficient, but then you compare like with like and can increase the clock back to the same as Arduino.

RE: Reading from slave chip (ADS1293) through SPI

Konrad213 — Fri, 12 Jul 2019 17:32:48 GMT

Thanks! I removed the manual ss controls and lowered the speed. Now I get this on the oscilloscope although the execution still gets stuck in the same place:

Here is what it looks like on the Arduino that works:

RE: Reading from slave chip (ADS1293) through SPI

hmolesworth — Fri, 12 Jul 2019 14:58:20 GMT

I think there is a misunderstanding about the use of the cs pin; either the spim manager handles it or your code does, not both. Here you tell the manager to control cs but then you take over:

        .ss_pin         = 3,
    };
    ret_code_t err_code = nrfx_spim_init(&m_spi_master_0, &spi_config, NULL, NULL);

  nrf_gpio_pin_clear(3);
  ret_code_t err_code = nrfx_spim_xfer(&m_spi_master_0, &xfer_desc, 0);
  nrf_gpio_pin_set(3);

Either do this:

        .ss_pin         = 3,
    ret_code_t err_code = nrfx_spim_init(&m_spi_master_0, &spi_config, NULL, NULL);

//nrf_gpio_pin_clear(3);
  ret_code_t err_code = nrfx_spim_xfer(&m_spi_master_0, &xfer_desc, 0);
//nrf_gpio_pin_set(3);

or this:

        .ss_pin         = NRF_DRV_SPI_PIN_NOT_USED, // ie 0xFF

  nrf_gpio_pin_clear(3);
  ret_code_t err_code = nrfx_spim_xfer(&m_spi_master_0, &xfer_desc, 0);
  nrf_gpio_pin_set(3);

I also think 4MHz is too high to start; once you have it working you can always try speeding up

RE: Reading from slave chip (ADS1293) through SPI

Konrad213 — Fri, 12 Jul 2019 06:39:14 GMT

I used an oscilloscope and changed the code slightly but still the same behavior as mentioned below. I used an arduino to talk to the ADS chip and it worked fine. It looks like the output to the slave select pin isn't happening at all (voltage stays same on oscilloscope) when using the nrf chip vs the arduino. Any ideas?

static uint8_t tx[2];
static uint8_t rx[2];

static const nrfx_spim_t m_spi_master_0 = NRFX_SPIM_INSTANCE(SPI_INSTANCE);

void ADS_init( void )
{
    nrfx_spim_config_t const spi_config =
    {
        .sck_pin        = 4,
        .mosi_pin       = 22,
        .miso_pin       = 23,
        .ss_pin         = 3,
        .irq_priority   = APP_IRQ_PRIORITY_LOW,
        .orc            = 0xFF,
        .frequency      = NRF_SPIM_FREQ_4M,
        .mode           = NRF_SPIM_MODE_2,
        .bit_order      = NRF_SPIM_BIT_ORDER_MSB_FIRST, 
    };
    ret_code_t err_code = nrfx_spim_init(&m_spi_master_0, &spi_config, NULL, NULL);
    SEGGER_RTT_printf(0, nrf_strerror_get(err_code));
     
}

uint8_t Read_reg(uint8_t addr)
{ 

  tx[0] = addr | ADS1293_READ_BIT; 
  tx[1] = 0x00; 
  nrfx_spim_xfer_desc_t xfer_desc = NRFX_SPIM_XFER_TRX(tx, 2, rx, 2);
  nrf_gpio_pin_clear(3);
  ret_code_t err_code = nrfx_spim_xfer(&m_spi_master_0, &xfer_desc, 0);
  nrf_gpio_pin_set(3);
  SEGGER_RTT_printf(0, nrf_strerror_get(err_code));
  return rx[1];

}

RE: Reading from slave chip (ADS1293) through SPI

Konrad213 — Fri, 12 Jul 2019 06:35:34 GMT

Thank you for the response! I've changed the code slightly but the issue is it getting stuck and not returning from the nrf_spim_event_check. Also as mentioned below the documentation for the 1293 differs from 1291/2.

RE: Reading from slave chip (ADS1293) through SPI

hmolesworth — Mon, 01 Jul 2019 23:36:19 GMT

Are you still using the code you posted above?

uint8_t Read_reg(uint8_t addr)
{ 
  tx = ADS1293_READ_BIT | addr;
  ret_code_t err_code = nrf_drv_spi_transfer(&m_spi_master_0, &tx, 1, &rx, 1);
  SEGGER_RTT_printf(0, nrf_strerror_get(err_code));
  return rx;
}

That will not work, as that code only reads and writes a single byte and so you never get to the registers ..

This is an example for the ADS1291/2:

// Scratch pad comms buffers, used for reading all AFE registers - want to preserve these for debugging
static uint8_t mTxBuf[] = {ADS_CMND_RREG, NUMBER_OF_ADS_REGISTERS-1, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00};
static uint8_t mRxBuf[sizeof(mTxBuf)];
static const uint8_t mRxTxBufLength = sizeof(mTxBuf);
STATIC_ASSERT(sizeof(mAFE_RegisterActualSettings)+2 == sizeof(mTxBuf)/sizeof(mTxBuf[0]));

#define AFE_HARDWARE_VERSION 0x53 // ADS1292

/**
 * @brief SPI command to read AFE registers
 * @param none
 */
static void ReadAfeRegisters(void)
{
    uint32_t spiAttempCount, mAfeSpiErrorCount = 0;

    for (spiAttempCount=0; spiAttempCount < AFE_MAX_ATTEMPTS; spiAttempCount++)
    {
        memset(mRxBuf, '?', mRxTxBufLength);
        mAfePacketTransferComplete = false;
        APP_ERROR_CHECK(nrf_drv_spi_transfer(&mAfeSpiInstance, mTxBuf, mRxTxBufLength, mRxBuf, mRxTxBufLength));
        // Hard delay per byte to transmit
        nrf_delay_us(mRxTxBufLength*AFE_SPI_PER_BYTE_uSECS);
        // Check for ADS1292 Id
        if (mAfePacketTransferComplete && (mRxBuf[2] == AFE_HARDWARE_VERSION))
        {
            // Correct Id for this AFE, post the results skipping cmd code & length bytes
            //memcpy(&mAFE_RegisterActualSettings, &mRxBuf[2], sizeof(mAFE_RegisterActualSettings));
            break;
        }
        else
        {
            mAfeSpiErrorCount++;
            // Extra hard delay per byte to transmit
            nrf_delay_us(mRxTxBufLength*AFE_SPI_PER_BYTE_uSECS);
        }
    }
    // Post the results (may be bad) skipping cmd code & length bytes
    memcpy(&mAFE_RegisterActualSettings, &mRxBuf[2], sizeof(mAFE_RegisterActualSettings));
}

My point isn't to use this code, but instead that there is nothing useful in the first returned byte and so a command requesting a single byte transfer isn't going to be useful. The first byte of the synchronous SPI transfer says what data should be returned on 2nd and successive bytes ..

RE: Reading from slave chip (ADS1293) through SPI

Konrad213 — Mon, 01 Jul 2019 21:22:16 GMT

Still haven't been able to solve it, the last function in the callstack is the following:

Any ideas?

RE: Reading from slave chip (ADS1293) through SPI

Konrad213 — Wed, 26 Jun 2019 10:44:42 GMT

I compared the documentations for the 1298 and 1293 and it looks like you're right about some of the things not being applicable (the wakeup commands etc.). I tried changing the frequency and SPI mode but that didn't help. Thanks either way!

RE: Reading from slave chip (ADS1293) through SPI

Konrad213 — Wed, 26 Jun 2019 10:43:18 GMT

Thanks for the answer! Won't have access to an oscilloscope for a couple of days, will do so as soon as I have but I'm hoping I can get past this issue before that. The execution gets stuck in this while loop (part of nrfx_spim.c):

 if (!p_cb->handler)
    {
        while (!nrf_spim_event_check(p_spim, NRF_SPIM_EVENT_END)){}

Here's nrf_spim_event_check (part of nrf_spim.h):

__STATIC_INLINE bool nrf_spim_event_check(NRF_SPIM_Type * p_reg,
                                          nrf_spim_event_t spim_event)
{
    return (bool)*(volatile uint32_t *)((uint8_t *)p_reg + (uint32_t)spim_event);
}

RE: Reading from slave chip (ADS1293) through SPI

hmolesworth — Tue, 25 Jun 2019 14:40:28 GMT

Edit: Looks like the ADS1293 is quite different (upgraded) from the ADS1291/2/8 so some of these comments may not apply.

There are a number of other issues with your settings. Start with a slower clock and SPI Mode 1:

// SPI INTERFACE
// The SPI-compatible serial interface consists of four signals: CS, SCLK, DIN, and DOUT. The interface reads
// conversion data, reads and writes registers, and controls ADS1291, ADS1292, and ADS1292R operation. The
// DRDY output is used as a status signal to indicate when data are ready. DRDY goes low when new data are
// available.
// Chip Select (CS)
// CS selects the ADS1291, ADS1292, and ADS1292R for SPI communication. CS must remain low for the entire
// duration of the serial communication. After the serial communication is finished, always wait four or more tCLK
// cycles before taking CS high. When CS is taken high, the serial interface is reset, SCLK and DIN are ignored,
// and DOUT enters a high-impedance state. DRDY asserts when data conversion is complete, regardless of
// whether CS is high or low

// Sending Multi-Byte Commands
// The ADS1291, ADS1292, and ADS1292R serial interface decodes commands in bytes and requires 4 tCLK cycles
// to decode and execute. Therefore, when sending multi-byte commands, a 4 tCLK period must separate the end of
// one byte (or opcode) and the next.
// Assume CLK is 512 kHz, then tSDECODE (4 tCLK) is 7.8125 us. When SCLK is 16 MHz, one byte can be
// transferred in 500 ns. This byte-transfer time does not meet the tSDECODE specification; therefore, a delay must be
// inserted so the end of the second byte arrives 7.3125 us later. If SCLK is 1 MHz, one byte is transferred in 8 �s.
// Because this transfer time exceeds the tSDECODE specification, the processor can send subsequent bytes without
// delay. In this later scenario, the serial port can be programmed to move from single-byte transfer per cycle to
// multiple bytes.
  
  spi_config.frequency = NRF_DRV_SPI_FREQ_500K;  // Ensure less than 4 clk cycles for ADS1291/2/8
  spi_config.bit_order = NRF_DRV_SPI_BIT_ORDER_MSB_FIRST;
  spi_config.mode      = NRF_DRV_SPI_MODE_1;

Note you need to stop continuous read mode before accessing registers by sending command ADS_CMND_SDATAC:

/****************************************************************/
/* ADS1x9x COMMAND DESCRIPTION and definitions */
/****************************************************************/
typedef enum {
  // System Commands
  ADS_CMND_WAKEUP    = 0x02,   // Wake-up from standby mode
  ADS_CMND_STANDBY   = 0x04,   // Enter standby mode
  ADS_CMND_RESET_CMD = 0x06,   // Reset the device registers
  ADS_CMND_START     = 0x08,   // Start/restart (synchronize) conversions
  ADS_CMND_STOP      = 0x0A,   // Stop conversion
  ADS_CMND_OFFSETCAL = 0x1A,   // Channel offset calibration - needs to be sent every time there is a change to the PGA gain
  // Data Read Commands
  ADS_CMND_RDATAC    = 0x10,   // Enable Read Data Continuous mode.
                               // - This mode is the default mode at power-up.
  ADS_CMND_SDATAC    = 0x11,   // Stop Read Data Continuously mode
  ADS_CMND_RDATA     = 0x12,   // Read data by command; supports multiple read back.
  // Register Read/Write Commands
  ADS_CMND_RREG      = 0x20,   // Read n nnnn registers starting at address r rrrr
                               //  - first byte 001r rrrr (2xh)(2) - second byte 000n nnnn(2)
  ADS_CMND_WREG      = 0x40    // Write n nnnn registers starting at address r rrrr
                               //  - first byte 010r rrrr (2xh)(2) - second byte 000n nnnn(2)
} ADS1x9xCommand_t;

Have a look at the answers we provided in this link: interface-an-ads1298-via-spi

RE: Reading from slave chip (ADS1293) through SPI

awneil — Tue, 25 Jun 2019 14:04:49 GMT

Again, have you used an oscilloscope to see what's happening on the wires?

Have you used the debugger to find exactly where it's getting stuck within the software?

RE: Reading from slave chip (ADS1293) through SPI

Konrad213 — Tue, 25 Jun 2019 13:55:21 GMT

Thank you! Now the execution gets stuck at nrf_drv_spi_transfer which never returns anything do you know any potential reasons for that?

RE: Reading from slave chip (ADS1293) through SPI

Simonr — Tue, 25 Jun 2019 12:59:35 GMT

It seems like you've mixed up the SPI pin configuration, as you should refer to the pin configuration chapter when configuring pins, and not the pin assignments. The sck_pin, mosi_pin, miso_pin, and ss_pin configured like this for the setup you're describing.

        .sck_pin        = 4,
        .mosi_pin       = 20,
        .miso_pin       = 19,
        .ss_pin         = 3,

Best regards,

Simon

RE: Reading from slave chip (ADS1293) through SPI

awneil — Tue, 25 Jun 2019 10:47:07 GMT

[quote userid="80983" url="~/f/nordic-q-a/48962/reading-from-slave-chip-ads1293-through-spi"]Is there any way to make sure that the pins I'm using are correct[/quote]

Get an oscilloscope and look at what's happening on the wires.