• State Resolved
  • Locked Locked
  • Replies 4 replies
  • Subscribers 62 subscribers
  • Views 8704 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

UC3843: How does the RC parameter of the current sampling pin of the UC3843 affect the loop stability?

zoujiangyilang
Genius 4560 points
Part Number: UC3843

Hello:

I made a flyback power supply in the same way as the reference circuit of UC3843. The block diagram is similar to the following:

In my actual circuit diagram, Rcsf=1k, Ccsf takes 330uF~1000uF. I found that the power supply had a self-oscillation when the Ccsf took 1000uF. The output ripple is increased from 30mV to 90mV, and the oscillation frequency is about 30kH. When Ccsf takes 330uF, the power supply is normal, the ripple is 30mV, and the frequency is 400kHz. The switching frequency is 400kHz.

I looked at the manual for the Rcsf, Ccsf value description. As shown below:

I see the manual only from the perspective of RC filtering, saying that the time constant of RC is less than the switching period.

From this point of view, Ccsf should take 330pF~1000pF to meet the conditions.

It should be the switching frequency fs=400kHz, and the switching period Ts=2.5us.

Rcsf=1k, Ccsf=1000pF, RC time constant t=RcsfCcsf=1us; Rcsf=1k, Ccsf=330pF, RC time constant t=RcsfCcsf=0.33us; all are smaller than the switching period T.

But why does Ccsf take a little bigger and cause self-excitement?

I suspect that Rcsf and Ccsf are actually in the inner loop of the current, which will affect the performance of the current inner loop.

However, I don't know how to quantify the parameters of Rcsf and Ccsf.

The analysis and calculation of the loop compensation parameters in the manual are only the harmonic compensation and the op amp-based compensation parameters.The parameters of Rcsf and Ccsf are not calculated, or Rcsf and Ccsf are not considered in the loop compensation.

However, the value of Ccsf in practice does affect stability. In the flyback power supply I made, Ccsf takes 300pF power supply stable, Ccsf takes 1000pF power supply instability, how can this be analyzed?

In addition to the RC time constant is less than the switching period, are there other restrictions for the parameter design of Rcsf, Ccsf?

Need to consider the current inner loop? How to analyze the calculation of the current inner loop? It feels very complicated. The examples in the manual do not seem to consider the design of the current inner loop compensation parameters.

  • 0
    TI__Mastermind 34465 points

    Hi ,

    You need to distinguish between uF and pF.
    All of the units used for Ccsf are in pF..

    The waveform on the ISENSE input needs to be a ramp signal because this signal and the COMP signal are used in the PWM comparator to generate the pwm OUTPUT gate drive to your power supply.
    If you make Ccsf too large then you will not be able to generate a clean ramp signal input  to the PWM comparator..
    The value of Rcsf and Ccsf is chosen as the minimum value required to filter out any high frequency edges due to circuit parasitics and pcb layout traces.
    You should check the ISENSE signal with a scope and verify that it is a clean input 

    Regards

    John

  • 0 in reply to John Griffin1
    Genius 4560 points

    Hello John，

    Thank you for your advice.
    It was wrong before, they are all pF capacitors.
    It seems that the main function of Ccsf and Rcsf is filtering, which has little effect on the loop.


    Later, I adjusted the value of Cvcc from 3.3uF to 4.7uF, and the self-excited phenomenon disappeared. I observed the ripple voltage on Cvcc. It doesn't make much difference between 3.3uF and 4.7uF, but it is true that after Cvcc is increased, the power supply is stable.
    I don't know why? A power supply filter capacitor can affect stability. I suspect that this should not be self-excited due to insufficient phase margin in the loop, which may be caused by unreasonable circuit state design. However, I really can't figure this out.


    When vin is greater than 35V, the input current average is below 60mA, and the primary current peak is 0.5A.

    Regards

    ZJYL

  • 0 in reply to zoujiangyilang
    Genius 4560 points

     I thought it clearly Later, I used the auxiliary power supply winding feedback, not the optocoupler feedback.

    Increasing the supply capacitor Cvcc actually increases the output capacitance on the loop, so it has an effect on the loop characteristics.

  • 0 in reply to zoujiangyilang
    TI__Mastermind 34465 points

    Hi ,

    Yes that will certainly affect the loop response.

    If you wish to update your schematic with the latest changes you will have a better idea of what is happening.

    At the moment Cvcc is shown as 120uF and the auxiliary winding is only used for bias power.

    Regards

    John