• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Subscribers 63 subscribers
  • Views 399 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.

LM3410: Flicker with LM3410

Basso
Prodigy 40 points
Part Number: LM3410

We have a product which has been on the market for a few years using the LM3410.  We're finding that when the brightness is turned to max (PWM=100%), some customers are reporting flicker.

I think it may be due to the ripple on the power supply caused by the heavy drive, but have't been able to solve it with the addition of extra input capacitance (going from 4.7uF to 22uF) or extra output capacitance.  We've also tried adding capacitance across the FB pin, to no avail.  The next logical place to look seems like the inductor value. running the calculations, it appears that the minimum inductor value is 6.5uH, and we are using 15uH.  We are pondering increasing that to 22uH.  

Perhaps you could you suggest an approach which might help the flicker.  Would a larger inductor help?  And would it have any significant impact on our efficiency which is running around 85% (measured) right now?

The schematic is attached below.  We recently added the over-current protection circuitry, although it had no impact on the flicker either way.

  • 0
    TI__Mastermind 28655 points

    Hello Basso,

    Have you been able to reproduce this?  What power supply is driving this?  Is it possible the input is dropping out?  At low input voltage the source needs to be capable of over 1A.

    What does the output voltage look like at full current?  Is it at 9V?  Has this problem always existed?

    Changing the inductor probably won't change this behavior.

    Best Regards,

  • 0 in reply to Irwin Nederbragt
    Prodigy 40 points

    Thanks for the help.

    Yes, we have occasionally been able to reproduce the flicker (even once during FCC testing), but it seems to come and go. Some units are rock-solid while others are more prone to it. (They are fabricated in the 1000s with pick-and place, so it is more likely to be some slight component variation.)

    We are driving this circuit with a 5V 2A USB supply or a 8000mAh Li-Ion battery, depending on whether it's plugged in or not.  The issue tends to occur more when plugged in, but we've seen it both ways.

    The output voltage is a solid 9V at full current. We are seeing a great deal of power supply noise (>1.5Vp-p on the 5V supply) on the Vss bus when full on.  The power supply is handling the current requirements without dropout. The battery provides better smoothing, since it's a very big capacitor anyway.

    We've tried C4 at 22uF and C7/C8 have doubled to 10uF each as well. (C7 and C8 are on opposite sides of the PCB to reduce microphonics.  We are basically dividing a 10uF in half and having them cancel each other out.  Doing so helps the audio noise quite a bit.)

    I read another post on this forum which suggested adding a 10nF or 100nF capacitor on the FB resistor.  We've tried a 100nF (0.1uF), but nothing smaller.

    Not quite sure where to turn next.

    Thank you.

  • 0 in reply to Basso
    TI__Mastermind 28655 points

    Hello Basso,

    As you know it will be very hard to figure out what's going on unless you can reproduce it and see what it is doing.

    I'm not sure what you've tried but you could try reducing the current sense resistor(s) to drive more current to see if it malfunctions more frequently.  Or if some other operating condition such as higher temperature makes if more frequent.

    As for the noise on the input, it it noise, current ripple or?  I'm not sure how you are measuring this.  If you are using a scope probe with it's ground try looking at the ground connection (probe where the ground is connected) to see if it is real or being picked up in the oscilloscope loop.

    Best Regards,

  • 0 in reply to Irwin Nederbragt
    Prodigy 40 points

    Thanks, Irwin.  I agree that it's difficult to isolate without reproducing it reliably.

    It may be that our circuit is getting into a state where the PWM varies due to effects on Vss by the LM3410, causing a feedback loop.  In other words, Vss noise causes our microprocessor to adjust the PWM to less than 100% (full on), which causes Vss noise.  (Ground is measuring dead-flat zero, BTW, so it isn't a scope loop.)  

    I've taken a scope trace of the 2 kHz PWM being sent from our microprocessor to the LM3410 at roughly 66% duty cycle, and the subsequent impact on Vss.  The yellow trace is the PWM signal. The blue trace is Vss.  Note that scales for the traces are different (Vss is shown at 1V per division to illustrate the impact better).  Vss is varying about 1Vp-p.  In this case, the "noise" is effectively ringing and power supply droop from what I can tell.

    Thanks for the help.

  • 0 in reply to Basso
    TI__Mastermind 28655 points

    Hello Basso,

    This is caused by impedance from the power source to the input of the LED driver.  If the output is 9V at 300 mA the average on-time input current is above 600 mA for 4.4V input (during the LED on-time).  It appears to be dropping about 0.6V which makes the input impedance around 1 ohm, this is high.

    To hold this up with capacitance it will require 1800 uF of capacitance to keep the dv to around 0.1V.  I would look at reducing this impedance since the dv will go up as the input voltage goes down.  You can try adding a large electrolytic capacitor to the input to see if it reduces this.  You have to keep the ceramic for high  frequency bypass.

    How is the power supply connected to the LED driver?

    Best Regards,

  • 0 in reply to Irwin Nederbragt
    Prodigy 40 points

    Irwin:

    Thanks, that's likely where we need to look.  We have a power supply MOSFET between the power supply and the driver, and although it shouldn't increase the impedance much, it's not helping. The extra long cord to our power supply is probably the major contributor.

    We're going to see if we can put the battery permanently inline with the driver so it acts as a large (although not wagon-filling) capacitor when wall-powered.  

    At this point, I think the problem lies outside of the LM3410.  I appreciate you walking through this with me.  It's a great help to have you available for questions!

    Best regards,

    Basso.