• State
  • Locked Locked
  • Replies 8 replies
  • Subscribers 63 subscribers
  • Views 635 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.

TPS564247: Design Review, Power Failures

Peter
Prodigy 10 points
Part Number: TPS564247
Other Parts Discussed in Thread: TPS25947, TPS566242,

Hi all,

I've been working to make a power management board (PMB) that uses the TPS564247 to provide 3.3V, 5V, and 6V. However, during use the TPS564247 regulators manage to keep breaking due to unknown causes and we wanted to get a sanity check on our design to see if there are any errors. For our next iteration we'll be moving to the TPS566242 regulators with TPS25947 e-fuses on each power rail. Here's some more information about out use cases and circuit loads.

  • Batteries (6-8.4V, 7.4V Nominal) During usage they never went below 7.5V
  • PSU and Battery inputs are never used at the same time
  • 3.3V, 5V, 6A loads should never exceed 2.5A max individually, typical is 0.5~1A per regulator
  • 3.3V: Microcontroller + low power RF transceivers
  • 5V: Microcontroller + sensors
  • 6V: Servo with 2.5A max stall current
  • We conformal coat all exposed surfaces
  • The enable pin for each regulator is configured to the auto-enable side of the jumper
  • Batteries have polarity keyed connectors that prevent reverse polarity situations

As it stands we've worked our way through about 7 regulators due to unknown failures on each voltage rail. The regulators do not release any magic smoke. They simply stop working, except for one specific instance. For example, our system was working nominally we powered it off for a day and then after the next power up the 5V regulator was non-functional. Here are some of our working theories.

  • The 6V regulator had too much current draw with our stalling servo or overheated due to extended usage. (We're fairly certain on this one since it emitted smoke once)
  • The 5V, 3V3 regulators may have experienced some sort of ESD
  • Ripple voltage or current may damage them
  • Start up has a too sharp slew rate

Any and all feedback is appreciated. Thank you!

PMB_TI_Review.pdf

  • 0
    TI__Genius 13451 points

    Hi Peter!

    1. Reviewed your SCH, it suggests to add 30pf feed forward cap, which is parallel with top FB resistor. It can increase the phase margin of control loop stability. But I don't this is the root cause of your issue.

    2. It suggests check the voltage waveform of VIN, EN,SW, when the regulator stop working.

    Shuai

  • 0 in reply to Shuai Fan
    TI__Genius 13451 points

    3. During the manufacture process of PCB, it is any configuration about the programming burning? We saw there is a programming Jumper in SCH. I mean, during programming burning, if input is floating(battery is disconnected.) but force a 5V voltage to the output side of TPS564247, this may cause the damage risk.

    https://www.ti.com/lit/an/slyt689/slyt689.pdf Please refer this application note.

    4. For failure power rail, could you remove the device from PCB and check the whether the internal MOSFET is damaged by measure the impedance between VIN and SW, SW and GND?

    5. Pls provide the layout for review.

    Shuai

  • 0 in reply to Shuai Fan
    TI__Genius 13451 points

    3. During the manufacture process of PCB, it is any configuration about the programming burning? We saw there is a programming Jumper in SCH. I mean, during programming burning, if input is floating(battery is disconnected.) but force a 5V voltage to the output side of TPS564247, this may cause the damage risk.

    https://www.ti.com/lit/an/slyt689/slyt689.pdf Please refer this application note.

    4. For failure power rail, could you remove the device from PCB and check the whether the internal MOSFET is damaged by measure the impedance between VIN and SW, SW and GND?

    5. Pls provide the layout for review.

    6. You said the startup slew rate is sharp, do you mean it is a hot swap startup operation? And what is the value of the slew rate?

    Shuai

  • 0 in reply to Shuai Fan
    Prodigy 10 points

    1. Thanks for the suggestion we've now added more capacitors in the updated design.

  • 0 in reply to Peter
    TI__Genius 13451 points

    Best regards

  • 0 in reply to Shuai Fan
    Prodigy 10 points

    2. We'll take a look at the regulators when connected to our batteries with an oscilloscope and get back to you.

    3. The Atmega_5V is not directly connected to the 5V power rail of the TPS564247. Atmega_5V is a different power rail created by the regulator shown below and is only used to power our Atmega MCU. We did not supply the Atmega from the 5V TPS regulator because we wanted to be able to turn the regulator on/off in the future via the enable pin. However, this does highlight another good point. We'll add a diode in the next design to the output of our Atmega_5V regulator. We did program the MCU and supply Atmega_5V with external power through the programming pins but the TPS regulators continued to immediately work afterwards and only failed weeks later.

    4. Once we finish test the device under load for point #2 we'll remove it and see if the internal MOSFET has been damaged. Great idea!

    5. I've attached the board layouts below. We know the regulators are not following the optimal design suggested in the datasheet. We don't think this is the issue but we're hoping to route them better in the next iteration anyway.

    6. We are not in a hot swap startup situation. The slew rate was mostly a reach in the dark, because we were out of debugging ideas.PMB_B_Cu.pdfPMB_B-Fab.pdfPMB_F_Cu.pdfPMB_F_Fab.pdf

  • 0 in reply to Peter
    TI__Genius 13451 points

    Hi Peter!

    Thanks for your information, I will help review layout the day after tomorrow. Because we have a vacation tomorrow.

    Shuai

  • 0 in reply to Shuai Fan
    TI__Genius 13451 points

    Hi Peter!

    Your input cap is too far away from IC.

    Shuai