• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Subscribers 64 subscribers
  • Views 403 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.

TPS62932: Large, lower frequency ripple on VOUT

Nicholas Junker
Prodigy 20 points
Part Number: TPS62932
Other Parts Discussed in Thread: TPS62933F, TPS62933O

Tool/software:

Hello,

I'm validating the design and layout of a 20V to 12V buck converter using the TPS62932, and am observing some unexpected behavior on the voltage output. At low loads (0-200mA) the voltage ripple shape looks as one would expect, with the ~30mVpp making somewhat of a sawtooth at a lower frequency due to PFM. When the load is increased to 250mA and above, one can observe a large, ~80mVpp sinusoidal waveform at about 10kHz. If zoomed in, you can see that the 10kHz waveform is made up of the normal looking ripple at 1200kHz. I am trying to chase down the source of this large, low frequency ripple. 

I assume that this behavior is somehow related to the buck converter entering DCM mode, but I am unsure if this behavior is to be expected or not. It doesn't effect our final use-case apart from an audible whine, but I wanted to double check if this is intended or not.

Thanks,

Nicholas

AC Coupled waveform with no load:

AC Coupled waveform with 500mA load:

AC Coupled waveform with 500mA load, zoomed in:

  • 0
    TI__Genius 17540 points

    Hi, 

    We are in Chinese holiday, will reply at 9/18. Thanks.

  • 0
    TI__Mastermind 27436 points

    Hi Nicolas

    As the load decreases, the device enters into DCM, and when load is even smaller, it enters into PFM, to have higher efficiency under light load. The Vo_ripple is larger under light load and the frequency decreases. This is normal. The TPS62933F operates under CCM mode across all load range, so there is no larger Vo_ripple and no potential to have audible noise due to the decreased fsw. But the light load efficiency is lower.

    You can check if the Vo_ripple and the audible noise is acceptable to your system, and please note that there is also TPS62933O which has high light load efficiency than FCCM ,and will not has audible noise.

    Thanks and best regards.

  • 0 in reply to Gui He
    Prodigy 20 points

    Hello,

    Thanks for the reply. The reason this throws me off is that the ripple grows in size as load increases, not as load decreases. I.e. at no load, we have 30mA ripple. At 500mA load, we have a sinusoidal 80mV ripple, which increases to almost 100mA at 2A load. The large ripple is also a constant frequency (~9.5kHz) across all loads, not adjusting to match load. This goes against the expected behavior from PFM.

    I did stumble across the thread linked below where another user had this issue and traced it back to the spread-spectrum feature, which modulates at the exact frequency I am seeing on the ripple. I also found a TI whitepaper that seems to confirm this is related to the spread spectrum implementation, specifically saying that the spread-spectrum feature will often leave a voltage ripple on the input and output at the exact modulation frequency. Leaving those links here for others who may come across this thread. I have the TPS62933F on the way to check if the ripple disappears.

    TI E2E thread about same issue

    TI Whitepaper about Spread-Spectrum pros and cons

  • +1 in reply to Nicholas Junker
    TI__Mastermind 27436 points

    Hi Nicholas

    The ripple should be caused by the spread spectrum, the modulation frequency is fsw/128, TPS62933F has no spread spectrum, should be able to solve your problem.

    Thanks and best regards.

  • 0 in reply to Gui He
    Prodigy 20 points

    Thank you for the reply - received samples of the TPS62933F and can confirm it solved the issue. Scope shots are attached for others who may stumble across the thread. Ripple is now between 20-25mA across all load conditions. If possible, I'd suggest adding a statement to the datasheet mentioning this behavior, since it is not mentioned anywhere despite TI having a whitepaper specifically about it.

    AC Coupled Waveform with no load:

    AC Coupled Waveform with 2A load: