• State
  • Locked Locked
  • Replies 3 replies
  • Subscribers 63 subscribers
  • Views 496 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.

TPS2121: Thermal event failure

Christopher Hess
Prodigy 95 points
Part Number: TPS2121

I have completed a design with the TPS2121 power mux and am working on prototyping it with a new board layout.  My design has an 18k current limit programming resistor to set the device current limit to its max of 5 A.  My bench top power supply current limit was set to 3A.  I verified the device working as intended with limited load, and continued to add more components to the board and again tried to power on.  The TPS2121 caught fire, there is a short 10 mil trace from pin 12 to a via to an internal ground plane, this trace also burnt up.  I believe there was an issue elsewhere on the board that could have caused a power to ground short, but I am unsure why this thermal event occurred.  It seems to me that bench top power supply should have limited current that could cause this damage, and that the TPS2121 should have shut down due to internal thermal monitoring.  What could have caused this thermal event?  Do I need to change my layout to increase the trace width to my ground plane?  I want to ensure this failure does not occur once the design is in production.  

  • 0
    TI__Genius 13760 points

    Hi Christopher,

    The TPS2121 is a very resilient device for it has several protective features that disable the device in the event that something could damage it. These protective features, such as over current and over temperature protection, disable the device for about 25ms. After the 25ms the device will restart and resume normal operation. One thing that could damage the device is being exposed to voltages that exceed the absolute maximum ratings through any pins of the device. The device will be damaged if ground was shorted and power was applied to the GND pin.

    I have a couple questions to find out if anything else could be the root cause of this event:

    1. What is your input voltage?

    2. Are OVx pins configured in your design?

    3. What is the output capacitance of the TPS2121?

    4. Was your load connected when the TPS2121 caught fire? If so, how much current does it pull?

    5. How long are the cables being used to input power from the bench top power supply?

    6. Do you have any capacitors on the input of the TPS2121?

    Best regards,

    Andy Robles

  • 0 in reply to Andy Robles
    Prodigy 95 points

    Both inputs can see between 12V and 16V.  The over voltage pins are configured to shutoff at 17.5V. 

    Output capacitance is 10,000uF.  A partial load was connected, it should not have pulled more than a few hundred milliamps -- but there may have been a short that caused more.  Again, the board was pulling the full 3A from the power supply.  

    The cables from the bench supply were about 3ft long.  

    I do not have any capacitors on the inputs.

  • 0 in reply to Christopher Hess
    TI__Genius 13760 points

    Hi Christopher,

    One concern I have here is with the 10,000uF capacitance at the output. The output capacitance of the output is a factor that affects the inrush current going through the device during start up or switchover that could potentially cause the device to restart [ I_Inrush = C_out * (dv/dt) ]. During start up the soft start feature will mitigate the amount of inrush current going trough the device but it could end up being a problem during switchover.

    Using the EVM with bench top power supplies could cause a ripple in the voltage applied caused by the inductance of the cables. When the device is turned on the channel will start conducting and current will start flowing through. In this transient the instantaneous change in current and the inductance of the cables will cause the voltage to ripple, potentially putting it over the over voltage threshold [ V = L * (di/dt) ].

    I don't believe either of these points would cause the device to catch fire, but minimizing that output capacitance could be a good start. The first thing that needs to get done is to find out if there is in fact a power to GND short in the board for this will directly affect the GND pin of the device.

    In the initial question you did mention that you started adding components to the board. Would you be able to share a partial schematic on the configuration of the board with the devices initially populated and with the devices you populated after the first test?

    Best regards,

    Andy Robles