Advertisement

Blog

Effects of Switching Frequency on Transient Response

Here is the first of three short tech nuggets (courtesy of Rich Nowakowski, TI) before our “Ask the experts” session on Wednesday, July 23 at 1:00 p.m. ET (10:00 a.m. PT). Hopefully these brief blogs will help stir your curiosity and questions, which you can post on our chat site here either before the session or during the session.

When designing with DC/DC converters, power supply designers face the compromise between small size and high efficiency. Hopefully, one is more important than the other in the application. Considering the trade-off is important. With advancements in MOSFET and process technology and shorter controllable minimum on-times, the envelope continues to be pushed. Here is an excerpt from an article that discusses the affects switching frequency has on the transient response of a DC/DC converter with 48V input and 5V output.

Transient response

Figure 1

Bode plots at 100 kHz, 300 kHz, and 750 kHz.

Bode plots at 100 kHz, 300 kHz, and 750 kHz.

Transient response is a good indicator of the performance level of a power supply. In this example, a Bode plot of each power supply was taken to show a comparison at higher switching frequencies (see Figure 1). As shown, the phase margin of each power supply is between 45o and 55o , indicating a well dampened transient response. The crossover frequency is approximately one-eighth of the switching frequency. When using a fast switching DC/DC converter, the designer should make sure the power IC’s error amplifier has enough bandwidth to support a high crossover frequency. The unity gain bandwidth of the error amplifier, in this case the TPS54160, is typically 2.7 MHz. The actual transient-response times are shown in Table 1 with the associated values for voltage peak overshoot. The overshoot value is significantly lower with the higher switching frequency, due to the wider bandwidth.

Table 1

Transient response

Transient response

Will improvements to increase switching frequencies affect pre-conceived notions on how DC/DC converters are chosen or used?

15 comments on “Effects of Switching Frequency on Transient Response

  1. Netcrawl
    July 23, 2014

    @Steve great post! thanks for that. The efficiency of DC/AC converter is still one of the most important attributes to consider when designing a power supply, where poor efficiency could translate into higher dissipation which has to be managed on the circuit board with heat sinks. 

  2. samicksha
    July 23, 2014

    I agree you @Netcrawl, i read about release from Intel wherein they were using some oil to keep their servers at good temperature, can we apply this theorem here as well rather than using Heat sink.

  3. amrutah
    July 24, 2014

    Steve: Thanks for the post and I am rueing for having missed the chat.

       Few of the effects of higher switching frequency is increased ripple content in the output voltage and also decrease in the efficiency due to switching loss.

  4. amrutah
    July 24, 2014

    Steve,

      It would be very helpful if you can share the realtion between the switching frequency, ripple voltage and phase margin.  From Table 1, I see that for 750kHz, the PM of the loop has reduced, but the ripple content is also reduced.  I thought lesser the PM, more should be the ripple.

      Is there any kind of superposition happening here?

  5. Myled
    July 25, 2014

    “The efficiency of DC/AC converter is still one of the most important attributes to consider when designing a power supply, where poor efficiency could translate into higher dissipation which has to be managed on the circuit board with heat sinks. “

    Netcrawl, it can create spikes also.

  6. SunitaT
    July 31, 2014

    @Steve, thanks for the post. Is it good to have integrated inductor and capacitors instead of on board even though the current requirements calls for huge inductor and capacitor ?

  7. SunitaT
    July 31, 2014

    @Netcraw, I agree that the DC/AC conversion is the main thing that reduces the efficiency.  If the efficiency can be solved then all the renewable energy sources like solar and wind energy can be used effectively.

  8. Steve Taranovich
    July 31, 2014

    @SunitaT0—There are always that difficult trade-off decision. Larger passive elements in a package means a larger and heavier package. This can be a problem in some cases like those that must undergo shock and vibration environments. It all depends upon the application.

    On the positive side, it is always best to have components as close to the IC as possible for lower stray capacitance, inductance and resistance errors. So all in the same package is a good thing for that case as well as the case of just plugging in one component at assembly and also not having to worry about how to choose the L and C values for best performance.

  9. RedDerek
    August 20, 2014

    @amrutah: Few of the effects of higher switching frequency is increased ripple content in the output voltage and also decrease in the efficiency due to switching loss

    Higher frequency translates typically to lower ripple voltage. But it is true that the efficiency will decrease, primarily at the low load conditions due to the energy going into the switch.

    If lower frequency is needed for higher efficiency, but ripple must stay low, then there is the multi-phase supply topology that is implemented for current computer mother boards. Interleaving a 3 to 8 phase supply can significantly drop the ripple. If the load drops, then, with an intelligent supply controller, you can start to cut out phases to increase efficiency.

  10. Davidled
    August 20, 2014

    One of solution might be adding or changing capacitor connected with resistor. Capacitor might absorb a few noises in charging cycle time. Value of capacitor might be depending upon the frequency of switch.  Please review PWN DAC design comment by Dennis Feucht, on 08/15/14. That might give some reference as understanding the formula.

  11. dassa.an
    August 21, 2014

    @Red Derek: When you transfer from high to low, the levels of the frequency and the signal strengths do get low isn't it? Any method to restore it back because if not it's no use of transferring to lower level isn't it ? 

  12. RedDerek
    August 21, 2014

    @dassa.an – not sure what you are asking in your question. Could you please clarify?

  13. amrutah
    August 21, 2014

    “Interleaving a 3 to 8 phase supply can significantly drop the ripple”

    @redDerek: This is good to counter the efficiency but the system will be more complex.  Having different phases of supply means more of the components increasing the BOM and board area.

  14. amrutah
    August 21, 2014

    “with an intelligent supply controller”

    @redDerek: If you have any examples or whitepapers please share.  It would be immensely helpful to study this.

  15. RedDerek
    August 21, 2014

    Texas Instruments, Maxim, and I believe Linear Technology has datasheets and papers. This knowledge is from my Vishay days when I visited Intel and AMD numerous times. I do not have any current references at my fingertips.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.