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.

TS3A24159: I2C error when operating the switch.

Part Number: TS3A24159

Hello! I'm a camera engineer for automobiles.

Currently, we would like to control I2C using TS3A24159.
However, when the serializer's GPIO7,8 connected to the IN pin of TS3A24159 is operated, the I2C communication error and serializer are turned off.
I share the conditions I tested. Please check what additional debugging points are there.
Please refer to the attached circuit.

1. When U5 and U4 are separated (R7 and R9 removed), I2C is detected in D1 and D3 pins of U5, and serializer operates normally.
2. Even if the pull-up resistors R38, R39, R37, and R40 of the U4 are removed, I2C is not operated and the serializer is turned off.
3. Even if the pull-up resistors R38, R39, R37, R40, R19, R20, R25, and R35 of U3 and U4 are removed, I2C is not operated and the serializer is turned off.
4. When TS3A24159 is completely removed, it is confirmed that GPIO 7 and 8 of the serializer operate normally.

 

 Circuit.pdf

Thank you!

  • Please note that the pull-up resistors of all enabled bus segments are in parallel. The resulting load of 1.5 kΩ might be too large for at least one of the devices.

    Anyway, please show an oscilloscope trace of the I²C signal.

  • Hi Choi,

    Thanks for providing this detailed information to help with the debug process. A few things I would suggest:

    1) It is advised to have pullup resistors on both sides of the switch to ensure the I2C lines are never in a floating state (I do not see pull ups on the SDA/SCL lines). Even though the switch is normally connected to one of the lines that do have pullups, there is a feature in many of our switches (including this one) that implements a break-before-make feature which disconnects the switch inputs and outputs entirely before switching to the desired I/O.

    2) Bullet point #1 is showing that I2C signal does work properly when the second multiplexer stage is disconnected indicating the loading could be problematic when putting these multiplexers in series. Note that since the RON for these devices are very low, it means they have pretty high capacitance as a consequence. Now, I2C has a max operating frequency at only 400kHz, so these should be able to handle being cascaded as these muxes have 23MHz of bandwidth. However, do you know how much capacitance/loading is on the entire SCL/SDA lines?

    3) When you do have multiple pullup resistors on the SCL/SDA lines, they will act essentially in parallel, reducing the overall pullup resistance for the I2C lines. Hence, they may be too strong and not allow the line to be pulled low at all. You may need to increase the resistance of the pullups to mitigate this and get the desired pullup resistance to properly function.

    4) Lastly, any scope shots on the SCL/SDA lines on the input to each multiplexer and on the output of each multiplexer would help greatly in helping resolve this.

    Thanks!

    Bryan

  • Thank you for reviewing it.
    Serializer's I2C Input capacitance is about 3pF. I don't know the input capacitance of the sensor and LED driver. South Korea currently has a long holiday, so I'll revise the I2C Pull Up resistance value next Thursday. If it doesn't work, I'll measure scope image.

  • Thank you for reviewing it.
    South Korea currently has a long holiday, so I'll revise the I2C Pull Up resistance value next Thursday. If it doesn't work, I'll measure scope image.

  • Sounds good Choi. Please update at your earliest convenience.

    Thanks!
    Bryan

  • Hi Clemens,
    As you mentioned, I lowered the parallel pull-up resistance value and it worked correctly.

    Thanks

    Choi

  • Hi Bryan,
    As you mentioned, I tried to increase the pull-up resistance value, but it still didn't work. Instead, it worked when R30 and R31 were lowered to 1.5K or less, that is, 0.9K was calculated for parallel resistance calculation.
    Anyway, the communication problem was solved well, and thank you for your help.

    Thanks

    Choi