• State Resolved
  • Locked Locked
  • Replies 9 replies
  • Answers 3 answers
  • Subscribers 64 subscribers
  • Views 282 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.

LM5146-Q1: LM5146-Q1 output voltage is not stable

Joseph Wang
Prodigy 20 points
Part Number: LM5146-Q1
Other Parts Discussed in Thread: LM5146,

Tool/software:

Hi, could anyone please help to review my DC/DC design with LM5146-Q1?  my A-sample has an issue is when their is no load, Vin=45V, Vout=27V, but when connect load Iout=10A then Vout=12V. so which parameters I need adjust or component i need to check to solve this problem make Vout stable? here is the calculation file.1123-4-LM5146-Q1-28V-02-TI.xlsm

Thank you for your support.

  • +1
    TI__Guru 67482 points

    Hi Joseph,

    Thanks for send the quickstart file.

    Are you running in diode emulation mode at no load? If so, check FPWM by pulling the SYNCIN pin high. I assume the issue here is when loaded (as Vout = 28V is entered in the quickstart file).

    Here are some notes on the design:

    1. The buck inductance is very low, giving a huge ripple current of 110% (should be 30-40%). Change from 1.8uH to 4.7uH.
    2. Check that Cout is actually 440uF (as ceramic capacitance derates significantly with voltage). This is important, as it affects compensation. I see the ESR is 1mΩ, so I assume it's not electrolytic.
    3. No need for a 150V low-side FET – 100V is adequate for 85V Vin-max spec.
    4. TO-220 and D2PAK FETs have high parasitic inductance and are generally older technology devices. Consider using conventional 100V 5 x 6mm package FETs (see the LM5146 EVM for reference, both in terms of FET choice and PCB layout placement). This gives more efficient switching, particularly when Fsw = 500kHz.

    If the issue persists, send on the schematic and layout for review.

    Regards,

    Tim

  • 0 in reply to Timothy Hegarty
    Prodigy 20 points

    Hi, Tim,

    Thanks a lot to give me direction to solve this issue, here are the information for your question, yes in my design SYNCIN pin is NC float, I will connect with VCC to see if could sove this issue, please get the schematic and layout(the layout I benchmarked LM5146 EVM but because of the board size the placement are not good):

    1: I will change L4

    2: there are 5 TDK(C2012X5R1V226MT000E) and 1 NCC(EMZR350ARA331MHA0G)

    3&4: thank you for the suggestion, I will change it as well,  I actually benchmarked the LM5146 EVM but didn't think over about MOSFET just selected from my library which looks like parameters satisfy the calculation

    Thanks

    Joseph

  • 0 in reply to Joseph Wang
    TI__Guru 67482 points

    Thanks, Joseph.

    The layout doesn't look too bad. I hope the input caps have a good connection back to the source of the low-side FET.

    The 330µF/35V electrolytic will dominate the output capacitance, as the 22µF/35V/X5R/0805 caps will derate significantly (possibly to only a few µF each).

    Note that the 330µF is an aluminum electrolytic, which complicates stability as its ESR will increase dramatically at cold. Better to use a 35V-rated polymer electrolytic with stable and lower ESR. If you keep the aluminum electrloytic, just update the ESR entry in the quickstart file as it will be much higher than 1mΩ. This definitely has an impact on stability, which is likely the issue here.

    Let me know if you see improvements with the changes. 

    Regards,

    Tim

  • 0 in reply to Timothy Hegarty
    Prodigy 20 points

    Hi, Tim:

         Thank you for the strong support, here was my test, I connected SYNCIN(PIN8) to VCC( also tried via 20Kohm resistor), when starting the testing without load, it was strange Vout=10V, disconnect then go back to 27V, with load the same as long as I pull up the SYNCIN then Vout=10V, any thought? 

    I will change the aluminum capacitor, MOSFET, inductence  according to your suggestion step by step to see if could solve this issue. please provide your valueable feedback to me.

    Thansk

    Joseph

  • 0 in reply to Joseph Wang
    TI__Guru 67482 points

    Just to confirm, Vout - 10V in FPWM but goes to 27V in diode emulation mode?

    This sounds like a stability issue, so let's get the buck inductor and output cap optimized first.

  • 0 in reply to Timothy Hegarty
    Prodigy 20 points

    Hi, Tim:

      Yes, as long as I pull up SYNCIN=FPWM then Vout=10V whatever with or without load.

    about output capacitors my calculation was 5*2mohm // 1*80m ohm=0.398m ohm, this was 1m ohm come from; you are right I did testing according to your last feedbck on ELYT CAP, after remove it from PCBA then compare remain Vout increase from 12V to 20V with 5A, without load Vout changed from 27V to 28.4V, so looks like here is the key issue, but here I am lack of theory to calculate it, please provide your professional suggestion to me, I am going to change the inductor and ELYT CAP according to your suggestion to see the result.

    Thanks

    Joseph

  • 0 in reply to Joseph Wang
    TI__Guru 67482 points

    Joseph,

    The ESRs don't appear directly in parallel, as it also depends on the capacitive impedance in series (hence the effective ESR value varies with frequency).

    This file may be useful:

    2072.Parallel Capacitors Zct LM5145.xls

  • 0 in reply to Timothy Hegarty
    Prodigy 20 points

    Hi, Tim:

       Thank you so much on the support, today I decided to review all your inputs and focus on output capacitors and inductor, after change to a 4.7uH inductor and relevant capacitors accordingly, Vout=28.5V without load and 26-27.5V with load, I should satisfy the requirement in calculation file at beginning.

    Thank you again for your professional guide.

    Joseph

  • 0 in reply to Joseph Wang
    TI__Guru 67482 points

    Hi Joseph,

    Thanks for the update. However, the output should be well regulated, < 1% load reg when operating in FPWM.

    Regards,

    Tim