• State Resolved
  • Locked Locked
  • Replies 8 replies
  • Answers 2 answers
  • Subscribers 65 subscribers
  • Views 279 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.

TPS543C20: Operation Mode

user4811888
Intellectual 530 points
Part Number: TPS543C20

Hi expert,

1. The API feature is used in load-step up, it claims that control circuit senses both speed and amplitude of undershoot.

Does it mean it will sense both dV/dt(slew rate) and delta V(compared with threshold) together, and these two elements should satisfy "AND“ relationship?

If so, the dV/dt threshold are fixed(we cannot modify the sensitivity of voltage change speed), right? Any spec of this slew rate?

2. Once API is triggered, it looks there'll be only one extra PWM pulse inserted(1 extra pulse during one switching cycle).

It is possible that >1 PWM pulses will be inserted(find API in several switching cycles)?

3. The BB feature is for load-step down, does the overshoot threshold setting totally the same as API?

  • 0
    TI__Guru 56925 points

     

    1) The Advanced Pulse Injection (API) feature is triggered by a combination of dV/dt and Delta-V but it is not a simple "AND" logic function.  The sensed feedback voltage is compared with a rolling average of the feedback voltage to create a differential gain signal which allows the API circuit to detect a smaller sudden drop than a slower drop in VOUT.

    I do not have the exact bandwidth of the averaging circuit to provide the minimum dV/dt required to trigger API, but due to the averaging, slower transient will require greater amplitude to trigger an on-time pulse.

    2) API on the TPS543C20 is limited to 1 additional pulse per switching cycle, this limits the resulting change in frequency and gain during an API event.  If multiple pulses are allowed, there are additional stability concerns related to the shift in gain and the reset time after a pulse is generated before another API event can be detected.

  • 0 in reply to Peter James Miller
    Intellectual 530 points

    Hi Peter James,

    Thanks very much for answering, it is more clear to my understandings！

    Still interested in below three:

    1. As stated in my ticket no.3 question, what about BB triggering dv/dt and threshold? similar with API?

    2. If you only consider about load transient, it seems dV/dt is not required to sense. Because amplitude of vo change is able to reflect load transient events happening. What kind of special working condition should be considered, to avoid mis-triggering by amplitude sense only?

    3. Related to no.3, if I adjust Vo by manually injecting voltage source into feedback loop suddenly, it is likely possible to mis-trigger API/BB, right?  

  • 0 in reply to user4811888
    TI__Expert 7500 points

    Hi,

    Peter is looking into this and will feedback to you soon.

    Thanks,

    Lishuang

  • +1 in reply to user4811888
    TI__Guru 56925 points

     

    1) Yes the Body Breaking Overshoot reduction circuit uses a similar current versus averaged feedback voltage threshold to trigger the turn-off of the low-side FET in order to increase the voltage drop across the inductor and reduce the inductor current more quickly in order to reduce the overshoot during a rapid drop in VOUT.

    2) The current - averaged feedback sense circuit is used to make the detection circuit less sensitive to offset and drift between feedback amplifier that is maintaining the accurate target voltage and the API/BB comparators.  Without this averaging circuit, comparator offsets could significantly change the sensitivity of the detection circuits which need to trigger quickly with very low input override to be effective.

    3) If rapidly changing the output voltage by dynamically adjusting the feedback divider or injecting/removing current from the feedback pin it is possible to trigger API or BB, yes.  If that occurs, the triggering of API or BB will allow the TPS543C20 to track the change faster than it would without these features.  In this case, it would not be considered a "mis-trigger" of these features, as the features are doing their job of speeding up the TPS543C20's ability to quickly track the changes commanded by the external manipulation of the loop.

    It is generally recommended that external injection into the feedback or dynamic changes to the feedback divider should have external slew-rate control in order to keep the loop within its regulation bandwidth during the dynamic change in order to prevent triggering PGOOD and/or OV/UV.  Such slew-rate control should also prevent API/BB triggering.

    NOTE:  Rapid decreases in VOUT via manipulation of the feedback divider or injecting current into the feedback loop can result in very high negative inductor currents as the TPS543C20 attempts to discharge COUT rapidly, this negative current will flow through the high-side FET back into the input node.  If there is not a discharge path for this energy it can charge the VIN bypass capacitor and create a potential for a damaging over-voltage on the input and/or switching node of the TPS543C20.  For this reason in particular, slew-rate control when decreasing the regulated output voltage is particularly important, especially when testing a single power converter in a laboratory setting where input source impedances are higher due to long power leads, and there are not other supplies available to dissipate energy recovered from the output capacitor during a commanded decreases in the output voltage.

  • 0 in reply to Peter James Miller
    Intellectual 530 points

    Thanks very much, Peter.

    2）About the second point, I agree that dv/dt is greatly helpful to detect transient timing,  so combination of speed and amplitude should be good enough because:

    ①If only delta V is used, it is enough functionally, but, offset and drift might cause false trigger, as you told.

    ②if only dv/dt is used, it is the best to react at the very beginning when Vo changes, but, very slow but large-step load transient might not be identified.

    Is my understanding completed of pros and cons in this comparison?  

    3）Understand what you are pointing about, API/BB should behave.

    About your reminder of negative current, yes I'm aware of that. I will be careful if I use external control to tune down Vout. 

  • 0 in reply to user4811888
    TI__Guru 56925 points

     

    Your understanding is generally correct on the potential issue of using only an absolute or delta voltage measurement without a dV/dt component.

    The issue with using only a dV/dt measurement with no absolute value component is the sensitivity of such a circuit to noise, especially the normal switching frequency ripple and high-frequency switching noise typically seen on the output voltage.  If API and BB triggered solely on dV/dt with no consideration of the amplitude, both circuits would be continually seeing false triggers.

  • 0 in reply to Peter James Miller
    Intellectual 530 points

    Got it, thanks very much!

  • 0 in reply to user4811888
    TI__Expert 7500 points

    Hi,

    We will close this thread. If it solved your question, please click the resolved button.

    Thanks,

    Lishuang