• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Subscribers 62 subscribers
  • Views 1504 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.

Instability on output of LM3150 regulator

Marshall Bussen
Intellectual 255 points
Other Parts Discussed in Thread: LM3150, LM25116

I am seeing a lot of instability on the output of the LM3150 regulator circuit on some prototype boards.  I was hoping to get some help from TI in trying to debug this because I have tried a number of different things to try and stabilize the output of this regulator to no avail so far.

Our circuit is shown in the attached image LM3150 circuit.png.  It is more less copied from the output of the TI Webench tool, which I have also attached above.  I’ve also attached an oscilloscope capture of the output of the voltage regulator as measured across C363 (TEK00054.PNG).  In the scope capture, the voltage is being offset by 1.038V (our target nominal output voltage) so that an “ideal” output voltage would look like a DC level at the midpoint of the screen.  You’ll notice that there is some DC offset which I am not as worried about right now, due to the FB voltage being located in the ball grid array of the IC that this regulator is powering.  However, the 50mV peak-to-peak sawtooth that is showing up on the output of the regulator is what I am trying to address and improve upon.

I have tried using additional output capacitance to replace C363.  Our output capacitor is a Nichicon RL80E102MDN1KX which I attempted to replace with a Nichicon RNU0E272MDN1PH 2.7mF capacitor, because my own hand calculations led to a minimum output capacitance of just over 1mF.  This additional capacitance actually just made the sawtooth amplitude even higher, which is a confusing result to me.  I also attempted to replace C373 (the feed forward capacitor CFF) with an 82pF capacitor because my own hand calculations showed that this cap should be closer to 78pF rather than the 1500pF that the Webench tool specified.  This didn’t appear to make a marked difference.  Finally, I tried replacing the 1uH inductor that we were originally using (Bourns SRP1235-1R0M as spec’d by Webbench) with a 0.68uH inductor (Bourns SRP1235-R68M) which was closer to the inductor that we should be using as indicated by the Inductor Nomograph in Figure 15 of the datasheet and my hand-calculated ET of 3.619 V*us.  This also did not improve the output voltage stability

2402.webench_1.038V_SAS.pdf

  • 0
    Intellectual 255 points

    Also, here is an image of the layout for your review.  Note that I made one error in the layout here which was to place the 4.7uF caps near the VIN pin of the LM3150 rather than the high-side FET Q14.  I have manually added 2x 4.7uF ceramic caps near C375 but it didn't appear to make a difference for my output stability.

  • 0 in reply to Marshall Bussen
    TI__Genius 15810 points
    We will forward your query to the right engineer.
    Regards,
    Akshay
  • 0 in reply to Marshall Bussen
    TI__Intellectual 2225 points

    Hi Marshall,

    The LM3150 controls it switching time by looking at the output ripple voltage.  So the 50mV of ripple you see is considered normal.

    To reduce the output ripple further, I would suggest you try adding a ripple injection circuit.  In webench, after you open the LM3150 design, there is a "output ripple option" menu where you can select a "Low O/P Ripple Solution."  This will add small circuit in parallel with the inductor that injects ripple into the feedback circuit. 

    Please try taking a look at this suggestion and see if it helps you. 

    Perry

  • 0 in reply to Perry Tsao96839
    Intellectual 255 points

    Thank you Perry.  I will look into the "Low O/P Ripple Solution" to see if this can help me minimize the output ripple.

    If this solution does not work or is not sufficient for our requirements, can you point me at another switcher that I could use instead?  Assuming that we need a max +/- 20 mV output ripple at a maximum 12.3A load.

  • 0 in reply to Marshall Bussen
    TI__Intellectual 2225 points

    Marshall,

    I would consider using the LM25116.  It has similar specs, but operates using constant current mode control with a fixed frequency instead of the constant-on time architecture of the LM3150. That makes minimizing ripple a little more straightforward.

    Perry