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.

PFCLLCSREVM034: Regarding the LLC gain curve selection

Part Number: PFCLLCSREVM034

Dear Sir/Madam,

I am hoping to modify PFCLLCSREVM034 to output 30V/5A/150W so I am currently design a LLC stage by the tool UCC25640x Design Calculator Rev4.0.

I will replace the transformer and the resonant inductor Lr will be replaced by leakage inductor of transformer. 

Here are some design results, their transformer parameters are different so the LLC gain curves are different as well.

1) 24V/8A transformer. operating frequency 70k-120k

In this curve, LLC switching frequency range is 72.3k-82.6kHz. The gain is between 1.302-1.510

2) 48V/5A transformer. operating frequency 70k-120k

In this curve, LLC switching frequency range is 207.6k-263.3kHz. The gain is between 0.67-0.77.

At this moment, I choose solution 1. The reason is LLC switching frequency is in the transformer operating range.

 

But I am hoping to know, is second solution acceptable IF transformer operating frequency range is very wide or cover the LLC switching frequency?

Moreover, I would like to ask is there a most suitable gain range when we are design the LLC stage?

Thank you for answering.

Regards,

  • Hi Sean,

    The first solution is good. The 2nd solution is not acceptable as the gain values are always less than the no load gain curve during high frequency range. Due to this, the converter will operate in burst mode during the higher loads.

    It's always better to operate the converter around the resonance where the gain is 1 (assuming discrete magnetics). Because converter will have less losses at both primary and secondary sides at the resonance.

    when you design the LLC converter, it's better to fix the operating frequency (same as resonant frequency) first. based on this, you can easily determine the number of turns of transformer. And then you need to select Ln,Qe based on the gain requirements. However, the design calculator automatically does these calculations for you. 

    The above steps are well explained in this document: Designing an LLC Resonant Half-Bridge Power Converter Article

     Please let me know if you have any further questions.

    Regards

    Manikanta P

  • Hi Manikanta,

    Thank you very much for the detailed explanation. It is very helpful for me to understand LLC design. but I have 2 more questions here.

    1) I totally agree with your comment "It's always better to operate the converter around the resonance where the gain is 1 (assuming discrete magnetics)."

        When I design LLC by UCC25640x Design Calculator Rev4.0, the gain is always very close to 1 (maybe 0.9~1.1) IF the transformer turn ratio is ideal (the same with the calculated result). However, I didn't use the calculated result because I can only find the existing transformer from market (Würth). That's why the gain in the solution 1 and 2 are higher than 1 or lower than 1 too much. I didn't use a suitable transformer for 30V output. Those transformers are used for 48V and 24V.

        But here is a question. You mentioned about " gain is 1 (assuming discrete magnetic)," What if the resonant inductor is NOT a discrete inductor? In my case I use the leakage inductor of transformer as resonant inductor. Is it still better setup gain around 1?

    2) I just realized that I made a mistake on leakage inductor. the transformer datasheet provides Lm and Lr values as shown below.

       

    I assume the transformer model of the above table should be the same with below picture (from the design document you provided). Please assume that Lr is LLk (leakage inductor). Then LLk and Lm values in the picture are 95uH and 500uH from above table. Am I right?

    However, the transformer model in the UCC25640x Design Calculator Rev4.0 is shown below

    I think Lm is still the same value. But How can I convert this model to the first model? is the Lr value in the second model equal to half of the Lr value in the first model? In this case is it 47.5uH + 47.5uH ?

    thank you very much for your help.

  • Hi Manikanta,

    I am sorry that schematic from the UCC25640x Design Calculator Rev4.0 is below

    Maybe I don't need to consider about it too much. Just directly use the Lm and LLK values from transformer datasheet as Lm and Lr? It is correct?

  • Hi Sean,

    Since you are using integrated transformer, you can consider below model for LLC design. 

    You can find the values of this transformer by using the formulas given in above screenshot.

    here Lp=500uH, LLk=95uH, n=(35/4)

    So, from above values, k= 0.9, Lr1=50uH,Lm=450uH.

    When you have the integrated transformer model, gain at resonant will be in the range of 1.1-1.2 which depends on the k value. 

    The LLC gain derivation when you use the integrated model is given in the following link: Analysis and Design of LLC Resonant Converter with Integrated Transformer | IEEE Conference Publication | IEEE Xplore

    If you want to convert the parameters from integrated model to simple APR (all parameters primary referred) model or vice versa, here are the good references:

    Design guideline for magnetic integration in LLC resonant converters | IEEE Conference Publication | IEEE Xplore

    https://verimod.com/presentations/Magnetic_Coupling_IEEE_Seattle_PELS.pdf

    Regards

    Manikanta P

  • Hi Manikanta,

    Thank you for answering my questions and providing so many useful articles. then I would like to double check one thing. Magnetic components manufacturers usually provide component parameters. When they say "inductance,"  does it mean Lm (magnetic inductance) or Lp (primary side inductance)?

    I also calculated the Lr by formulas in the sheet last week. But I used L=500uH as Lm directly, not as Lp, and LLk=95uH as well. Then I solved those equations to obtain K=0.9087. Lr=49.77uH. These two values are very close to yours, but Lm is totally different (500uH and 450uH).

    May I ask which one is correct?

  • Hi Sean,

    Ideally the magnetic component vendors should mention the test conditions along with the inductance values (you can reach out to customer support to know the test conditions). If they don't, you can make the measurements in the lab.

    For example, in the above screen shot, inductance values are mentioned along with the test conditions.

    Here, test condition is that inductance is measured between 3 and 4 and all other windings are open. Which means it is open circuit primary inductance (Lp). Similarly, leakage inductance is also measured between  3 to 4 by short circuiting all other windings. which means this is representing LLK.

    So, I followed the similar method as above to calculate Lm and Lr. 

    Regards

    Manikanta P

  • Hi Manikanta,

    Thank you very much. your answer is very clear and helpful!

  • Hi Manikanta,
    could you please elaborate more about the first sentence (The 2nd solution is not acceptable as the gain values are always less than the no load gain curve during high frequency range. Due to this, the converter will operate in burst mode during the higher loads)?

    In the high frequency region, the no-load curve will always be higher than the gain attainable in any other load condition. Do you mean that the no-load curve must intercept the minimum required gain in order for the controller to work fine?

    I ask because on any prototype we built so far, anytime the required gain from LLC stage drops below unity, controller start skipping pulses and output voltage regulation is lost. Basically, the converter is working in burst mode, even if a continuous switching with the proper frequency will guarantee the desired output voltage.


    We got around that issue by increasing the transformer turns ratio, but we are still trying to investigate reason behind that behavior. Your explanation seems a plausible cause, so we would like to get a deeper understanding.

    Many thanks
    Francesco



  • Hi Francesco,

    Its good to hear from you.

    The following slides should give you a better picture.

    Here you can see how high frequency gain is effected by the value of k (Lm/Lr ratio)

    Here you can see the design constraint at no load (see the last point in the following slide). So, from the gain curve shown in the below slide, if the minimum gain doesn't intercept Q=0 gain curve, that means gain of the design is always high at the high frequency region (f> fr1 series resonant frequency). Due to this, converter always operates at the burst mode when the voltage is high and when load is low.

    Please let me know if you have any further questions.

    Regards

    Manikanta P

  • Hi Manikanta

    thanks for your prompt reply and nice explanation.


    One thing is not clear to me: is that constraint inherent to the LLC topology or is it related to control implementation?
     

    Best regards

  • Hi Francesco,

    It is inherent to the LLC topology.

    Regards

    Manikanta P