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.

LM2574: 450mA / 37V Test

Part Number: LM2574
Other Parts Discussed in Thread: LMR14010A, LMR33610, LMR36510, LMR36520

Hi,

we have a complete line of transmitters (pressure, humidity, temperature, CO and CO2). Now we're implementing Wifi option, 350mA peak, in out product line.

As we would have a max current of 450mA (100mA peak from CO2 and 350mA from wifi), i'm testing some regulators for that situation. From most of the time this circuit will consume ~160mA.

Out first bet was a regulator from another brand, due to it's small footprint, BD9G101G, i burnt around 30 units, tried even a small heat sink, and 2 designs, and then gave up.

As we already worked with LM2574 in our previous design, we tried it. We were not doubting it would work, so we manage to fit in our new PCB, and with the finished project i tried make it work with 450mA, and it burned.

It heats a lot with 300mA, but it's been turned on for a while, and maybe the RMS value would be always bellow 300mA (need to make some more tests with the Wifi). But it's capacitors has a lot of heat too, so i'm concerned that it would drop a lot the life of the product.

Anyway, how does this regulators are tested for 500mA? With an infinite heat sink? I'm using the datasheet's design, for fixed 5V, 330uH.

Should i use a higher current reg. even knowing i have only some peaks near the maximum regulator working range? The heating is a problem for me beacause of the temperature sensor, can you indicate other regulator, with less heat, small footprint, low cost?

  • I was looking on LMR14010A for an alternative. It could work with up to 1A, so i would have a good margin.

    But about the heating problem, trying to see in terms of efficiency and thermal resistance.

    Webench says i'll have almost double efficiency with LMR1410A:

    LMR14010A    89,6%

    LM2574          79,7


    But i'll have almost double Junction-to-ambient thermal resistance on LMR14010A.

    LMR14010A    102°C/W

    LM2574          60,4°C/W

    So i'm not so confident with it.

    I measured the LM2574 temperature 2 hours before i turned on with 310mA, i could measure 118°C. So i think i need a lot more efficiency with 450mA.

  • Hi Rodrigo,

    Yes it seems that it need more efficient converter. 

    Another option is LMR33610. It is easy to use and has a thermal pad to dissipate the heat. 

    http://www.ti.com/lit/ds/symlink/lmr33610.pdf

    It is synchronous converter, therefore you do not need to have an external diode. 

    It will also use a smaller inductor since the operating switching frequency is higher. 

    We can take a look at the layout of your LM2574 to see how you layout the converter. 

    Thanks

    -Arief

  • Hi,

    i don't think that LMR33610 is a good option, it works with maximum 36V, i'll work at 37 maximum, so i need at least 40V input.

    In it's efficiency chart it says could get 85% with 500mA and 75% with 200mA @ 36V, so i still think that LMR14010A  could be a better option.

  • Arief,

    I saw that i could use a similar component, LMR36510.

    84,1% eff. from webench, not as high as LMR14010A , but it really has a very good thermal resistance RθJA = 42,9°C/W.

    Can i have a quick sample of this component from TI?

    Thanks

  • While i'm waiting my LMR36510 sample approval i'm designing the PCB.

    I know that it may be a controversy question, but i usually don't use GND plane bellow the inductor, even on the opposite layers of it. I saw that TI indicates on LMR36510's datasheet to use ground plane bellow the inductor. Can i use it without a problem? It's interesting for me in this case as i'm worried about power dissipation and i want to use GND polygon whenever i can. I should use solid polygon, right? What about exposed cooper near the LMR36510?

  • HI Rodrigo,

    Sorry for the delay, i have been caught up with internal request for the past few days. Let me know if you need any help ordering sample or EVM for this product. The EVM is for LMR36520 but it is just a higher current option (2A vs 1A)

    Yes the LMR36510 is a good candidate. It is just recently released also. I would be more comfortable with this part for your operating conditions. 

    I think the datasheet layout guideline will provide a good recommendation from both electrical and thermal perspective. http://www.ti.com/lit/ds/snvsbd7a/snvsbd7a.pdf#page=30

    In respect to GND underneath the inductor, my opinion is that if you put a ground plane underneath inductor it will reduce your EMI issue since you have a reference plane underneath the SW node which is the inductor. 

    The important part is the GND/DAP underneath the IC (exposed copper), this is where most of the thermal is being release from. Therefore i would put a solid polygon especially on the same layer of the IC. 

    My rule of thumb is to provide as much as ground pour on the external layers compared to internal layer because these are the layers that is exposed to ambient air and the heat can dissipate faster to ambient. 

    We can also review your layout when you are done with the design

    Thanks

    -Arief

  • Hi,

    samples tested.

    LMR36510  worked with 500mA+ without a problem, max temperature observed ~45°C. My inductor heated more than the regulator ~55°C.