This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

SN65DSI83: SN65DSI83 I2C communication not working

Part Number: SN65DSI83


Hi,

I am having some trouble communicating with my SN65DSI83. I am using i2ctool (i2cdetect and i2cdump), but I never receive an ACK. My I²C bus has a 2.2k pull-up, and I can communicate with all other devices on the bus.

 

root@evk-kit-ai:~# i2cdetect -y -r 2
     0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f
00:                     -- -- -- -- -- -- -- -- --
10: -- -- UU -- -- -- -- -- -- -- -- -- -- -- -- --
20: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
30: 30 -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
40: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
50: UU -- -- -- -- -- -- -- UU -- -- -- -- -- -- --
60: -- -- -- -- -- -- -- -- UU 69 -- -- -- -- -- UU
70: -- -- -- -- -- -- -- --
root@evk-kit-ai:~# i2cdump -y -f 2 0x2d
No size specified (using byte-data access)
     0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f    0123456789abcdef
00: XX XX XX XX XX XX XX XX XX XX XX XX XX XX XX XX    XXXXXXXXXXXXXXXX
10: XX XX XX XX XX XX XX XX XX XX XX XX XX XX XX XX    XXXXXXXXXXXXXXXX
...
f0: XX XX XX XX XX XX XX XX XX XX XX XX XX XX XX XX    XXXXXXXXXXXXXXXX

Screenshot 2025-11-21 125243.png

Screenshot 2025-11-21 125119.png

To test further, I tried connecting the EN pin to ground through a 220nF capacitor, but it didn’t help. I also added a 25MHz clock on the REFCLK pin, but that didn’t help either. I tested the voltage on both the ADDR pin and EN pin, and both are HIGH. To be sure, I tested the Vcore (1.135V) and Vcc (1.8V).

Could you please help me troubleshoot this problem?

Thank you and best regards,

Gabriele

 
  • Hi Gabriele,

    What address I2C are you reading from? Since the ADDR pin is pulled up here, the read address should 0x58 (8-bit). Please try reading from this address.

    Best regards,
    Ikram

  • Hi Ikram,

    Based on the pull-up on the ADDR pin, I believe the 7-bit address should be 0x2D, giving a write address of 0x5A and a read address of 0x5B. Could you confirm if that looks correct?

    To be sure, I tried both 7-bit addresses 0x2c and 0x2d.

    Here is the result of my i2cdetect:

    root@evk-kit-ai:~# i2cdetect -y -r 2
    0 1 2 3 4 5 6 7 8 9 a b c d e f
    00: -- -- -- -- -- -- -- --
    10: -- -- UU -- -- -- -- -- -- -- -- -- -- -- -- --
    20: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
    30: 30 -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
    40: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
    50: UU -- -- -- -- -- -- -- UU -- -- -- -- -- -- --
    60: -- -- -- -- -- -- -- -- UU 69 -- -- -- -- -- UU
    70: -- -- -- -- -- -- -- --


    I also tested with i2ctransfer:

    root@evk-kit-ai:~# i2ctransfer -y -f 2 r1@0x2d
    Error: Sending messages failed: No such device or address


    root@evk-kit-ai:~# i2ctransfer -y -f 2 r1@0x2c
    Error: Sending messages failed: No such device or address

    As you can see, using both 8-bit read addresses (0x5B and 0x59) does not return an ACK on the 9th bit.

    To confirm further, I tried write transactions on both addresses, and here are the results:

    root@evk-kit-ai:~# i2ctransfer -y 2 w1@0x2d 0x0d r1@0x2d
    Error: Sending messages failed: No such device or address


    root@evk-kit-ai:~# i2ctransfer -y 2 w1@0x2c 0x0d r1@0x2c
    Error: Sending messages failed: No such device or address

    Again, using the 8-bit addresses (0x5A and 0x58) I don’t get an ACK.

    Despite testing both the 7-bit and 8-bit addresses with i2cdetect and i2ctransfer, I consistently receive 'No such device or address' errors and no ACK. Could you advise on what I might be missing or suggest the next steps to debug this?

    Thank you for the support,

    Gabriele

  • Hi Gabriele,

    Have you checked this on multiple boards and do they all show the same behavior?

    Could you also please control the EN pin and make sure it is enabled after the power rails are stable.

    If possible, if you have another I2C-capable device available, could you wire and/or solder on the I2C lines and check if it can detect the device?

    Best regards,
    Ikram

  • Hi Ikram,

    I tested all 5 of my prototypes and observed the same behavior on each of them. Regarding the EN pin, I verified it by removing the pull-up resistor, shorting it to GND, and then releasing it while the board was already powered on. The behavior remained unchanged.

    Unfortunately, I don’t have another I2C-capable device available to connect to the bus. What I find puzzling is that I can successfully communicate with all the other devices on the I2C lines, but not with this one.

    Thank you very much for your support.

    Best regards,

    Gabriele Vezzali

  • Hi Gabriele,

    Could you probe the I2C pins close to the device and check the signals. Please also check the layout for this, the I2C traces and connections, device footprint, and power rails.

    The I2C speed here is 100 kHz right? This is within the 400 kHz limit as well. Since this is seen on all the prototypes, and there is no known issue with I2C like this, there could be a system issue, with the layout and connections. 

    Best regards,
    Ikram

  • Hi Ikram,

    The i2c speed is indeed 100 kHz, and the layout of the board seems good to me.

    This is a probe done as close as i can on the pin of the i2c (yellow=sda pink=scl).

    regards,

    gabriele vezzali

  • Hello,

    The team is out due to a public Holiday in the US. Responses will be delayed until Monday.

    Thank you for your patience

  • Hi Gabriele,

    Thank you for sharing these. I could not yet identify the cause for this since the pinout compared to the layout looks correct. Is the part soldered on correctly? Is it possible that it was rotated somehow?

  • Hi Ikram,

    We identified a hardware error on our side. To continue testing, we have purchased an EVM board for now.

    Thank you for your support.

    Best regards,

    Gabriele

  • Thank you Gabriele. Please let us know there are any further questions.

    Best regards,
    Ikram