• State TI Thinks Resolved
  • Locked Locked
  • Replies 3 replies
  • Answers 1 answer
  • Subscribers 34 subscribers
  • Views 302 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

TCA6408A: Question about RESET input

Daniel Benkman
Intellectual 525 points
Part Number: TCA6408A


We have a design that uses the TCA6408A expander chip.  The current design has its VCCI power input connected to a 1.8v supply voltage, and the VCCP power input connected to a 3.3v supply voltage.

We are having trouble figuring out why the chip wont respond to its I2C bus, while another chip (an accelerometer chip) responds perfectly well to the very same I2C bus.  AThis happens on every board we test.

We noticed that the TCA6408A expander chip's RESET input is connected to the 3.3v supply through a 100k resistor in our design.  We also noticed that the spec says that there are clamps inside the chip that implies that diodes are connected from the input pins to both the supply voltage and to ground.

One question is if the protection diode going to the supply voltage on the RESET line goes to VCCI or VCCP?  If it goes to VCCI, could that create a problem, and even damage the chip, because the RESET input's pullup to 3.3 v would be turning on the RESET input's protection diode to VCCI?

Could this be causing the TCA6408A expander chip not to function correctly, or even damage the chip?

All the other inputs to the TCA6408A expander chip seem to be OK, even though we can't talk to it over I2C.  Can you suggest any other reasons why the chip may not communicate over I2C on our boards?

Should we move the RESET input's pullup from VCCP to VCCI?  Should we also replace the TCA6408A expander chip?


Thank you for any help you can give.

Dan Benkman
Innowave
db4772@gmail.com

  • 0
    TI__Mastermind 34590 points

    Hello Dan,

    I don't believe you will damage the device by having the reset pin connected to VCCP through a 100k resistor as this is withing the voltage tolerance rating of the device.  However, the Reset pin should be connected to the VCCI rail instead of the VCCP.  Have you tried moving it to the VCCI rail to determine if the communication is successful? 

    There is another thread that discusses the Power up sequence of VCCI and VCCP on the 16pin version of this device (TCA6414A) and is also relevant to the TCA6408A.  The VCCP rail controls the reset/initialization of the device and shoudl be powered on after the VCCI rail.  Is this the case for your application?  If not, then the reset pin would most likely need to be toggled to initialize the device after power up.  However, if the pullup resistor on the Reset pin is connected to the VCCI rail, then there is no power up sequence dependency either rail can be powered up first, and the device should be initialized properly to allow for communication.

    Since you have the Reset pin connected to VCCP instead of VCCI, I am thining you might bet getting stuck in a POR/Initialization situation.  Can you try moving the pullup resistor to VCCI and check communication again?

    Regards,

    Jonathan

  • 0 in reply to Jonathan Nerger
    Intellectual 525 points

    Jonathan,

    Thank you for your help!  The power on sequencing issue when the Reset pin is connected to VCCP was completely unknown to us.  We will definitely try this change.  Since our board is small and has surface mount components, we wanted to ask before getting a rework expert to make the change.

    I'll not mark this as resolved until we try it, so that you and readers looking for an answer to similar problems can see if this solution works.


    Dan Benkman
    Innowave

  • 0 in reply to Daniel Benkman
    TI__Mastermind 34590 points

    Hello Dan,

    Yes, I don't see any cause for damage to the device, and while I'm speculating a little that there is a stuck reset issue, I believe it to be a possibility since we know the pullup should be connected to VCCI by design and a dependency could be established.  I hope simply moving the resistor resolves the issue, otherwise I can dive deeper into the application and look for other possible causes.

    Thanks and Regards,

    Jonathan