• State Resolved
  • Locked Locked
  • Replies 4 replies
  • Answers 1 answer
  • Subscribers 62 subscribers
  • Views 1358 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.

LM3421: LM3421 Boost Converter oscillates when it is switched off and the high level of the output current falls down when difference between Vin and Vo becomes higher at dimming

Aaron_Rangel
Intellectual 440 points
Part Number: LM3421


Hello,

I have designed a Boost converter with LM3421 according the design guideline provided in the datasheet and it seems to operate well but I get an oscillation when the converter is switched off and I see a strange waveform when dimming. Io  is set up to 500mA and the output regulation is quite well adjusted (less or equal 1% at the most).

I attach you the schematic if that can help you to identify better what is going on. (D2 is a Schottky Rectifier with very low forward voltage drop and Main NMOS connected to the gate pin 11, is a Nexfet type).


When the converter switches off, I see an oscillation in the oscilloscope as you can check in the following capture. That means that when I switch off, I see as if the converter would switch on and off for a short period of time and hence that becomes a bit visible when I look at the leds connected to the output. What is happening here? How could I avoid this?

If I zoom the capture, maybe it can be seen more clearly. (Ch2= Io).

I also see some strange captures for output current when I am dimming whatever frequency I use of those that are expected to be used as per my design ( 5KHz, 20KHz). In fact I can see that better, when the difference between the input and the output becomes higher (i.e.: Vinput = 12V and Voutput = 33V). I attach you two captures, the first one is for 12V input, 33V output (11leds) and 20KHz dimming frequency and the second one, is for 18V input, 33V output (11leds) and 20KHz dimming frequency. (Ch1 = Vsw, Ch2 = Io).

When the output current is in "high level", so to speak, it falls down and swings and finally a peak is produced. I am a bit worried about this unexpected peak.

Even though the circuit dimms well and the second capture seems to be normal (I am not completely sure, but I think so), why do I have these oscillations (Vsw) when the transistor is off?. I think that the current peak I get at this moment has relation with this ringing but I do not know what an why I am getting this. When I use the circuit in fixed output current the Vsw waveform is very clean and free of ringing.

Could you please help me with these issues?

Thanking you all in advance.

Regards.

  • 0
    TI__Mastermind 25500 points

    Hello Aaron,

    I am not sure about the oscillation when you turn off the part. How are you turning it off? Are you just pulling nDIM low and seeing this? If so have you checked the nDIM pin for oscillations? If not does it happen if you remove the snubber across the main switching FET?

    For the dimming I have seen that kind of peaking before. If you are concerned about the peak at turn-off it is just likely that the disconnect FET is turning off too quickly. You can help that by adding some gate resistance. Then you can add a schottky in parallel with it to ensure a fast turn-on of the FET still.

    The peaks at the rising edge are due to the control loop. The 560k you have from COMP to ground will help, but it may need to be a lower value for lower input voltages. But what helps the most (along with a high value resistor) is a lower value COMP capacitor to increase the error amp bandwidth. That in conjunction with more output capacitance (to keep the voltage higher and also reduce closed loop BW to avoid the RHP zero) will usually do the trick.

    Regards,

    Clint

  • 0 in reply to Clinton Jensen
    Intellectual 440 points

    Hello Clint,

    I want to clarify that I am always disconnecting the converter through the main power supply of the lab.
    But, anyway, I have recently discovered what made this kind of oscillations when I switched off the converter. I switched off pressing the ON/OFF button of the main power supply and it seems this button produces a bounce when pressing it. I can confirm you this because I changed and tested with some other power supplies and it did not happened anymore. Therefore, it was due to a problem of this power supply.

    Of course, it did not have relation with the snubber yet I tried to check that by removing it and even checking nDim pin. It did not have relation with this, fortunately.

    On the other hand, I proved what you said about output capacitor and Ccmp capacitor. That is a rule of thumb for converters and it is a good tip that really functions well, you were right!. The problem here is to adjust the values but I got it without much attempts. I am saying "the problem", because according to my design I got a Ccmp = 1.56uF for an inductor of 57uF and I chose 1uF and 47uF as you can check in the schematics. And finally I have choosen Ccmp = 100nF and a resistance in parallel of 750KOhm, which really works even more fine.

    Just one issue I could not solve yet, I am sorry but I cannot reduce the peak at turn-off. I have tried it by adding a RD like in the main transistor, testing it with different resistances that goes from 10R to 4k7! which I thing is too much for charging the gate. Please do correct me if I am wrong. I could only check with this, indeed, that the turn on was delayed more when the resistance was increased but the turn-off time is produced always at the same time.

    So, at the end, the peak is still remaining here. Any idea for this? Why is still happening if it is not due to the nDim turning-off time?

    Thank you very much for your tips one more time, that help me to understand better how LM3421 works.

    Regards

  • 0 in reply to Aaron_Rangel
    TI__Mastermind 25500 points

    Hi Aaron,

    Are you connecting the RD in the same way as you are the main FET? If so you will be slowing the turn on and not the turn off speed. For the dimming FET you need to reverse the diode so that you slow the turn off but keep the turn on fast. Let me know.

    Thanks,

    Clint

  • 0 in reply to Clinton Jensen
    Intellectual 440 points

    Hello Clinton,

    Yes, I connected RD in the same way as in the main FET.

    After connecting diode in reverse mode and adjust the resistance the peak disappeared and the dimming works even more fine.

    Thank you very much, I appreciate so much your attention.

    Regards,

    Aarón