• State TI Thinks Resolved
  • Locked Locked
  • Replies 4 replies
  • Answers 1 answer
  • Subscribers 63 subscribers
  • Views 1090 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.

PSFB: impact transformer/FET resonance frequency

evs
Genius 5910 points
Other Parts Discussed in Thread: UCC28951-Q1

Hi,

I'm designing the transformer for my PSFB. I'm selecting the leakage inductance and  I'm wondering what is the impact is  of the resonance frequency?

I expect that the resonance frequency is the transition time to switch between the switching states.

 I have thr resonance freq. at 2.5Mhz ->400 ns. So if that is the transition and the switching speed is 250khz I only have a duty of only  0.8(1 -  400ns/2000ns).   That is a bit much. I can raise the resonance freq. But then I enter PSFB at 60% of max current. So that is also not very useful.

Is is this correct?

Any suggestions how to solve this? lowering switch freq. Adjust transformer turn ration.

Thanks!

Edit: I realized I didn't make it clear max duty = 50% . So a duty of 0.8 is actually 0.8 * 0.5. = 0.4

  • 0
    TI__Mastermind 31580 points

    Hi,

    If the resonant frequency is 2.5MHz, then the period is 400ns. The time taken to complete switch the switch-node from say the positive input rail to 0V is 1/2 the period, so 200ns.

    In addition the resonant frequency depends on the leakage inductance and the circuit capacitance or switch node capacitance which is dominated by the output capacitance, Coss, of the power devices in the bridge.

    In some cases, when using super junction mosfets, the Coss is non-linear. At higher voltages the Coss value is much lower and consequently the swith-node voltage will have faster dv/dt than when the drain voltage on the device drops below say 30V for a 650V device. Hence the transition time can be less than calculated if a fixed value of Coss is used.

    Also the efficiency hit can be minimal if the power devices are turned on with incomplete ZVS once the voltage across the device is significantly reduced during turn on.

    Note at 250kHz the transformer design will require closer attention to minimize core loss and copper loss. There are a couple of sections in the TI library on magnetic design which may be useful,see link below.

    www.ti.com/.../login.shtml

    Regards

    Peter

  • 0 in reply to Peter Meaney
    Genius 5910 points

    Peter thank for the reply!

    Of course it is 1/2 of the duty. 

    I'm aware what the choices are for the design of  forward transformer. I don't know what the impact is for different design choices for a ZVS forward transformer. I didn't find any good documentation all of it didn't go in that much detail.

    I hope you can answer some of my questions:

    The design : 1kw 375/350V with full brigade rectifier. PSFB. transformer is PQ 40/40 N97 ,  controller  C2000 with average current control.

    - For a PSFB this size 250k is not a strange value?

    - There are 4 Coss in a full bridge 2 of them a shorted. So the resonance freq = 1 / (2*PI*Sqrt ((2*Css + Ctransformer)+Ls)). Is this correct?

    - The energy stored in LS needs to be bigger then 2*Coss + Ctransformer  for ZVS?

    - Why isn't there any mentioning of the impact o the bigger f Ls in the transfer ratio?

    - What is normal duty cycle for a PSFB in normal operation.

    Thanks!

    Edit: also imported to note: I like a 2kw peak power. Thermal limited.  But I'm under the impression that ZVS is not very suitable for peak power.

  • 0 in reply to evs
    TI__Mastermind 31580 points

    Hi

     

    I was OOO for a couple of days, apologies for the delay.

     

    250kHz is higher than many of the applications we see, typically they are 100kHz or less. This keeps the fundamental switching component beneath 150kHz where the EMI standards start. The choice of wire size/type for the transformer winding is a little easier as the skin depth is larger at 100kHz Vs 250Khaz so the trade off between AC resistance and DC resistance is less at 100kHz versus 250kHz. The core size of the transformer will be reduced at 250kHz which will be a benefit for high density applications.

     

    We have an analogue PSFB controller, the UCC28951-Q1, and it has the formulae for calculating the LS, (61), including the Coss. There are are also links to a Mathcad calculator in the datasheet, http://www.ti.com/lit/ds/symlink/ucc28951-q1.pdf

     

     

    The normal duty cycle during normal operation will be determined by the minimum input voltage at which the power stage needs to keep the output voltage in regulation, this point of operation will set the maximum duty cycle and turns ration for the design. As Vin is then increased from the minimum input value then the duty cycle will decrease. The variation in duty cycle will follow the input voltage range.

     

    Regards

     

    Peter

  • 0 in reply to Peter Meaney
    Genius 5910 points

    Peter,

    Thanks for your reply