• State Resolved
  • Locked Locked
  • Replies 3 replies
  • Answers 1 answer
  • Subscribers 63 subscribers
  • Views 268 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.

TPS53355: Is the capacitor from LL to VOUT in the ripple injector really wise?

Peter Schmidt-Nielsen
Prodigy 20 points
Part Number: TPS53355
Other Parts Discussed in Thread: TPS54KC23

Tool/software:

Let's take a look at the application schematic:

Specifically, let me call your attention to C1, C2, R7. These components are referred to as the ripple injector, and are recommended in the datasheet if you have low ESR MLCCs on the output. I have placed these components, and am getting relatively good control loop stability up to 20A, but I am seeing more ripple than I'd like at the output (about 100 mV pp on this 1.2V rail), and I can't help but think that C1 is partially to blame. R7 and C1 together form a path with impedance 3kΩ + 10MΩ/jω from LL directly to VOUT, which seems ill-advised, and I assume is partially responsible (although I haven't tried desoldering C1 yet, as I don't have very many boards to spare).

Can you folks chime in on the purpose of C1 specifically? I understand how R7 and C2 together increase the ripple on VFB, but I'm not understanding the purpose of C1. Is C1 just about transient response, and giving a faster path from VOUT back into VFB? If so, can I omit C1 if I'd like to trade off worse transient response for less ripple and more control loop stability?

Overall, I should say that while the TPS53355 seems to be pretty good, I have found its control loop to be more sensitive than I would like, ideally. I am driving some sensitive digital FPGA rails, and need very low ripple. Can you folks recommend any other TI parts I could look at that can go down to 0.85V, do at least 30A, and have rock-solid "idiot proofed" control loops? I would much rather give up some transient response to get a rail that never goes down.

Thanks,

-Peter

  • 0
    TI__Genius 11155 points

    Hello Peter,

    I will respond to this in few days.

    Thanks,

    Calan

  • 0
    TI__Genius 11155 points

    Hi Peter,

    Hope this helps.

    If this helped answer your question, please don't forget to click "This resolved my issue".

    Thank you,

    Calan

  • 0 in reply to Calan Elaya
    Prodigy 20 points

    Got it, thanks for the info. Based on reading that document, it sounds like C1 (called C_A in that document) is not just about transient response, and actually is about producing an appropriate ripple? So I guess I shouldn't be removing C1? Note that I assumed it was about transient response, so I hooked up C1 to the output all the way at the very end, after all of the bypass caps. Here you can see the little RCC network highlighted in green, and that long highlighted trace is me hooking up C1 to VOUT at the very edge of the board. I assume this is pretty bad, and C1 should be connected directly to inductor's output?

    Thanks for recommending that other part, it does indeed look pretty good for me, and nice and simple. I'm curious when it'll be in stock at JLCPCB, but I guess I can just order some there.

    Thanks,

    -Peter