• State Resolved
  • Locked Locked
  • Replies 8 replies
  • Answers 2 answers
  • Subscribers 64 subscribers
  • Views 3531 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.

TPS40057: TPS40057-based power supplier suffers high droop at heavy load

user3534421
Prodigy 50 points
Part Number: TPS40057
Other Parts Discussed in Thread: TPS40055, , TPS54331, OPT8241

Hi all,

My new board uses TPS40057 as the main switch controller. The schematic follows the recommended design in figure 15 in TPS40055.pdf, except some parts changed:

(1) MOSFET: AOD478, 150mR Rds_on

(2) RILIM: set to 162KR

(3) RKFF: set to 68KR

(4) Inductor: MHCC10040-100M-R7, the saturated current is larger than 8A.
All other parts are almost the same, except for the input and output capacitors, and that the input voltage is between 16V to 28V and the output voltage is 12V.

In our design we are requested to provide clean power to a device and it's required to withstand up to 20ms power loss, so the input capacitor reaches 10000uF and the output capacitor reaches 5000uF.

The ripple rejection is done by adding several 10uF MLCC directly after the inductor.

Currently the problems are:

(1) At no load the output is around 11.92V,

(2) while load at around 5A the output reduced to 10.4V,

(3) VFB is 0.5x instead of 0.7V,

(4) VCOMP is 3.8V then.

Out of 3 module boards however there's one of them works fine in given condition -- it worked happily at input 22V and output at 6A@12V! This board has VCOMP only at 3V when idle and 3.1V at light load.

Can anyone helps me to discover the issues and solve them? Thanks very much.

  • 0
    TI__Mastermind 23845 points
    We have received your question. I'll try to see if i can route your question properly.

    Thanks
    -Arief
  • 0
    TI__Mastermind 40040 points

    Hello,

    Please can you share a schematic?  Thank you in advance.

    Kind regards,

    David Baba.

  • 0 in reply to David Baba
    Prodigy 50 points

    Thanks very much. Here's the schematic:

    Part R123 and D35 is used for back-end current limit however I've tested TPS40057 current limit function so it's useless and will be removed then. This resistor consumps a lot power (the efficiency would be reduced to 60% at 6A load).

    The MOSFET RDON is 0.15V typical so the current limit resistor is calculated to be 150K. 162K is used actually.

    In the test board D35 is not populated; VFB is around 0.6V during heavy load (~5A) when output voltage only reaches 10.4V, and 0.68V during light (~1A) load when output voltage reaches 11.5V. I'll try to capture several waveforms later.

  • 0 in reply to Arief Hernadi
    Prodigy 50 points
    Thanks. I've posted the schematic. If there's anything I can do to help identify the problem, please leave me a message.
  • 0 in reply to user3534421
    TI__Mastermind 40040 points

    Hello,

    I suspect, you are hitting current limit?  Have you tried increasing the Rlim resistor to increase the current limit set point?  The current limit can be layout sensitive.

     

    Have you an LC filter on the 24V input bus, if so, this will need to be dampened?  this is typically implemented with a large cap with a relatively high ESR.  If not, I suspect the traces of Vin and Switch are not optimal in layout (not routed differential to their respective pins, causing excessive ringing and premature tripping of current limit?

     

    Some other less drastic things you can do to minimize the noise on the current sensing.

     

    1. Place an RC snubber from V Switch to Gnd, Try a 1 ohm and a 470pF and place right across the Bottom side MOSFET.

    2. Place a 5ohm series resistor in the Ho (pin 13 ) and the gate of the MOSFET. 

     

    Please can you check if things improve with these modifications?

     

    Kind regards,

     

    David Baba.

  • 0 in reply to David Baba
    Prodigy 50 points

    Hi!

    Thank you very much for your help.

    My PCB does use a common-mode inductor and TVS diode for input protection. Previous design uses TPS54331 however it is very easy to get burnt.

    I checked the design and found out the SS needs to be extended - I selected 0.47uF and it boots happily with 6A load. When SS is ok, selected RKFF and RILIM seems to work fine, and the output is able to keep 12V.

    Four 10uF/50V MLCCs are tied directly to D of the high side MOSFET, and two 4700uF aluminium caps around the power plane, so there won't be much ripple voltage in the VIN. There're also four 22uF/25V MLCCs at the out pin of the inductor. The output ripple is analyzed to be +-7.5mV when idle and at heavy load.

    The transient response seems to be slower (around 40 us), see the figure below:

    The max droop reaches 1.4V at 6A load, seems to be measurement issue. The result showed higher respond time with smaller feedback resistors (40 us to 100 us). There're also phase issues that causes ringing. 

    We have another design (OPT8241 related) that drives a high current illumination board, which requires up to 15 us respond time. What can we do to help reduce its response time? I tried increasing RILIM, no luck. I also tried to increase SW freq, however no luck either (the above figures).

    Thanks.

  • 0 in reply to user3534421
    TI__Mastermind 40040 points

    Hello,

    I suggest adding an RC in parallel with the top FB resistor, as it appears that you do not have these compensation components installed.  Suggest using 4.7nF and 100ohm.  This will increase the BW of your loop.  By increasing the loop BW, you allow the feedback network to respond faster minimizing the droop on the output.  Also what is the ESR/ESL of the 4700uF output cap? This also will effect the droop.

  • 0 in reply to David Baba
    Prodigy 50 points

    Hi,

    I added an RC as you told and now the output looks good. Instead of rising and ringing at heavy load (6A designed), the output now rises continuously for around 100 us (from VO - 150mV to VO) and the witnessed response time is able to reach 15 us.

    Though the droop is still the same with previous configuration, the ringing effect is however removed completely.

    The droop is around 150mV at 6A load. The ESR of MLCCs might be very small, and the aluminium capacitor may be some milli omega I suppose. The peak voltage drop seems to be ESL effect because I'm probing them from a plug.

    When load at only 3A, the output is incredibly stable! It recovered normal operation within only ~10 us! In this case I think the heavy load recovery curve should be the production of both high Rdson of MOSFET, and larger inductor value (calculated to be around 2.2 uH), and some routing issues (the bottom side MOSFET is not optimal routed -- the GND plane is splitted from the main GND plane).

    I'll use high switching frequency and very low Rdson MOSFETs and 1.0 uH inductor in the next design. I expect 15 us recovery at very heavy load at around 8V 15A. With these knowledges I suppose that won't be a problem.

    Thanks for your help.