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.

TPS63020: L2 short to GND

Part Number: TPS63020
Other Parts Discussed in Thread: TPS63070

I've had 3 TPS63020 parts running for several weeks in my prototypes without any problems.

First 10 boards of new production some are failing.  Some run for hours other fail immediately.

Defective device L2 pin can be measured as almost 0 Ohm to GND.

Input Voltage is 3.3V to 5.5Volts

Output 5.4 Volts Load 1.5 Amps.

PCB layout is almost identical to EVM pcb. 

Production code on original chips are different to the 10 with failures they all have the same production code.

Is there a known issue with TPS63020 ?

  • Hello H,

    Could you please tell me if there are any other differences between your design(s) and the EVM? What kind of inductor is used? Do you have load transients?

    In the meantime I can try running the EVM under the same conditions.

    Best regards,
    Milos
  •  L1 = Coil Craft  XFL4020-102MEC
    Output Caps: C6,7,25,31 are 22uF 10V X7R C9=1uF 25V X7R, C10=10nF 50V NOP
    Input Caps: C8 C16 are 22uF 10V X7R
    R1=1M
    R2=100K  for Vout for 5.4V
    Pin PS/SYNC is tied to VINA which is decoupled with 100nF X7R 25V
    Enable Pin is pulled down to GND by 100K and enabled by Microcontroller with 3.3Volts.
    Your comments are welcome.

  • Hello H,

    Thank you for the layout. The connections look fine, however I have a few remarks:
     - The output capacitors should be placed closer to the device, ground ends especially. If that is not possible, at least ground vias should be used to improve the connection between GND ends of the I/O capacitors and the device GND.  
    - Thermal vias should be added under the exposed pad to prevent from thermal protection limiting the output current. 
    - It is not a good idea to share a ground path between the power part and the control (feedback) part. You can brake this by cutting the GND layer here:




    I ran the EVM under the same conditions and did not experience any problems. For the  start let's try a few things:
     - Do you have load transients, and if so, could you show scope measurements of the I/O voltage when the output current change, and during startup?
     - Can you try running the circuit at lower output currents? 
     - Do you still have some devices from the previous batch that you could try?

    Best regards,
    Milos

  • Thanks for your reply. No I've used all the chips from that bad batch. I ordered new and they have a different batch number. Read on … I doubt I will even use the new chips..

    Frankly I've lost confidence in this chip and have decided to move the DC-DC section to the TPS63070 . It's not a efficient as the first choice, but time is now running out.

    I ordered the TPS63070 EVM wired it for my needs and ran it all night.

    My test switches on a 1.5Amp load for 250 mS and then OFF for 5 seconds the TPSEVM is still running after 11 hours...

    I have to make a new PCB and doing that to the TPS63020 and possibly still having a failure is daunting...


    Point taken about the FB GND... Thanks
  • Hi H,

    With its extended voltage range, the TPS63070 is a good choice. In any case I will run the load transient tests to see if I can reproduce the same behavior. I suspect that the load transients could cause excessive oscillations on output voltage which might trip the overvoltage protection, since the device is operating close to the maximum recommended value. But still, that should not damage the device. If you still try the TPS63020 on your current boards, it would be nice if you could post scope images of output voltage during load transients.

    Best regards,
    Milos
  • I agree and when I first made the choice I figured I was on the edge of thin ice with the input Voltage.

    If I ever spend time to put the different batch chip on the boards I will post the scope image as you suggested.

    Thanks again for you help