CC1352P: i2c slave mode not working

Ofir Cohen

Part Number: CC1352P
Other Parts Discussed in Thread: CC2650, SYSCONFIG

Hi,

I have a setup of 2 CC1352-P2 boards which are connected by i2c (a total of four wires: Pin 21 for i2c SCL, Pin 5 for i2c SDA, and two extra: GND and 3.3v).

Following this thread: "RTOS/CC1352P: i2c slave mode" (and the code snippet in its related threads) I tried to create 2 CCS projects:

First board acts as the master
Second board acts as the slave

The master tries to send the slave (address 0x74) two bytes ['h', 'e'] and the slave upon receiving them sets its two LEDs (green and red).

Attached are the 2 CCS projects with the code I used.

Unfortunately it doesn't work (i.e. the slave never receives the 2 bytes message) and I'm not sure why, please advise.

Thanks,

Ofir

i2c_master.zip i2c_slave.zip

over 5 years ago

0 Severin Suveren over 5 years ago

TI__Mastermind 23730 points

Hi Ofir,

Just so I understand you correctly, have you connected 3v3 and GND between the two LaunchPads directly with wires?

0 Ofir Cohen over 5 years ago in reply to Severin Suveren

Prodigy 100 points

Severin Suveren said:
Hi Ofir,
Just so I understand you correctly, have you connected 3v3 and GND between the two LaunchPads directly with wires?

Correct.

Any idea what is wrong with the code / setup?

0 Severin Suveren over 5 years ago in reply to Ofir Cohen

TI__Mastermind 23730 points

Could you take a photo of the setup?

0 Ofir Cohen over 5 years ago in reply to Severin Suveren

Prodigy 100 points

Below is a photo of the setup (note that in my *actual* setup I disconnected one of the USB cables below so the power of one of the board actually comes from the other board via the 3v3 pin).

Please let me know if you need more info.

Thanks,

Ofir

0 TER over 5 years ago in reply to Ofir Cohen

TI__Guru**** 317180 points

From a hardware perspective it makes sense to use common ground on the two boards but disconnect the 3V3 connection since you are basically shorting the output of two voltage regulators together.

0 Ofir Cohen over 5 years ago in reply to TER

Prodigy 100 points

Hi TER,

Thanks for the reply, good to know.

As I said above though: "note that in my *actual* setup I disconnected one of the USB cables below so the power of one of the board actually comes from the other board via the 3v3 pin"

Has someone looked at the attached CCS projects (master + slave) and the code inside them?

Thanks,

Ofir

0 TER over 5 years ago in reply to Ofir Cohen

TI__Guru**** 317180 points

Sorry, I was only looking at the picture. When we ask about a picture, my assumption is that the picture shows the actual test setup without any exceptions.

0 Ofir Cohen over 5 years ago in reply to TER

Prodigy 100 points

My bad =D.

So what would be the best way to tackle this issue?

Is there any official / (TI) internal i2c slave sample for CC1352P that we can refer to?

I'm guessing the problem is not in the i2c master sample because this code is taken from the CCS samples.

Thanks,

Ofir

0 Severin Suveren over 5 years ago in reply to Ofir Cohen

TI__Mastermind 23730 points

Hi Ofir,

No sadly not, we don't have any I2C slave example. However, it should be fairly trivial to convert the normal I2C sample application into a I2C slave example. Just use the I2C driver documentation. Also, please use a debugger to check where in the code it hangs/crashes.

0 Ofir Cohen over 5 years ago in reply to Severin Suveren

Prodigy 100 points

HI Severin,

Thanks for your reply.

I've posted it here because it is not fairly trivial at least to me and most likely to many other people due to this issue being asked in many other TI threads:

https://e2e.ti.com/support/wireless-connectivity/sub-1-ghz/f/156/t/816289

https://e2e.ti.com/support/legacy_forums/embedded/tirtos/f/355/t/482090

So someone was able to test it on CC2650 but no one has ever confirmed that it works on CC1352P.

I am stuck at the moment and I would highly appreciate more help.

IMHO, It's not about building / merely attaching a debugger to the code, it's about debugging the protocol and/or TI-RTOS drivers/APIs that don't behave as expected.

Note that the attached code is legit and based on samples from an official i2c TI sample for the master, and a non-official TI sample (from the forum posts) for i2c slave,

and not something I simply made up.

I would highly appreciate follow-up and an advice from a TI expert to look into the attached code and samples, and at least confirm whether or not the code works on CC1352P.

Thanks,

Ofir

0 Severin Suveren over 5 years ago in reply to Ofir Cohen

TI__Mastermind 23730 points

I was able to quickly throw this together as a proof of concept for I2C slave:

#include <stdint.h>
#include <stddef.h>
#include <string.h>
#include <unistd.h>

#include <ti/drivers/GPIO.h>
#include <ti/drivers/Power.h>
#include <ti/drivers/power/PowerCC26XX.h>

#include <ti/devices/DeviceFamily.h>
#include DeviceFamily_constructPath(driverlib/i2c.h)
#include DeviceFamily_constructPath(driverlib/ioc.h)
#include DeviceFamily_constructPath(inc/hw_memmap.h)

#include "ti_drivers_config.h"

#define Board_I2C_SCL0   IOID_4
#define Board_I2C_SDA0   IOID_5

#define I2C_SLAVE_ADDR    0x74

void *mainThread(void *arg0)
{
    Power_setDependency(PowerCC26XX_PERIPH_I2C0);

    IOCPinTypeI2c(I2C0_BASE, Board_I2C_SDA0, Board_I2C_SCL0);

    I2CSlaveInit(I2C0_BASE, I2C_SLAVE_ADDR);

    char i2cData[10] = { 0 };

    while (1)
    {
        size_t index = 0;

        while (I2CSlaveStatus(I2C0_BASE) != I2C_SLAVE_ACT_RREQ_FBR);
        i2cData[index] = I2CSlaveDataGet(I2C0_BASE);

        for (index = 1; index < 2; index++)
        {
            while (I2CSlaveStatus(I2C0_BASE) != I2C_SLAVE_ACT_RREQ);
            i2cData[index] = I2CSlaveDataGet(I2C0_BASE);
        }

        i2cData[index] = '\0';

        if (strcmp(i2cData, "he") == 0)
        {
            GPIO_toggle(CONFIG_GPIO_GLED);
        }
        else
        {
            GPIO_toggle(CONFIG_GPIO_RLED);
        }
    }

    return NULL;
}

I tested this with a normal I2C Master example, and both I2C master and slave is able to correctly complete the I2C transfer.

0 Ofir Cohen over 5 years ago in reply to Severin Suveren

Prodigy 100 points

Hi Severin,

Much appreciated the sample and the time you took to build and test it!

I was able to build it locally, I'll test it soon and update.

Thanks,

Ofir

0 Severin Suveren over 5 years ago in reply to Ofir Cohen

TI__Mastermind 23730 points

I configured CONFIG_GPIO_GLED and CONFIG_GPIO_RLED with SysConfig. You can replace these with the equivalent Board_GPIO_* define in your Board.h.

0 Ofir Cohen over 5 years ago in reply to Severin Suveren

Prodigy 100 points

Hi Severin,

Thank you so much, I'm able to confirm that it is working and the issue is resolved, much appreciated for the time and working sample code!

I will mark this issue as resolved.

Small question, the toggling of the LEDs on the board happen very fast, so I have set a breakpoint at the strcmp() of 'he' and in the slave code

and let it resume.

It stops at the first time but doesn't stop again (although the master.infinitely tries to send 'he' to the slave).

Any idea?

Thanks,

Ofir

Sub-1 GHz

Sub-1 GHz forum

CC1352P: i2c slave mode not working