• State TI Thinks Resolved
  • Locked Locked
  • Replies 16 replies
  • Answers 7 answers
  • Subscribers 64 subscribers
  • Views 515 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.

Original question:

LM5117: Unable to output much current. Device keeps turning off.

LM5117: Vout drops to 0V when connects to load

Steven Wang
Prodigy 230 points
Part Number: LM5117
Other Parts Discussed in Thread: LM5148,

Hi, dear TI team

I’m using a digital potential meter as the feedback resistor to control the Vout of LM5117 buck. Vin = 48V. f_sw = 185 kHz.

  • When there is no load, Vout=32V is correct. C38 (capacitor at SS) is very hot when no load, about 80 degrees.
  • When connect to a load (48Ohm, should draw 0.6A). Vout is immediately drops to 0V, and no current output. C38 (capacitor at SS) is less hot with load (but no current), about 50 degrees.

Webench simulation circuit

Below are the measurements: 

Vout, RT, SS, FB, RES, CM. VCC

RAMP

COMP

HB

HO - SW

LO - SW

  • 0
    TI__Guru 59611 points

    Hello Steven

    I think there is a mistake on your layout, or your measurements are wrong. 

    V(HB-GND) should be zero voltage when V(VCC-GND) is zero, but it is not when the load is applied. 

    -EL 

  • 0 in reply to EricLee
    Prodigy 230 points

    Sorry, V_HB is 7V as expected. 

  • 0 in reply to Steven Wang
    TI__Guru 59611 points

    Hello Steven

    V(HB-GND) should be 0V when V(VCC-GND) is 0V. In your measurement, V(HB-GND) is ~7V while V(VCC-GND) is 0V. 

    -EL

  • 0 in reply to EricLee
    Prodigy 230 points

    Hello Eric, 

    There were some issues during the oscilloscope measurement. I have re-uploaded the measurement result.

    Thank you for the support

  • 0 in reply to Steven Wang
    TI__Guru 59611 points

    Hello Steven 

    Can you confirm if this is true ? 

    • DEMB pin is floating.
    • DAP is floating 
    • AGND node is not connected to PGND node. 
    • Sense resistor is connected to AGND 
    • No output capacitors 
    • CM capacitor is connected to PGND 
    • No input capacitors 

    Also, please check again without U7. 

    -EL

  • 0 in reply to EricLee
    Prodigy 230 points

    Hi Eric,

    • DEMB pin is floating.

    Yes

    • DAP is floating 

    Yes, LM5117 chip is at room temperature. 

    • AGND node is not connected to PGND node. 

    AGND connects to PGND via the net label (basically a short small wire).

    • Sense resistor is connected to AGND 

    Yes

    • No output capacitors 

    There are multiple output capacitors (10uF, 4.7uF, 1uF, 100nF * 2)

    • CM capacitor is connected to PGND 

    CM pin connects to STM32 ADC pin. 

    • No input capacitors 

    There are multiple input capacitors (47uF, 4.7uF, 1uF, 100nF)

    • U7 

    I removed U7 and used a fixed 1kOhm resistor. But the same issue. When no load, Vout is correct. When load is connected, Vout drops to 0V and no current. 

  • 0 in reply to Steven Wang
    TI__Guru 59611 points

    Hello Steven 

    • Connect DEMB pin 
    • Connect DAP 
    • I couldn't find the AGND-to-PGND connection. 
    • Connect the ground connection of the sense resistor to power ground. 
    • Please show the schematic including COUT and also send me the completed excel quick start file.
    • CM capacitor is connected to PGND now. It should be connected to AGND.

     -EL

  • 0 in reply to EricLee
    Prodigy 230 points

    Hi Eric, I have updated the schematic. It includes all circuit information now. 

    • Connect DEMB pin 

    I connected DEMB to PGND, but the Vout drop with load issue still exists. 

    • Connect DAP 

    DAP is at the bottom side of LM5117 chip. So I cannot manually connect it the GND. It's currently floating. 

    • I couldn't find the AGND-to-PGND connection. 

    I connected AGND to PGND via a net tie. The net tie is basically a wire. 

    • Connect the ground connection of the sense resistor to power ground. 

    I used a jumper wire to connect current sense resistor to PGND. 

    • Please show the schematic including COUT and also send me the completed excel quick start file.

    Schematic is updated. Webench simulation circuit is provided as well. 

    • CM capacitor is connected to PGND now. It should be connected to AGND.

    I removed the C19, so C pin connects to ADC directly now. 

  • 0 in reply to Steven Wang
    TI__Guru 59611 points

    Hello Steven

    You have to design the PCB properly to ensure the device functions properly. I can review your layout if you share it with me.

    -EL 

  • 0 in reply to EricLee
    Prodigy 230 points

    Hi Eric.

    I have attached the schematic and layout to this reply. Please note that I did the following things in the layout: 

    1. Cut the wire between UVLO and VCCDIS

    2. Use a jumper wire to connect the current sense resistor GND side to PGND. 

    I simulated the circuit in PSpice, the simulation result seems to be good. 

    Some new findings:

    My Vin=48V, Vout=32V. Although Vout>14.5V (I should use external VCC in the datasheet), I'm using internal VCC = 6V. Could this cause the issue? 

    If I use a fixed resistor to replace the I2C digital potentiometer, still the same issue.

    When <=0.2A load, the buck is fine.

    When >0.2A load, buck shuts down and VCC drops to about 2.8V. When load is removed, Vout cannot recover to 32V. I must re-start Vin=48V, then Vout becomes 32V again. 

    LM5117_Schematic_Layout.zip

  • 0 in reply to Steven Wang
    TI__Genius 17110 points

    Hi Steven, 

    Because of the Thanksgiving holiday in U.S., so I will respond to your questions in the meantime. I saw your AD files and there are 3 pcs LM5117 in your design. Is it only the U4 path that has the problem when Io>0.2A? And is VCCDIS in a floating state now? Looking forward to your reply. Thanks.

    Aurora

  • 0 in reply to Aurora Tie
    Prodigy 230 points

    Hi Aurora, 

    Thanks for the support. Yes, VCCDIS is floating now, I cut the wire.

    All LM5117 have the issue that Vout is correct when no load, but drops to 0V when connecting to a load. 

    Only U6 could draw max 0.2A current. 

  • 0 in reply to Steven Wang
    TI__Genius 17110 points

    Hi Steven, 

    Can you re-test SW waveform? From the previous waveform, SW seems to be unstable. I'm not sure if there is some problem with loop design. In addition, there seems to be a lack of connection between the AGNDs of the three LM5117s in the layout. In addition, regarding the layout design, I think the driver problem is relatively big. You can refer to the layout section of Section 9.4 of our LM5148 datasheet for specific design considerations. 

    You can also try adding a 5ohm drive resistor to the HS mosfet to see if there is any improvement or connect DEMP to VCC. What's more, considering the piezoelectric effect of ceramic capacitors, the output capacitor can also be increased. Thanks.

    Aurora

  • 0 in reply to Aurora Tie
    TI__Guru 59611 points

    Hello Steven 

    Additionally, I found the connection from the low-side MOSFET to the sense resistor is too thin, and the ground connection of the sense resistor is floating now. 

    -EL 

  • 0 in reply to EricLee
    Prodigy 230 points

    Thanks Eric and Aurora, I have tried all methods above. The buck could provide stable Vout when Iout<=0.7A. I guess it's the layout issue. My plan is to buy a LM5117EVAL/NOPB evaluation board to validate. Meanwhile, I will draw a new layout. I will update the result with you later (probably in 1-1.5 months).

    Thank you very much. You guys are very professional ~

  • 0 in reply to Steven Wang
    TI__Guru 59611 points

    Hello Steven 

    Thanks for choosing TI. Please feel free to contact us if you have any question later. Please create a new post when the layout is ready for review.  
    -EL