• State Resolved
  • Locked Locked
  • Replies 8 replies
  • Answers 3 answers
  • Subscribers 63 subscribers
  • Views 506 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.

UCC256302: ways to avoid burst mode

Hiroshi Doi35
Expert 2845 points
Part Number: UCC256302

Hi Ben-san,

My customer's question is continued. Please answer to below.

Industrial market customer want no burst.

==

I am looking for ways to avoid burst mode.

Then, although it is a matter of adjusting the peripheral constant whether to enter the burst mode at the end or not, please do the following question.
 
Question 1.
In the data sheet, the control frequency range of this IC is the upper limit of 35 kHz to 1 MHz.
This is because when the load is not loaded (minimum load), output or input voltage change
Is it said that IC monitoring allows continuous control up to 1 MHz?
 
Question 2.
Why is it shifting from continuous control to burst mode operation when raising only the input voltage even when setting continuous control operation at no load at present?
(The output is not a synchronous rectification method, and it is confirmed also by the diode rectification method.)
 
Also, is this solution difficult as long as the Np / Ns turn ratio of the conversion transformer is not optimally adjusted?
In short, does this IC have a narrow dynamic range of continuous output power?

  • 0
    TI__Mastermind 27135 points
    Hi Doi-san,

    Nice hearing from you.

    1. Typically, you need to push the switching frequency much higher during startup in order to achieve soft start and get a smooth ramp up of the output voltage. I would say for most designs, ~300kHz to 350kHz resonant frequency is possible. This will leave enough margin in order to push the switching frequency greater than resonance to get soft start and smooth ramp up of the output voltage.

    2. The burst mode threshold will change with input voltage because as you increase the input voltage, the magnitude of the resonant current decreases and the AC voltage on the resonant capacitor will also decrease in magnitude. This means the VCR peak to peak voltage will decrease and make burst mode more likely. If you want to avoid burst mode, I suggest first depopulating the resistor between RVCC and LL/SS. Next, you will want to decrease the bottom VCR capacitance. What this will do is make the peak to peak VCR voltage larger. To avoid burst mode, you want the peak to peak VCR voltage to be greater than the burst threshold (BMT=0.7V when the resistor between RVCC and LL/SS is depopulated). The IC itself is not limited in its dynamic range of output power.

    Best Regards,
    Ben Lough
  • 0 in reply to Benjamin Lough
    Expert 2845 points

    Hi Ben-san,

    My customer's return:

    Based on the advice, the voltage-dividing resistor between the LL / SS terminal and RVCC was removed, the operation test of light load to no load was performed again but the burst mode could not be avoided.
    Also, regarding reducing the capacity of the lower capacitor of the VCR pin, we do not know how far it is to reduce the capacity, so we are concerned about the other actions.
    Currently the evaluation is in progress with the default 0.015 μF lowered to 0.01 μF.
    There was also memory that loop control became unstable when operating this with the capacity of this VCR further reduced by rush current countermeasure before, this is not implemented again.

    And I want ask for you:

    Can you stop burst mode perfectly by your EVM at 1W load?

  • 0 in reply to Hiroshi Doi35
    TI__Mastermind 27135 points
    Hi Doi-san,

    Both the top and bottom VCR capacitors would need to be reduced in order to avoid burst mode. After VCR is adjusted, it is likely that the compensation would also need adjustment.

    I will need to test the EVM in order to determine what values of VCR are sufficient to get rid of burst mode.

    Best Regards,
    Ben Lough
  • 0 in reply to Benjamin Lough
    TI__Mastermind 27135 points
    Hi Doi-san,

    With R1=DNP, C9=100pF and C10=5.6nF, I am able to run UCC25630-1EVM at 1W load with no burst mode. There is a little bit of instability at light load but this is because I did not adjust the compensation.

    If you want to send me your customer's schematic, I can suggest VCR and compensation values to help avoid burst mode at light load.


    Best Regards,
    Ben Lough
  • 0 in reply to Benjamin Lough
    Expert 2845 points

    Hi Ben-san,

    Your recommended "R1 = DNP" means BLK zero DC Level?

  • 0 in reply to Hiroshi Doi35
    TI__Mastermind 27135 points
    Hi Doi-san,

    Sorry, I meant to say the resistor between LL/SS and RVCC. R14=DNP not R1.

    Best Regards,
    Ben Lough
  • 0 in reply to Benjamin Lough
    Expert 2845 points
    Hi Ben-san,

    My customer's comment is,
    ===
    We have modified the instructions to the unmodified EVM that we have.
    Then turn on the power.
    Next we slowly increased the load current from 0A.
    Then we measured the load current from burst mode to continuous mode.
    The results are as follows.
     
    Unmodified: Continuous operation current of minimum load _ _ 0.25 A (2.97 W)
    At remodeling: Continuous operation current of minimum load _ 0.19 A (2.26 W)
    At this time, the output voltage is 11.90 V.
    Furthermore, when starting up the power supply, once it enters the burst, it shifts to continuous, so the load current required 0.2 A (2.38 W) as a result.
     
    Therefore, improvements
    2.97 - 2.38 = 0.59 W
    And there was no effect up to the expected 1 Wmin.
    Is this the variation of the capacitance value of C9, C10 used or the individual difference of EVM?
    In any case, it was far from the objective that we wanted continuous motion.
    ===

    my comment is,
    They want continuous mode with load zero W.
    "Because it is impossible in principle, how much preload is allowed?" I'm asking them.
  • 0 in reply to Hiroshi Doi35
    TI__Mastermind 27135 points
    Hi Doi-san,

    I've sent you an email.

    Best Regards,
    Ben Lough