• State Resolved
  • Locked Locked
  • Replies 12 replies
  • Subscribers 64 subscribers
  • Views 4140 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.

LMZ14202EXT: Regulation with light loads problem with light loads

Michael Denny
Prodigy 60 points
Part Number: LMZ14202EXT

I've got a design with the LMZ14202EXT with 22V to 36V input and 5V output right around 400mA.  As the datasheet shows, the regulation suffers with loads less than ~400mA and I'm on the edge.  However, I built eight engineering models and all of them passed with flying colors over extreme temperature.  No shift in DC regulation or increased DCM noise. 

Now, my production units are having problems at ambient temps.  It doesn't enter CCM without an additional 30mA load. The only difference between the production and EM is that this is a new LOT of parts. (procured later in time).  Is there anything in the fabrication of these parts that would change the critical conduction point?  What is the critical conduction point for this part?  If I'm right on the edge (as in my application); what do you recommend I try to improve it's performance. 

Thanks,

Mike

  • 0
    TI__Genius 13700 points

    Hello Michael,

    I will see what I dig up for an answer to your question.

    Outside of an acceptable manufacturing variation that is within the specification of the datasheet and production line validation testing there should be no difference.

    Can you past a copy of the schematic showing the input and output capacitors for the LMZ14202EXT? This would help the process.
    Also are you using a feed-forward capacitor in your set up?

  • 0 in reply to Ryan May
    Prodigy 60 points

    Sure.  I was just wondering if something on the newer lots has changed.  I should have clarified my question.  What tolerance do you expect for the DCM/CCM boundary?  I was always close, but I felt safe after testing so many units.

    Anyway, here is my schematic. 

  • 0 in reply to Michael Denny
    TI__Genius 13700 points
    Hello Michael,

    Thanks for the schematic.

    Looking at equation (11) in the datasheet and using 49.9kohm I ended up getting about 771Khz for the CCM switching frequency. Then using that value and the min and max input voltage range in equation (17) I would expect that the boundary be between about 250mA to 280mA.

    The primary question that I have is what purpose s CR12 serving? I am wondering if replacing that with a shunt that and configuring R196 and R197 to a simpler resistor divider ration will help to address the issue.
  • 0 in reply to Ryan May
    Prodigy 60 points
    Yes, I figured the diode would raise some interesting questions. For this application, I needed the diode to slope the supply's output over temp. This is because of my odd customer requirements for this product. That being said, I wouldn't expect the diode to cause regulation issues.
  • 0 in reply to Michael Denny
    TI__Genius 13700 points
    Hello Michael,

    Have you tried swapping the high-side feedback resistor and the diode and then placing the Cff capacitor across only the high-side feedback resistor?

    I am working on digging into other potential ideas but right now based on the design your DCM/CCM crosspoint should be much lower and the diode is an interesting variable in the feedback network.

    The other question I have is what type of capacitors are the 33uF output caps?

    I am talking this application over with one of my colleagues to see what their thoughts are as well.
  • 0 in reply to Ryan May
    Prodigy 60 points

    The caps are ceramic X7R types.

    CAP,CHP,CER,X7R,33UF,20%,16V,1812 SIZE

    I'll just try removing the diode by shorting it.

     

  • 0 in reply to Michael Denny
    TI__Genius 13700 points
    Hello Michael,

    Since the output capacitors are ceramic with low ESR and there are a few in parallel there may not be enough output ripple coming back through the feedback network even with a feedforward capacitor due to the low output ripple that results from such low ESR capacitors.

    Increasing the output ripple slightly should improve the response of the control loop. This can be accomplished by placing a resistor in series with the output capacitors to increase the output voltage ripple enough to more effectively trigger the the COT control loop at lower output currents.
  • 0 in reply to Michael Denny
    Prodigy 60 points

    Removing the diode didn't help.  We also tried reducing the output capacitance significantly (< 10x) and that didn't help either.  The thing that has worked is increasing RON resistance to about 61K ohms.

    It's not easy for me to add resistance in series. 

    Mike

  • 0 in reply to Michael Denny
    TI__Genius 13700 points

    Hello Mike,

    Is running the Module at a lower frequency with a higher Ron an acceptable solution for you?

  • 0 in reply to Ryan May
    Prodigy 60 points

    My latest is an error in the datasheet.  The block diagram on page 11 shows 6.8uH inductor, but the equations for design (equation #16 for example) uses 10uH.  Which inductor is correct?

    Also, could you provide the tolerance of that inductor? (any info on lot-to-lot variation).  Is this a chip inductor or a hand wound inductor?

    We can get stable regulation by changing RON either higher or lower in value, but our current 49.9K seems to be right on the boundary with the new LOT.  My management wants to know that this problem won't reoccur on the next lot, hence my questions above.

    Thanks,
    Mike

  • 0 in reply to Michael Denny
    TI__Genius 13700 points
    Hello Mike,

    Those are great questions. I am pulling in our product line to better answer the issue you raised about the discrepancy in the datasheet for the inductor value and to provide a better explanation of potential process variance than I can provide.
  • 0 in reply to Ryan May
    TI__Mastermind 19070 points

    Hello Michael, 

    I confirmed that the inductor value is 10uH. 

    The DCM/CCM boundary will depend on the inductor value (and tolerance), input voltage, switching frequency setting (and tolerance).

    The worst case for you would be at maximum input voltage (36V) and the lowest inductance value. 

    I put together a spreadsheet with some calculations for your input voltage range, including the tolerances - see attached. 

    LMZ14202EXT.xlsx

    Please double check the calculations, but I think if you target CCM operation at 350mA and consider the tolerances and max Vin, the Ron resistor value should be around 43k. 

    Let me know if this is helpful and if you have other questions. 

    Cheers, 
    Denislav