• State Resolved
  • Locked Locked
  • Replies 9 replies
  • Answers 2 answers
  • Subscribers 62 subscribers
  • Views 659 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.

Original question:

TPS40061 KFF pin's application

TPS40055: KFF pin for large duty cycle applications per SLUA310

Nathan Wineera
Prodigy 90 points

Part Number: TPS40055
Other Parts Discussed in Thread: LM5146, LM5145,

I have an application in which Voltage Feedforward needs to be disabled due to limitation of maximum duty cycle as a result of programmable UVLO.

I have followed SLUA310 for this purpose with option 2 via section 4.2 to disable voltage feedforard by tieing KFF to BP5 through a resistor Rsub.

What is the effect of tolerance on Rsub ?. i.e. if Rsub is such that Ikff is lower than that calculated by Eq 7 of SLUA310 ? Can you get an unstable situation with irregular PWM ramp ?

Switching frequency of 566kHz. So Rt is 82k as per Eq (1) of the datasheet. Rkff is 27.5k as per Eq (2) of the datasheet and for Vuvlo (Vin(min)) of 8V.

Is Vin in Eq 7 of SLUA310 nominal Vin, or Vin(min). Assuming Vin(min) of 8V, Ikff is 163.3uA. So Rsub is then 9161 ohms according to Eq 8 of SLUA310. We used 9.1k. Unstable PWM operation is observed looking at the SW node.

My main question is what happens if Rsub is greater than 9161 ohms so that Ikff is effectively less than 163.3uA ? Will unstable PWM ramp generation result ? For reliable operation should Rsub be made much less than 9161 ohms ? Should Rsub be recalculated using Vin nominal in Eq 7 of SLUA310 ? Vin nominal is 24V and Vout is 15V for our application. i.e., if Rsub is recalculated for Vin of 24V then Ikff becomes 745uA and Rsub becomes ~2k.

Feedback would be appreciated.

Thanks.

  • 0
    TI__Guru 64311 points

    Hello Nathan

    I will reply as soon as possible. 

    -EL

  • 0 in reply to EricLee
    Prodigy 90 points

    Thanks Eric

    Just some further information to add. It appears that the particular board that I was testing was unstable due to an untuned compensation loop. Unstable by observing pulse skipping at the SW node and also increased output voltage ripple which appeared periodic. A second board was tested which behaved properly. The gain in the control loop of the original board was reduced significantly and the instability was elliminated. So the control loop now needs to be redesigned to provide more margin.

  • 0 in reply to Nathan Wineera
    TI__Guru 64311 points

    Hello Nathan

    1) I think the higher Rsub will make the Ikff lower and the UVLO level higher. There might be an issue if the UVLO level is too close to the VCC regulation target.
    2) In the section 4.2, 'Vin' in eq7 should be replaced with Vbp5 since a way to disable the feed-forward function and programmable UVLO is discussed in the section 4.2
    3) Unfortunately, the loop can be unstable by using the scheme in the section 4.2.

    Please consider using LM5145/LM5146 if new design. 

    -EL

  • 0 in reply to EricLee
    Prodigy 90 points

    Thanks Eric for your answers. However I am inclined to disagree.

    1. For a given programmed switching frequency there is a minimum programmable Ikff via Rsub for a minimum UVLO of 8V. If Rsub is made larger, then Ikff will be less than Ikff (min) for UVLO (min) and the device will not start (possibly)

    2. With the option of 4.2, BP5 does not track Vin like BP10 does as in 4.1. So Vin replaced by BP5 is incorrect since the device will only start with a minimum Vin around 8V (min UVLO) as per Figure 6-4 of the datasheet. It makes more sense that Vin in eq7 should be Vin (UVLO setpoint).

    3. 4.1 and 4.2 can both be unstable if the loop compensation has not been designed accordingly. The input voltage feedforward is effectively equivalent to current mode control to result in a first order response to simplify the compensator to a type II instead of type III required for normal voltage mode control. If using 4.1, feedforward is only applied up to Vin of 12V. In our application, minimum Vin is greater than 12V so our loop compensation is designed for voltage mode control using a type III compensator to get the required phase boost where needed to ensure stability.

    Will definately consider the newer devices you suggest for new designs. This is a legacy design so we are stuck with using the TPS40055 for now. 

  • 0 in reply to Nathan Wineera
    TI__Guru 64311 points

    Hi Nathan 

    1. Yes, you are right. You need a margin to make the UVLO level higher than the  8V. 

    2. 4.1 is about the idea using BP10. 4.2 is about the idea using BP5. 4.2 idea doesn't provide voltage feed-forward and just uses 8V fixed UVLO. 

    -EL 

  • 0 in reply to EricLee
    Prodigy 90 points

    Thanks Eric

    2. Yes. For 4.1 you get feedforawwrd up to 12V input voltage since BP10 tracks Vin with an offset of 2V as per Figure 6-4 of the TPS40055 datasheet. But if your system minimum input voltage is always greater than 12V you get no benefit from the feedforward feature.

    I have evaluation kits for the newer improved devices you suggest. I will have a play when time permits.

  • +1 in reply to Nathan Wineera
    TI__Guru 64311 points

    Hi Nathan

    Your understanding is right. Both 2.1 and 4.2 idea have a limitation. 

    Thanks for choosing TI.

    -EL 

  • 0 in reply to EricLee
    Prodigy 90 points

    Hi Eric

    Thanks for all your help on this.

  • +1 in reply to Nathan Wineera
    TI__Guru 64311 points

    Thanks for choosing TI. Please feel free to contact us if you have any question.
    -EL