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.

LM5161-Q1: Fly-Buck transformer heating

Part Number: LM5161-Q1
Other Parts Discussed in Thread: LM5160DNTFBK-CALC

Hi TI,

I have an application which will use LM5161-Q1 in a fly-buck configuration. I'm testing a few custom transformers and have the design working but I'm not sure if the temperature rise of my transformer is "normal" for the given load.

Vin: 12-50

Vout1 (N1, non-iso): 5V

Vout2 (N2, non-iso): 12.7V

Vout3 (N3, iso): 6.6V

Vout4 (N4, iso): 12.7V

Freq: 210kHz

Transformer specifications:

L(pri): 60uH

Isat: 1.4A

I(N1): 1A

I(N2): 100mA

I(N3): 1A

I(N4): 100mA

DCR(N1): 125m

DCR(N2): 125m

N1:N2 = 1:1.34

Test procedure:

Vout1 loaded with 200mA

Vout3 loaded with 600mA

Vout2 and Vout4 were not used. The transformer temperature rise was about 35°C from an ambient temp of 24°C. This seems excessive to me. The application will run in high ambient temperatures so a 35°C rise with a little more than half the rated current on the isolated side is too much. The saturation current is on the low side but it is higher than the min. high side current limit of 1.3A. I have run the numbers in the LM5160DNTFBK-CALC and max. current+ripple do not exceed 1.3A.

I also tested one of our flyback designs with a much smaller transformer and loaded it with even more than the Fly-Buck and the temp. rise was minimal. What could be causing the transformer to heat? The copper losses for the given load shouldn't be that high, could the core losses be the main heating source? And if so, why?

Thanks,

Tilen

  • Hello Tilen,

    I am forwarding your question to someone who is good with Fly-Buck converters.

    Regards, Robert
  • Hi Tilen,

    Yes the core losses could contribute to this heating. Seems like you have already reduced the frequency quite low to 200kHz to minimize some of the switching loss.

    Do you have the schematic of the design? Could you try to run a higher frequency to minimize the peak to peak ripple current and see if it improves your thermal?

    You can compare the transformer with the one from this EVM. In terms of size,DCR and leakage

    Thanks

    -Arief

  • Hi Arief,

    Do you have thermal data of the EVM board?

    Yes, I tried different frequency's (100-400kHz) but the temperature difference was minimal ~3°C. The biggest difference in temperature rise was today when I replaced the secondary Schottky with a one with a significantly smaller capacitance and this dropped the temp. rise by 5°C. So the temp. rise with the new diode is now ~30°C from ambient, but even this is too much for such a small load. I just noticed that my transformer supplier mentioned in the datasheet that the transformers are for flyback and not Fly-Buck? I said I need them for Fly-Buck but they could have missed this specification. Is there a difference between a flyback and Fly-Buck transformer?

    The schematic is my test setup. RON is of course different for my test freq. of 208kHz.

    Thanks,

    Tilen

  • Hi Arief,

    I still haven't solved this issue. Do you have any suggestions on what to do, to bring down the transformer temperature?

    Tilen
  • Hi Tilen,

    Do you have the datasheet for the transformer and how big is the physical size for the transformer? When you switch the diode the temperature drop is because the reflected current charging and discharging capacitance is amplified on the primary side.

    So the temperature increase that you have seen is when the other secondary side is not loaded?

    Thanks
    -Arief
  • Hi Arief,

    The temperature rise is when I load the secondary with 0.7A and the primary with 0.2A.

    Datasheet:

    Thanks,

    Tilen

  • Hello Tilen,

    The transformer needs to be a flyBack transformer, so I think it is suitable (somewhat). the issue I see is, the leakage inductance, this need to be kept at a minimal level as a percentage of Lpri. Recommend using a Lleakage of 1% or lower, you are currently using a transformer who's leakage is much higher than this (3%++). This is not acceptable for your needs. This causes a very large negative current in the transformer during the 1-D period which increases core losses. Also as mentioned the step up transformer is not helping you, when taking into consideration the Cj of the rectifying diode (as Arief alluded to), suggest using an UF diode with very low Cj.

    I think the biggest issue is the leakage inductance on the transformer. Please select alternative and retry.

    hope this helps?

    David.
  • In addition to David's comments, here are some transformer that we have used on our Flybuck design

    www.ti.com/.../slpt047.pdf

    Thanks
  • Hi David,

    Thank you for the explanation. I'm already waiting for some new transformers will keep my eye on the leakage inductance. I will also try your diode suggestions.

    Tilen.