• State
  • Locked Locked
  • Replies 7 replies
  • Subscribers 64 subscribers
  • Views 950 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.

LM3478 audible noise and Imin

bryan simmons1
Prodigy 40 points
Other Parts Discussed in Thread: LM3478EVAL, TL5001, LM3481, LM3478
  • I have prototyped a solution using the LM3478EVAL board. 
  • The device produces audible noise in the 5-20kHz range due to OVP under light loads.
  • I am skeptical that reducing L and utilizing DCM can reduce the output current to below my minimum current requirements which are likely to be in the 100mA range due to the minimum D=0.325 (325ns blanking / 1000ns PWM period.)and provide adequate throughput to produce the 5-10A of output current required at high loads. 
  • Is there another solution in the TI controller family that can accommodate this range of output current requirements without producing audible noise?

Thank you in advance for your help.

Bryan

  • 0
    TI__Intellectual 770 points

    Hi Bryan,

    For a similar solution, please can you send some more information about your requirements, such as input voltage range, output voltage and peak load current ?

    Regards,

    Doyel

     

  • 0 in reply to Doyel Ghosh
    Prodigy 40 points

    Doyel,


    Thank you for the reply:

    input 6-9V

    output 12V 7Amax 10mA minimum

    at loads less than say 1A the OVP kicks in on the part and cycles on/off at audible frequencies.

    Thanks

  • 0 in reply to bryan simmons1
    TI__Guru 63205 points

    Bryan,

    Is this the only known condition which the audible noise is present?

    In other words, have you found other operating conditions that also create the audible noise?

  • 0 in reply to BrandonAzbell
    Prodigy 40 points

    No other conditions.  The exacerbating condition is when the feedback voltage sags to below Vfb + Vovp - Vop(hys).  The OVP comparator lets up on the PWM logic block and the device starts feeding the output..  Because the minimum output PWM duty cycle is limited by the 325ns blanking period, the voltage quickly shoots back up to the OVP voltage and shuts the output down.  Rinse and repeat.

    You can re-create this with an out-of-the-box LM3478EVAL board at say 100mA and 6V input.


    It happens that at low loads <1A this above cycle is repeated every 50-200 microseconds producing 5k-20k audible noise.  I can't isolate the source of the noise, but I suspect it's the inductor.

    I can play games with the size of the output capacitance, etc. but it just pushes the window of offending load levels around.

    On an up note, you can vary the pitch of the noise by changing the input voltage.  With a little practice I could play a tune.  My labmates liken it to the sound of a mosquito, though.  None of this would matter, but our customers tend to be young and selective and have excellent high-frequency hearing.


    I'm starting to think that an old-fashioned voltage mode controller like the TL5001 is the answer, but I don't have the time to jack with compensation circuits, right now.

    Thank you for your help!

  • 0 in reply to bryan simmons1
    TI__Intellectual 770 points

    Hi Bryan,

    LM3481 can be a possible solution to your requirements. 

    Regards,

    Doyel

  • 0 in reply to Doyel Ghosh
    TI__Intellectual 2225 points

    Hi Bryan,

    Just to reiterate, we did test the LM3481 which has very similar functionality to the LM3478 and it appears to not have this issue with audible noise at light load.  We encourage you to try that part and see if that meets your needs.

    Perry

  • 0 in reply to Perry Tsao96839
    Prodigy 40 points

    Perry,


    OK, thank you.  I will evaluate the 3481, also.  From a review of the specs, I don't see why its fundamental operating mode would be any different, but the proof would be in the pudding.

    Thanks,

    Bryan