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.

LM3409 start up malfunction

Other Parts Discussed in Thread: LM3409, LM3409EVAL, LM3409HV


I have purchased LM3409 Demonstration Board (LM3409EVAL/NOPB, SV600307-001).
I use board for testing purposes for our future product.

Our product will be LED controller for 27V and 1,6A LED stack. At one PCB it will be 40 LED drivers.
We have deal done for 500 pcs of this controllers that mean we will buy 20k LM3409 drivers.

I have big problem with burning LM3409 at power up.
There were some components replaced and I write this component to .pdf attached.

Generally components was replaced for better efficiency and cost optimization.

Demonstration board had switching frequency of 700kHz which I have decreased to 530 kHz, with replaced R1 (15,4k) to 22k.
I also use different MOSFET, Diode, shunt and inductor.
Problem is next:
-> Power supply is 36V. Vout is 27V, Iout 1,6A nominal (Imax 1,8A in final functionality)
-> When I get Demonstration board I have run it and measure switching frequency and efficiency.
-> Then I have started with replacing components and also measuring efficiency.
-> Once at start up LM3409 burned out. I did not know why. There were no short circuit because few times before it work properly.
    Result was that MOSFET and LM3409 was burned out. I have replaced and it worked for a while.
-> After few days I was making new measurements with new components (MOSFET and diode). Demonstration board
    worked for few times then again LM3409 burned out. Then I have replaced MOSFET and LM3409 and it burned out 3 times in line.
    After that I have replaced: shunt, inductor, diode, MOSFET and LM3409 and then it worked.LM3409_Demonstration_Board.pdf
-> Yesterday I was going to measure final measurements with selected MOSFET. At demonstration board I had original MOSFET ZETEX with
    custom diode, shunt, cout and inductor, because I wanted to measure efficiency with this set up.
     Power supply voltage was set to 36V, supply current limit to 2A, load of 21R. Device was in disabled mode
    with jumper places to corresponding position. AND when I have connected to power supply LM3409 burned out! Again. I have replaced LM3409 again  and it worked. Then I have replaced MOSFET with DMP6050 and it worked. I have measure efficiency. 

-> Today I was going to lower switching frequency. I connected power supply and it worked with load of 15. Then I have added 0R6 resistor in series with 15R and when I connected to power supply LM3409 burned again. FET is OK. 

Till now I have burned 10 LM3409 all at start up. Obviously there is something wrong at start up.

I have removed R5 potentiometer. Instead of J1 jumper I have soldered resistor divider of 150k and 12k so at EN pin is 2,66V (@36Vin) and 1u capacitor directly to pin 3 and GND.

Can you give me answer what is wrong?
Problem is because we have PCB in design and we will make 20000 LED drivers and now I have problem that there is very big chance that LM3409 will burn out at start.
I replaced original LM3409 with LM3409HV (see ordering code in .pdf).

Thanks for your answer in front.

Best regards.

  • Is there any way you could provide information on the components you changed, such as the part number for the inductor, FET, diode, etc...?

    Also, without that information I can't be sure what could have happened initially. I'm also not sure what kind of soldering setup you have. But if the DAP of the device is not soldered solidly down you could cause damage.

  • Hello,

    Firstly thanks for answer.

    Ordering codes of all components was written as comments in .pdf file attached in mail.
    Here are ordering codes:
    -> LM3409HV
    -> C1 and C5: 1206 4u7 50V, Taiyo Yuden: UMK316AB7475KL-T
    -> Q1: Diodes: DMP6050SSD
    -> D1: Fairchild: FSV360FP
    -> L1: BOURNS: SRN1060-220M
    -> R4: 0R250 || 2 pcs => R=0R125; Stackpole Electronics Inc.: CSR1206FKR250

    Regarding your answer: IF DAP is not soldered solidly it cut cause damage.
    -> I'm not 100% sure that DAP was soldered correctly. I soldered it with hot air, but after removing DAP, DAP and pad looked it was soldered.
    -> Question is: If DAP is not soldered correctly it would damage also if EN tided to GND - device disabled?
    -> Demo board was operating at least 1 hour. LM3409 temperature was measured with thermal camera and it was 45°C, MOSFET temperature was 62°C. Then next day when first power supply was attached LM3409 burned.

    Thanks again for your answer.
    Best regards.
  • The components you are using seem like they should be ok. But I may have an idea. Some comments:

    1. You should check the thermals on the diode as well. I know it's rated for enough current, but it's a small package. If it heats up too much it will avalanche and destroy the FET most likely, which can also take out the LM3409 depending on what the gate does upon destruction or what transients that causes on CSP/CSN. This likely isn't the problem, but you should check to be sure.

    2. The DAP could cause issues, but not with EN low. But if you solder it down and it's marginally connected it could easily start up, work for a while, and expand so contact is good. But when you power down and let it cool the contractions could cause a weak connection to break so you would have issue on the next startup.

    3. I noticed you didn't mention changing the UVLO resistor values. This should be done as the UVLO is currently set at about 10V. You should have it around your lowest expected input voltage. In this case maybe try R2=4.02k and R3=100k to give you a startup voltage of around 32V. Basically you want to have it above the LED string voltage. Otherwise it will turn on at 10V and then it goes into 100% duty cycle mode, basically a short to the output, until the input voltage is high enough that it starts switching and regulating. In cases where the input voltage and output voltage are not extremely high this generally will just cause a blink at turn on. In cases where there is no output cap it will blink but there are no large current surges. In your particular case you have an output cap and it is turning on at 10V. So the output cap is charging with the slew of the input voltage. Depending on the rise time this could cause very high current in the FET and you could exceed its SOA. You may even saturate the inductor making things even worse. Adjusting your UVLO level will prevent that. This is the most likely cause of damage.

    So I would recommend upping the UVLO and making sure the DAP is down solid and you shouldn't have issues. But you can let me know if you still do.

  • Hello,

    I'll try and I let you know the results.

    Best regards.
  • Hello,

    1. Diode was OK.
    2. DAP was OK.
    3. UVLO was a problem. I use 120k and 4k7 resistors, which give Lock in voltage of 32,4 V and Lock out 30,3 V. With 4k7 resistor I have place 10 n capacitor. Now is working.

    Thanks for all your help!

    Best regards.