• State Resolved
  • Locked Locked
  • Replies 12 replies
  • Answers 5 answers
  • Subscribers 64 subscribers
  • Views 635 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

TPS55165Q1-EVM: Suspected TPS55165-Q1 Converter Failure - No Output Voltage

Matthew Jones4
Prodigy 70 points
Part Number: TPS55165Q1-EVM
Other Parts Discussed in Thread: TPS55165-Q1, TMS570LS0432

I have a TPS55165-Q1 running on an TPS55165Q1-EVM Evaluation Module. After I set it up on the bench the converter was working correctly and it was powering the load correctly. I went through a few power cycles without issue. I then powered the device up and there was a 'pop' noise. Now there is no output voltage. I suspect the converter has failed somehow.

My setup is:

  • The input voltage (VBAT on development board) was running off a lab PSU set to approx. 13.5 V.
  • The output voltage selection jumper (J6) was set to 5V.
  • The output voltage (VOUT) was being used to power a micro-controller development board (TI Hercules TMS570LS0432 LaunchPad). 
  • The PS jumper (J5) was in the PWM position.
  • The PG_DLY (J4) jumper was not fitted.
  • The SS_EN jumper (J3) was fitted.
  • The IGN_PWR jumper (J2) was not fitted. Instead the IGN_PWRL was being controlled via the Hercules micro-controller (using a NPN transistor to drive the pin low to unlatch the supply).
  • The IGN jumper (J1) was not fitted. Instead the IGN pin was wired to VBAT (13.5 V PSU) via a switch.

During the process of testing I was turning the lab PSU on and off several times. Usually I would start up VBAT with the IGN switch open, then after a moment closing the IGN switch to start up the converter. On the last time I powered up the device (when it failed) the IGN switch was already closed when I turned on the PSU. This might be coincidence or it may be related to the failure.

I see from the TPS55165Q1-EVM schematic that VBAT is filtered (VBAT_F) before it is is connected to VINP and VINL. With my bench setup, IGN is connected directly to the PSU (i.e. VBAT not VBAT_F). As VINP and VINL are filtered and IGN is not, then IGN could rise faster than VINP and VINL. Could this cause damage to the device? Is this the likely case of the failure? If IGN comes from a remote source do you recommend filtering or additional protection on the IGN input?

Thank you in advance of any help. Please let me know if you need more information to help diagnose the issue.

  • 0
    TI__Mastermind 40080 points

    Dear Sir,

    Did you do the test based on the TPS55165EVM board? What is the load current during the test?

  • 0 in reply to Helen chen
    Prodigy 70 points

    Dear Helen,

    Yes, the issue occurred running on the TPS55165Q1-EVM board. I did not get chance to measure the 5V load current before the failure. But the 5V load (TI Hercules TMS570LS0432 LaunchPad) is designed to be powered over USB at < 500 mA, so I understand that it is within the rating of the TPS55165-Q1. Shortly before the converter failed, I measured the input current to the converter (VBAT) to be approx. 80 mA @ 13.5 V.

    If it is critical that you know the exact load current, please let me know and I'll measure it.

    Best regards,

    Matthew

  • 0 in reply to Matthew Jones4
    Prodigy 70 points

    Dear Helen,

    Just to add to my post above, I've measured the 5V load at 150 mA.

    Best regards,

    Matthew

  • 0 in reply to Matthew Jones4
    TI__Mastermind 40080 points

    This is a small load.

    Could you check which pin is damaged?  WHat is the impedance from SW1 to GND, SW2 to GND.

  • 0 in reply to Helen chen
    Prodigy 70 points

    Dear Helen,

    I was not aware of any minimum load requirement for the TPS55165-Q1. Figure 3 in the datasheet (5-V Output Regulation vs Load Current) shows output regulation all the way down to 0A output current. 

    I've checked the datasheet and I cannot see the SW1 and SW2 pins you are referring to. Please see extract from datasheet below:

    I thought you may have been referring to the L1 and L2 pins (for the inductor). I have measured these:

    •  L1 to GND = 3 MOhm
    •  L2 to GND = 3 MOhm

    Please let me know if you are any other pins you would like me to check to help you diagnose.

    Best regards,

    Matthew

  • 0 in reply to Matthew Jones4
    TI__Mastermind 40080 points

    Thank you for the update Matthew.

    Yes, L1 means SW1.

    There's another way to check the IC. You can check the ESD diode from the GND net to each pin.  If the IC damaged, generally,  the related ESD diode can't be detected out.  Put the multimeter's + to GND, put the multimeter's -  to each pin.

  • 0 in reply to Helen chen
    Prodigy 70 points

    Dear Helen,

    I've checked the ESD diodes on all pins on the IC as requested. No jumpers fitted to development board when taking measurements.

    • Pin 1 (PGND) = 0 V
    • Pin 2 (L1) = ~0.5 V
    • Pin 3 (BST1) = ~0 V
    • Pin 4 (VINP) = ~0.6 V
    • Pin 5 (VINL) = ~0.6 V
    • Pin 6 (IGN) = Open Circuit
    • Pin 7 (PS) = ~0 V
    • Pin 8 (IGN_PWRL) = ~0.6 V
    • Pin 9 (SS_EN) = ~0.6 V
    • Pin 10 (PG_DLY) = ~0.6 V
    • Pin 11 (VREG_Q) = ~0 V
    • Pin 12 (VREG) = ~0 V
    • Pin 13 (GND) = 0 V
    • Pin 14 (VOS_FB) =~0.6 V
    • Pin 15 (PG) = ~0.6 V
    • Pin 16 (VOUT_SENSE) = ~0 V
    • Pin 17 (VOUT) = ~0 V
    • Pin 18 (GND) = 0 V
    • Pin 19 (BST2) = ~0.6 V
    • Pin 20 (L2) = ~0.5 V

    I suspect means the following pins are likely damaged: BST1, PS, VOUT, VOUT_SENSE, VREG and VREG_Q. whilst doing the test I have noticed a small hole in the IC package, on the top between the PS and IGN pins. I've attached a picture below although I apologise for the poor quality. In the picture it looks like a black mark, but it is definitely a small crack/pinhole in the package.

    Do you have any idea what may have caused the device to fail in this way? Is there any issue with the IGN being switched remotely? For example could damage occur if IGN > VINP/VINL?

    Thank you for your assistance.

    Best regards,

    Matthew

  • 0 in reply to Matthew Jones4
    TI__Mastermind 40080 points

    Hi Matthew,

    Thanks for the detailed information. it looks the IGN pin also damaged.

    I think there should be a long wire between the VBAT and the IGN. I noticed there's no extra filter capacitor for the IGN pin on the EVM board. So at the time the switch closed, the voltage ring at the IGN pin will be very high, and the dv/dt will also very high, which might cause the IGN failure and also the wrong startup sequence, hence cause the other pin failed.

    On the original pcb board, I checked the pcb layout, the whole trace length from the VBAT to IGN is about 6.5cm.

    Could you please take a whole picture of how the switch connected? what about the wire length?

  • 0 in reply to Helen chen
    Prodigy 70 points

    Hi Helen,

    Thank you for the help. 

    I can't take a picture at the moment, but I have drawn a sketch showing the setup including lengths. The wire lengths are very long (up to 5 meters) and the wiring is is 0.5 mm2. The IGN switch is located at the PSU,  so is approximately 4.9 meters from the TPS55165Q1-EVM. As you say, this will  not be helping the issue with ringing and high dV/dt on the IGN pin. 

    I feel we are close to having this issue resolved. I have another TPS55165Q1-EVM board ready to try, but I don't want it to fail in the same way. I clearly need to make some changes to the setup before I try again. I just have a couple of questions you can hopefully help with:

    - I will have long cable lengths (up to 5m) with a remotely switched ignition in the real application. I can't change this, so I need to protect the IGN pin. To try to protect the IGN pin I am going to use a low pass (RC) filter and also a TVS diode to clamp to < 36V. What would you recommend?

    - Can damage occur if the IGN voltage is applied before VBAT? I can't guarantee that the IGN SW won't be left closed when the PSU is first turned on. There is then a 'race condition' between the rise of IGN or VINP/VINL. A low pass filter on the IGN pin should help as it will slow the rise of the IGN signal.

    Best regards,

    Matthew

  • 0 in reply to Matthew Jones4
    TI__Mastermind 40080 points

    Hi Matthew,

    Thanks for your question. Please check my below reply:

    - I will have long cable lengths (up to 5m) with a remotely switched ignition in the real application. I can't change this, so I need to protect the IGN pin. To try to protect the IGN pin I am going to use a low pass (RC) filter and also a TVS diode to clamp to < 36V. What would you recommend?

    --> Helen: Yes, RC filter can be helpful for the issue. Suggested RC filter is 10k/(100nF-470nF), please put this filter as close to the IC as possible. Make sure the high frequency noise can be filtered before goes into the IGN pin. you can leave a TVS on the pcb board. But I think generally a RC filter is OK.

    - Can damage occur if the IGN voltage is applied before VBAT? I can't guarantee that the IGN SW won't be left closed when the PSU is first turned on. There is then a 'race condition' between the rise of IGN or VINP/VINL. A low pass filter on the IGN pin should help as it will slow the rise of the IGN signal.

    --> Helen: Yes, damage can occur if the IGN voltage is applied before VBAT. This must be avoid. So the connect of the VBAT should be after the input cap or at the L1 ,  RC filter can also help the time delay.

  • 0 in reply to Helen chen
    Prodigy 70 points

    Hi Helen,

    Thank you for your suggestions and support with this issue. I will implement these changes to protect the IGN pin.

    Many thanks.

    Best regards,

    Matthew

  • 0 in reply to Matthew Jones4
    TI__Mastermind 40080 points

    You're welcome, Matthew.

    Good luck!