• State
  • Locked Locked
  • Replies 5 replies
  • Subscribers 62 subscribers
  • Views 1065 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.

Buck converter output capacitor failures

JEFFREY BLEDSOE
Intellectual 290 points

I have two buck converters, with Vin = 28V and Fsw ~ = 150 kHz, that are connected together for a +/-15V power supply. The RMS current in the output capacitors (3 X 22 uF tantalums in parallel on each rail) is calculated to be 152 mA, so ~ 50 mA RMS per capacitor.

The failure rate of the capacitors has been higher than expected, possibly due to 25V rating on the caps (slightly underrated voltage for 15V output) and possible cap manufacturing issues. They are being replaced with 35V caps (15 uF) and the results so far are promising.

However, there have been a few cases where all 3 caps on a rail were blown, even with the new and improved caps. Also, I've looked at the failure statistics and there are ~50% higher failures from one set of loads as compared to another set of loads. I haven't done the research to determine the differences in load types and magnitudes. The temperature environment is more benign for the case where more failures occur.

Two questions:

1. What event(s) would blow 3 parallel caps on an output?

2. Could RMS cap current be remarkably higher with some load types compared to others? What load types?

Thanks,

Jeffrey Bledsoe

  • 0
    TI__Mastermind 34465 points

    Hi Jeffrey,

    How do you connect together the two power supplys and what does the schematic look like?

    Ti has an article written on this topic at

    http://www.ti.com/lit/ml/slup094/slup094.pdf

    and you may find it informative.

    Regards,

    John Griffin

  • 0 in reply to John Griffin1
    Intellectual 290 points

    John,

    They're connected in series: the return of the +15V supply is connected to the high side of the -15V supply. The article addresses parallel-connected current-sharing arrangements.

    I don't have a detailed schematic. The schematic I have has some hybrids but I don't have their detailed schematics.

    There are details of a transformer. It's an isolated DC-DC converter, with the input DC power connected to one of the transformer windings.

    I can't upload the schematic as it will be proprietary to the power supply manufacturer.

    Thanks,

    Jeff

    .

  • 0 in reply to JEFFREY BLEDSOE
    TI__Mastermind 34465 points

    Hi Jeff,
    Sorry, my misunderstanding.
    It sounds then like the +15V and -15V loads are independent and presumably both power supplys can regulated independently as well.

    I would monitor the startup voltage and current of both outputs.
    Solid tantalum capacitors have a reputation for startup current surge failure. Perhaps you can soft start the power supply or replace the tantalums with multilayer ceramics or low ESR electrolytics. If this is not possible I would evaluate different tantalum technologies and manufacturers.

    Regards,

    John

  • 0 in reply to John Griffin1
    Intellectual 290 points

    John,

    The supplies have soft start on the output rails.

    To check the soft start function, I made some rail voltage slew rate measurements a few months ago; calculated cap currents based on dv/dt and cap values. Max inrush currents calculated at 275 mA per cap.

    However, I need to repeat the tests on more supplies and perform more startups.

    Thanks,

    Jeff

  • 0 in reply to John Griffin1
    Intellectual 290 points

    John,

    I attended TI's Power Supply Design Seminar last Wednesday (11/19) in Richardson; should have bounced this question off the presenters and attendees.

    Thanks,

    Jeff