• State
  • Locked Locked
  • Replies 5 replies
  • Subscribers 34 subscribers
  • Views 411 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.

Unexpect output state of PCA9557

Evan Wu
Prodigy 10 points
Other Parts Discussed in Thread: PCA9557

Hi all,

We encountered an unexpected behavior with PCA9557.

An I2C write command was issued by our controller followed by an I2C Write/Read command to do the validation.

Sometimes, we got a validation failure case (compare the value between write and write/read).

For example:

Step 1. Issue a write command to set the output register to 0xef

Step 2. Write/Read the output register to check the value. Sometimes we would not get 0xef but 0x01.

Does TI side have any errata on this I/O Expander?

  • 0
    Guru 317370 points

    The I²C protocol does not have any error checking.

    Please show an oscilloscope trace of the failed writes and read.

  • 0
    TI__Guru 73971 points

    Hi Evan,

    From a device perspective, our PCA devices do have an issue with bad power ups in the case where a bad ramp or power up from non GND occurs, the device can randomly assign values to all internal registers after power up. My guess is you had a bad power up and the polarity registers got flipped. When you see a case where you read 0x01h, can you also read the polarity registers and see what they are set to?

    -Bobby

  • 0 in reply to Clemens Ladisch
    Prodigy 10 points

    Reasonable. But, it's hard to do that in my case because the machine is on our customer's side.

  • 0 in reply to BOBBY
    Prodigy 10 points

    The failure case is in the runtime and the polarity registers will be initialized while controller FW boots up only.
    If I have a bad power-up situation in this case, I don't think the polarity registers will be set to the correct value (because I only do it when FW booted up).
    That means I should always read 0x01 at that time. And I believe the I/O Expander output state is changed to 0x01 indeed because the downstream devices are power cycled (I use PCA9557 as a power control pin, which is connected to a Gate of MOSFET). I will list this one in my checklist also.

  • 0 in reply to Evan Wu
    TI__Guru 73971 points

    If the FW is initializing all of the registers after power up then I suspect this isn't why you're seeing incorrect data being read/written. 

    We may want to start looking at scopeshots to see if the data is properly being sent and ACK'd when you send 0xef. 

    -Bobby