Failed to read/write ADXL362 data using nRF52840

RE: Failed to read/write ADXL362 data using nRF52840

Marko W — Wed, 17 May 2023 06:20:55 GMT

I recently had the same problem with the ADXL362 on an nRF52832DK board. I connected the ADXL362 (on a developer board "ADXL362/346") with a short cable (about 5 cm).

After many attempts in all possible directions (clock, source code, waveform, resistances in signal lines), I found a solution:

I removed the cables and soldered a pin strip directly to the "ADXL362/346" developer board (without VIN, this is connected to VDD on P1 by cable) and plugged it directly into P4 (pin 0.19 to GND). Since this is plugged in directly, the ADXL362 works without any problems.

I don't find that plausible, but I was able to continue my project with it.

I'm not sure if it will help others, but it's worth a try.

And it would be nice to hear if this has helped others as well.

Best regards

Marko

RE: Failed to read/write ADXL362 data using nRF52840

Håkon Alseth — Mon, 27 Jul 2020 08:05:45 GMT

Hi,

If you create a new case, with logic traces, and explain your specific setup (which GPIOs you're using etc), we can help you out. This case is over 2 years old, and is likely outdated for your use-case (SDK used etc).

Kind regards,

Håkon

RE: Failed to read/write ADXL362 data using nRF52840

rkorn — Thu, 23 Jul 2020 15:54:14 GMT

Hi,

I'm experiencing the same issue with the ADXL362 and the NRF52840DK. Were you able to resolve the issue? I suspect this is an issue with the ADXL362, not the Nordic. I'm going to try breaking up the writes next.

RE: Failed to read/write ADXL362 data using nRF52840

Håkon Alseth — Thu, 29 Nov 2018 10:18:36 GMT

I cannot see any difference in the way the communication is performed, other than the nRF sends the 3 bytes without any delay.

You could try to split up the transfer into 1 byte transfers, like this:

        for (int i = 0; i < 3; i++){
            APP_ERROR_CHECK(nrf_drv_spi_transfer(&spi, &m_tx_buf[i], 1, &m_rx_buf[i], 1));
            spi_xfer_done = false;
            while (!spi_xfer_done)
            {
                __WFE();
            }
        }

However; the sensors datasheet states that this should work without delay in SPI communication.

Could you please provide a scope/logic analyzer trace of all SPI lines at the same time, with both arduino and the nRF solution? It is close to impossible to see whats happening without having all SPI lines in the same trace.

Kind regards,

Håkon

RE: Failed to read/write ADXL362 data using nRF52840

beparas — Thu, 29 Nov 2018 05:29:21 GMT

Hello Sir,

Thanks for replay.

1. Could you post your latest source code that performs the SPI read/write?

Ans: Please find spim_adxl362.zip file contains source code for reading 0x01 register.

3. Could you post the arduino code used?

For Arduino I follow this document

1. Open ADXL362_SimpleRead example sketch.

2. To read 0x01 register modify ADXL362 library. added readDevIDData() function in it. ADXL362.cpp file is attached here.

3. called readDevIDData() in loop.

Note: SDK version: nRF5_SDK_15.0.0_a53641a

Thanks in advance.

devzone.nordicsemi.com/.../4237.ADXL362.cpp devzone.nordicsemi.com/.../6747.ADXL362.h devzone.nordicsemi.com/.../3247.spim_5F00_adxl362.zip

RE: Failed to read/write ADXL362 data using nRF52840

Håkon Alseth — Wed, 28 Nov 2018 14:19:42 GMT

1. Could you post your latest source code that performs the SPI read/write?

2. Could you try only reading the 0x01 register. Do not run the Soft-reset or powerctl routine.

3. Could you post the arduino code used?

Kind regards,

Håkon

RE: Failed to read/write ADXL362 data using nRF52840

beparas — Mon, 26 Nov 2018 04:30:54 GMT

Hello Sir,

As per ADXL362, here is command structure for read register

" </CS down> <command byte (0x0A or 0x0B)> <address byte> <data byte> <additional data bytes for multi-byte> … </CS up>"

So I am sending same command for Arduino & nRF52840. The command is
0x0B 0x01 0x00

You can check this in above attached waveform.

Thanks in advance.

RE: Failed to read/write ADXL362 data using nRF52840

Håkon Alseth — Fri, 23 Nov 2018 11:35:16 GMT

Hi,

Judging by the scoping of CLK+MISO, the resonse is incorrect. This may indicate that there's a difference between the inputted data on MOSI.

Is there any difference in the MOSI data compared to the arduino?

Kind regards,

Håkon

RE: Failed to read/write ADXL362 data using nRF52840

beparas — Fri, 23 Nov 2018 09:50:05 GMT

Hello Sir,

Any update on this.

Waiting for replay.

RE: Failed to read/write ADXL362 data using nRF52840

beparas — Thu, 22 Nov 2018 07:18:37 GMT

Hello Sir,

I connect ADXL362 to Arduino & check waveform on oscilloscope.
I am reading register 0x01 (having device MEMS ID)
So data I am sending on MOSI is 0x0B (Read register command) , 0x01 (register address), 0x00 (dummy data).

Here is the waveform for CLK & MOSI

waveform for CLK & MISO

Arduino adds 1us delay in CLK after sending 1 byte. but nRF52840 sending continouns CLK signal (please refer previous post)

What is the reason to add delay on CLK?
If I want to add delay in CLK, Where do I add this in SPI driver?

Note: SDK version is 15.0.0_a53641a

Thanks in advance.

RE: Failed to read/write ADXL362 data using nRF52840

awneil — Tue, 20 Nov 2018 13:24:00 GMT

So use your scope to compare what happens in the Arduino (working) case against what happens with your nRF...

Note that the Keysight scopes can capture screen images to a USB stick - much better than taking photos!

RE: Failed to read/write ADXL362 data using nRF52840

beparas — Tue, 20 Nov 2018 13:09:04 GMT

Hello Sir,

As per your guide line I set SPI frequency to 1MHz and try to read register 0x01 (having device MEMS ID).
I am getting wrong value 0x18 instead on 0x1D (default value).

Here attached waveform of CLK & MOSI pin.

Waveform for CLK & MISO pin

Data we are sending on MOSI is 0x0B (Read register command) , 0x01 (register address), 0x00 (dummy data).

Next I connect my ADXL362 evaluation board to Arduino and follow this document.

Logs from Arduino device showing correct data for register value 0x00, 0x01, 0x02

So my ADXL362 evaluation board is not damages.

But when I connect ADXL362 to nRF52840 DK I am getting wrong data.

Thanks in advance

RE: Failed to read/write ADXL362 data using nRF52840

Håkon Alseth — Mon, 19 Nov 2018 14:17:08 GMT

Hi,

Have you scoped the signal to see how it looks? You might be running on too high SPI frequency.

You can change the frequency in spi_config.frequency = NRF_DRV_SPI_FREQ_1M.

Best regards,

Håkon

RE: Failed to read/write ADXL362 data using nRF52840

beparas — Mon, 19 Nov 2018 13:50:31 GMT

Hello Sir,

Thanks for you replay.

I recheck the pin connection (MOSI <--> MOSI, MISO <--> MISO), are ok.

I simplify above program & try to read Analog device ID (register address 0x00), MEMS ID (0x01), Part ID (0x02) but I am getting wrong values.

As per ADXL362 datasheet default values for register address 0x00 is 0xAD, for register 0x01 is 0x1D, for register 0x02 is 0xF2.

But I am getting wrong values for these register.

The values I am getting for register address 0x00 is 0x90, for register 0x01 is 0x18, for register 0x02 is 0xF0.

Could you please explain the reason for wrong values?

Thanks in advance.

void spi_event_handler(nrf_drv_spi_evt_t const * p_event,
                       void *                    p_context)
{
    spi_xfer_done = true;
    NRF_LOG_INFO("Transfer completed.");
    if (m_rx_buf[0] != 0)
    {
        NRF_LOG_INFO(" Received:");
        NRF_LOG_HEXDUMP_INFO(m_rx_buf, strlen((const char *)m_rx_buf));
    }
    else
      NRF_LOG_HEXDUMP_INFO(m_rx_buf, 3);
}
		
int main(void)
{
    bsp_board_init(BSP_INIT_LEDS);

    APP_ERROR_CHECK(NRF_LOG_INIT(NULL));
    NRF_LOG_DEFAULT_BACKENDS_INIT();

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

    NRF_LOG_INFO("#########################################################");
    NRF_LOG_INFO("SPI example started.");

    while (1)
    {
        // Reset rx buffer and transfer done flag
        memset(m_rx_buf, 0, m_length);
        spi_xfer_done = false;

        // </CS down> <command byte (0x0A or 0x0B)> <address byte> <data byte> <additional data bytes for multi-byte> … </CS up>
        m_tx_buf[0] = 0x0B;       // Read Register
        m_tx_buf[1] = 0x00;
        m_tx_buf[2] = 0x00;       // dummy data

        APP_ERROR_CHECK(nrf_drv_spi_transfer(&spi, m_tx_buf, 3, m_rx_buf, 3));

        while (!spi_xfer_done)
        {
            __WFE();
        }

        NRF_LOG_FLUSH();

        bsp_board_led_invert(BSP_BOARD_LED_0);
        nrf_delay_ms(200);
    }
}

RE: Failed to read/write ADXL362 data using nRF52840

Håkon Alseth — Fri, 16 Nov 2018 13:19:56 GMT

Hi,

[quote user=""]Code is also attached here, will you please guide me, If I did any wrong in SPI configuration?[/quote]

You configuration looks to be correct, MSB first, SPI Mode 0.

Are you sure that you have connected the pins correctly to the SPIS device? MISO to MISO, MOSI->MOSI, etc.?

Try reading a register, for instance PARTID (0x02), as this is non-zero in reset-value.

[quote user=""][/quote]

I am little bit confuse of "tx_buffer_length" & "rx_buffer_length" value passing in nrf_drv_spi_transfer() function.

Does the values of "tx_buffer_length" & "rx_buffer_length" are correct at line no 101 & 118 in main.c file?

This is related to the amount of bytes to send (tx) and the expected number of bytes you read (Rx) from the sensor. This may differ from register-to-register on your SPI slave device.

This function looks a bit suspicious:

uint8_t adxl362_read_one_reg(uint8_t reg_address)
{
uint8_t tx_buff[3] = {0};
uint8_t rx_buff[3] = {0};

// </CS down> <command byte (0x0A or 0x0B)> <address byte> <data byte> <additional data bytes for multi-byte> … </CS up>
tx_buff[0] = XL362_REG_READ;
tx_buff[1] = reg_address;
tx_buff[2] = 0x00; // dummy byte

APP_ERROR_CHECK(nrf_drv_spi_transfer(&spi, tx_buff, 3, rx_buff, 1));
NRF_LOG_INFO("** %d %d %d", rx[0], rx[1], rx[2]);
NRF_LOG_HEXDUMP_INFO(rx, 2);

return rx_buff[2];
}

Here you are sending 3 bytes, then receiving one (rx_buff[0]), then returning index 2 in rx_buff. Given your print log, it is still 0x0, but this function is likely not working as you intend it to.

Best regards,

Håkon