• State
  • Locked Locked
  • Replies 6 replies
  • Subscribers 64 subscribers
  • Views 1073 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.

LM3492: The chip is corrupted if use MCU DIM control

Иван
Intellectual 370 points
Part Number: LM3492

Hello! I make the scheme, what use in DRL. I use two serial string of LED (12 LED in string, 150mA for one LED). Input voltage 12-14 V. I make the schematic, shown below. I get the big problem: if I try programming controller, and then start program (the program run control DIM and slowly increase lighting with PWM), sometimes LM3492 breaking. The IC don't work, use very big current. The resistece between VOUT and GND is 5 Ohm. Only sometimes It's happened. If I remove R19 and R20, I don't get this problem. Why, if programmer connect I get this problem. I lost near 5 IC for this problem.

  • 0
    TI__Mastermind 25500 points

    Hello,

    I am not sure at the moment. I am asking somebody that know this device better than I do. But if he doesn't have an idea we can investigate further. Often with switchers it's a layout issue but I can't be sure at this time. Let's see if I get a response whether this had been seen before. In the mean time I would be happy to check the BOM and layout if you could provide them.

    Thanks,

    Clint

  • 0 in reply to Clinton Jensen
    Intellectual 370 points

    Hello! In this file shown PCB and BOM. And can you tell, what temperatur work for this IC on it evalution board LM3492EVAL? I saw that they don't have thermal radiator. My device are very hot, and I use external radiator. But if I use two channel with 200 mA (for test), the IC is overheating and go to thermal shutdown.

    SL_LSeries_test.pdf

  • 0 in reply to Иван
    TI__Mastermind 25500 points

    Hello,

    I am not sure about the EVM thermal characteristics, but your board is quite small so there isn't much heatsinking. It also looks like you are using a pretty high switching frequency. It is possible with a high switching frequency and high output voltage the AC losses are too high. They are directly proportional to switching frequency and Vout^2. It may be in your particular application you need to use a lower switching frequency for thermal stability.

    I also wanted to ask are your inductor values really only 10nH? That seems very low and would result in very high peak currents, you may even be current limiting if so. Lower values also increase AC losses.

    I would try increasing L (if needed) and reducing the switching frequency to around 350kHz to keep it much cooler. That frequency is also a good one to pick if you need to pass any EMI requirements.

    Regards,

    Clint

  • 0 in reply to Clinton Jensen
    Intellectual 370 points

    Hello! This frequency was recomended by WEBENCH designer. As inductor I use CDR127 47 uH. Now I make new PCB, what are larger and have all components on one side, and bottom side are clean, and can place external radiator. Also I will try change work frequency.

    With new PCB I am also test my project for error, what damaging LM3492. Maybe it is realy breaked because of bad layout.

  • 0 in reply to Иван
    TI__Mastermind 25500 points

    Bad layout is a common issue and can absolutely cause damage due to voltage spikes exceeding abs max ratings. An improperly soldered DAP to ground can also cause damage. There are layout guidelines in the datasheet as well as multiple app notes regarding switching power supply layout on ti.com. If you are using 2 layers with all the components on top I would recommend using the bottom layer as a solid ground plane and tying all power ground points directly to it through multiple vias. A ground plane helps immensely, that includes heatsinking if you make sure you have vias at the IC DAP. I hope this helps.

    Regards,

    Clint

  • 0 in reply to Иван
    TI__Mastermind 25500 points
    BTW, Webench is a great starting point for sure. But it will not always take into account thermals if you are trying to dial in the smallest solution possible (higher switching frequency).