• State TI Thinks Resolved
  • Replies 7 replies
  • Answers 1 answer
  • Subscribers 31 subscribers
  • Views 135 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

TAS6424E-Q1: behavior of FAULT pin and overcurrent fault register

Jay Sun
Prodigy 60 points
Part Number: TAS6424E-Q1

Tool/software:

Hi TI team,

After performing stress test(monkey test) on audio system design with TAS6424E-Q1 audio amplifier, over current fault is detected on four channels: address = 0x10, value = 0xF0

In the mean time, the fault pin(no.26) is inactive. (Pull high)

It seems that the FAULT pin and OC fault register status is not matched logically.

It is supposed to be either: FAULT pin is active + fault register is detected or FAULT pin is inactive + fault register is not detected.

Besides, there is no reasonable condition to make OC happen based on the following setup:

a. The system is supplied from only an 2A power source. The input power is not enough to produce high current and trigger OC threshold. 

b. Channel output connect to Hi-z node for output monitoring.

There are two questions we need TI's input:

Q1: Is it possible that FAULT pin is inactive while fault register is detected?

Q2: Is there any possibility to cause OC fault in our condition?  

  • 0
    TI__Mastermind 35815 points

    Jay,

    I will need to understand your stress test before I can comment.  Also, is this on our EVM or on your board.

    Regards,
    Gregg Scott

  • 0 in reply to Gregg Scott
    Prodigy 60 points

    Hi Gregg,

    Thanks for the reply.

    In the stress test, we perform random input(for example, click the UI randomly) on our system. The test will last about 100hrs or higher and then we check the system's function status after test. That is when we found the overcurrent fault.

    From amplifier's perspective, it may receive random register change according to the system software behavior.

  • 0 in reply to Jay Sun
    TI__Mastermind 35815 points

    Hi Jay,

    With your random UI changes, are you sure you are making proper changes that would not cause an Over current event?  Do you have code that will allow the device to recover from an Over current event?   

    There are two ways for the OC to trigger.  A short to ground and short to power.  Your UI may have created an overcurrent event unknowingly. The software would need to clear this fault to allow the device to go to play mode.  I assume after the 100 hours the audio was not playing.

    Regards,
    Gregg Scott

  • 0 in reply to Gregg Scott
    Prodigy 60 points

    Hi Gregg,

    We do see some improvement on our experiment setup. It is possible that someone may cause OC unintentionally during manual operation, and we will prevent it to happen again.

    The main topic I would like to discuss with TI is about the FAULT behavior. After the test, we found that the FAULT pin is inactive while fault register is detected. However, it is aligned with the behavior mentioned in datasheet:

    Is it a possible behavior from your perspective?

    From our understanding, FAULT pin should present active since the fault register is not cleared yet, even the fault environment is rectfied.

  • 0 in reply to Jay Sun
    TI__Mastermind 35815 points

    Jay,

    If the OC is set in the registers, the Fault pin should be active low.  I do not see why you would have a fault in the fault register and the Fault pin is high.

    Regards,
    Gregg Scott

  • 0 in reply to Gregg Scott
    Prodigy 80 points

    Hello Gregg,

    Thanks to your reply. We are trying to reproduce the same fail mode on our side. In the meantime, could you please kindly help to share any guidance or suggestion of our debug steps below? We want to make sure we don't miss anything. Thank you.

    Below is our debug steps:

    1. Check Fault pin VDC level.

    2. Check Warn pin VDC level.

    3. Check STANDBY pin VDC level.

    4. Check MUTE pin VDC level.

    5. Check every channel's output VDC level.

    6. Apply all register dump. (check fault and channel state register's value.)

    Many thanks,

    Peter Wu

  • 0 in reply to JHENG-WEI WU
    TI__Mastermind 35815 points

    Peter,

    This is good practice.  I would keep a scope probe or voltmeter on the Fault pin to see when it triggers during your stress tests.  The way you can learn what in your code could be causing this unusual circumstance.

    Regards,
    Gregg Scott