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.

Problem with my design of LM3492HC

Other Parts Discussed in Thread: LM3492HC

Hello,

We've made a design for a LED driver for 2 x 10 Leds with the LM3492HC, based on the reference design in AN-2231.

Unfortunately when we connect Vin supply of 12V, we get not output voltage at all at the VCC (from the 5.5V LDO).

The EN pin is currently not connected at all. The LED load is also disconnected until the fundamental problem of the LDO will be solved... 

Looking at the Block diagram in datasheet it seems that the LDO VCC output shouldn't be affected at all by any input pin?

We've spent hours and couldn't understand what's wrong. 

 Thank you,

Ofir

  • I will check with someone who knows this device a little better than I do, but assuming no cold solder joints and the proper signals VCC should be present. Leaving EN open is fine, but are you making sure you apply >2V (but less than 6V) to DIM1 and DIM2 as well? If those are low the device will be disabled and VCC will not come up.
  • Hi Clinton,

    The DIM1/DIM2 are connected to a 5V pwm microcontroller driven outputs. We'll test it again but I think it made no difference when we tried different pwm values.

    Can you please refer me to the paragraph in the datasheet which indicates that DIM1/DIM2 are disabling the device? because from the block diagram it seems that the 5.5 LDO is directly driving the Vcc?

    Thanks,

    Ofir

  • I was taking a guess at a possible cause due to the EVM user's guide specifying applying a voltage to DIMx, but I am not 100% sure and I could not find any info regarding this in the datasheet either. I am trying to find out if there are any conditions other than EN low that would disable VCC because I know not everything that affects it is shown in the block diagram. I will let you know what I can find out. In the mean time is there any way you could check VCC with an oscilloscope when the input voltage is applied? I'm curious to know whether it comes up briefly and then is shut off by some event or if it just never comes up at all. Is there current draw from the input supply even though VCC is 0V? Have you checked for cold solder joints or solder bridges between pins? I am waiting to hear back from a couple people since I am not extremely familiar with this device and will let you know what I can find out.
  • Hello. A question from a co-worker: Is R40 populated, and if so what is the value? Thanks.

  • The R40 is not populated, as well as R41.

    We we look with oscilloscope at the Vcc pin during power up, there's no transient signal visible. It's just shows 0[V].

    I'm not sure if it's related but during power up the Vin signal is rising from 0 to 12[V] quite slowly (about 10 [ms]) due to capacitors in the PCB design.

    I'm not sure what is the current draw of the LM3492HC alone, because it's already assembled on a PCB with our mircocontroller(ATMega2560) and regulators. The total current of the PCB is now about 32[mA].

    We've tried to put EN to ground and back to 5[V] with the microprocessor, but there's no effect on the Vcc signal, and when we put EN high to 5[V] the extra current is about 6 [mA]. 

    Thanks, Ofir

    P.S

    If the output load (LEDs) is not connected to the PCB, does it draws current? 

  • Have you tried swapping out the part or checking for a short from VCC to ground? If VCC does not come up at all there must be some sort of manufacturing problem somewhere or the device must be bad for some reason.
  • Tried replacing the part, got the same result. no short circuit from VCC to ground, I get about 100KOhm resistance.
  • Another question: If the DIM1,DIM2,EN are connected to the microcontroller and are in undefined states during power-up or other error, can it harm the device?

    Thanks
  • It would likely be possible if they are in an unknown state. None of those inputs have hysteresis so if you are near a threshold it could bounce from one state to the other very fast and cause all sorts of issue. It is always recommended that if the input could be in an unknown state that you add a pulldown resistor to each so that they stay low unless you directly force them high.
  • OK, and what about my question If the output loads (LED strings) are not connected between the Vout-Iout pins? does it draws current? or can it harm the device?
    I don't want to destroy another part before making sure :)
  • No, having an open string will not damage the part. You should not have to worry about that. It will only report a fault condition if you do a read/write.

  • Hi Clinton,

    For testing I've got a PCB populated only with the LM3492HC and it's peripherals. Red and Black wires are Ground and 12V from supply connected to Vin.

    The current consumption is about 5[mA]. The DIM1/DIM2/Enable pins are not connected externally. Vcc and EN reading is 0[V]. We added large capacitor on the output but it has no effect and no load is connected.

  • New finding:
    We tested a new PCB without capacitor C36 (Cfb), it seemed to work fine with Vcc at 5[V] and Vout switching. EN/DIM1/DIM2 pins were pulled-up to the Vcc output. We used current limit at the supply of 50[mA].
    After we added the capacitor C36 (10pF) and powered up the board, the current limit immediately saturated and the Vcc output afterwards went again to 0[V].
    Testing the resistance between Vcc and GND now is again faulty at few Ohms, meaning the part is probably destroyed like the ones before.

    Please Advise,
    Thanks, Ofir
  • Did this damage occur with an output capacitor? I just noticed on your schematic that you do not have one. A boost regulator cannot run without an output capacitor and the type is important. Without an output capacitor, and with C36 installed, you may be pulling FB above its abs max rating which would damage the device. Alternatively if you use a high ESR output capacitor you could pull it up as well. Regardless a boost regulator requires an output capacitor. I would recommend around 10uF and use ceramics for low ESR.

  • It happened with the output capacitor (33uF 200V), again with no LED string.
  • It sounds like it could be an overshoot issue then due to L6 and the capacitor, so I guess in certain conditions you could cause damage without an LED string to absorb the extra energy. Have you tried it with the LED strings connected? That would be the next logical step.
  • Didn't try it yet. So far we have easily destroyed 7 ICs... Each one of them measures now just few Ohms between Vcc and GND - which supposed to be about 120KOhm.
    I'm reconsidering the use of this IC in our design if it is so sensitive. In our design the device can be powered up without the LED strings connected.
  • I can understand that. It seems that it is a FB pin abs max violation since it doesn't happen when CFB is removed. CFB is not absolutely required so you could go without it if you would like. Another option is to zener clamp the FB pin so that it does not exceed abs max. If you decide to use another IC just keep in mind that starting up with an open output will cause some transients, so be careful of low voltage pins that could be AC coupled to the output or you may have similar issues. Good luck.
  • Thanks a lot Clinton for the support.
    If you can check about other designs using the LM3492HC that might have experienced the same sensitivity to the 5.5 LDO I'll be glad to hear.