• State Resolved
  • Locked Locked
  • Replies 3 replies
  • Subscribers 63 subscribers
  • Views 1367 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.

LM3424: Thermal problem with MOSFETS driving 8 LED's @ 3A

Mario Jimenez
Prodigy 200 points
Part Number: LM3424

Hello,

I want to use the LM3424 to drive 8 LED's @ 3Amps (Cree® XP-L2, 8x2 series-parallel arrangement)with a 24V battery. I bought the LM3424BKBSTEVAL, which already gives 1A output.

I tried an alternate design provided on datasheet (AN-1967), which gives 2.5A output. It indeed worked, but the PCB, MOSFET's(Q1) and main diode (D1) gets hot as hell.

I ran webench but it couldn't suggest a part number for Q1.

I already changed the inductor and diode to a bigger ones, but Q1 keeps getting hot with barely 5 seconds on. I even tried two power mosfets in parallel. I changed Q1 to IRFP048NPBF,  IRFS4115TRLPBF, but both had the same problem.

I also attach my LED array:

8206.SIN000065-01_Helios_2x8_SCH.PDF

Could you suggest me some power mosfets to use in my design? Some thermal dissipation tips/ layout guidance with copper area would be useful too.

Thanks in advance.

  • 0
    TI__Mastermind 25500 points

    Hello Mario,

    The two IRF devices will not give you any better results. With the switch voltage you have most of your losses are not likely to be Rds(on) conductions losses but rather AC losses. Those two devices have very high Qg and a very long plateau in the gate charge versus Vgs graph. So you will want to minimize total gate charge and use a FET with a very small plateau. The EVM FET is actually quite good in this regard. But in some cases where you are trying to drive a lot of power, and this is one of them, you will need to consider running at a lower switching frequency and using a higher value inductor if needed. If you run at half the switching frequency you will cut the AC losses in half as well. This will also help diode power dissipation. 500kHz is a bit high for >60W of output power. You could also use a FET in a larger case to help with the heat dissipation but you still need to be careful about the gate charge.

    Regards,

    Clint

  • 0 in reply to Clinton Jensen
    Prodigy 200 points

    Hello Clinton,

    Thanks for the quick response.

    So my main problem are AC losses. I was looking for a FET with larger case but with similar Qg characteristics & low Rds . I found AOB416. Do you think this FET would do the work?

    Also, what you mean with a higher value inductor? A higher current rated inductor? Or the inductance parameter itself?

    I didn't realized switching frequency could be the problem because almost all examples use 500kHz or more. I'll try cutting it at half and see what happens.

    Thanks!

  • 0 in reply to Mario Jimenez
    TI__Mastermind 25500 points

    Hello Mario,

    That is a pretty good FET according to the datasheet. You still may need to reduce the switching frequency. The AC losses are directly proportional to the switching frequency, but more importantly here to the drain voltage squared. That is why high power, high switch voltage applications often require lower switching frequencies.

    The higher inductor value is optional and I meant inductance value, it was just to keep the inductor ripple current the same with a lower frequency. If you can tolerate more LED current ripple or you can add output capacitance it isn't necessary. Of course the saturation current rating is always important as well.

    Regards,

    Clint