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.

TPS54332 power supply using design generated on webbench has output voltage drop significantly when loaded even lightly.

Hi,

Under idle conditions, the circuit outputs 11.5V, when putting on a 12V fan rated for 0.1A, output goes down to 10V.  Other loads similar drag the output down.

I am using a 16V input.  There are two variations from the webbench design. 

1) I am using a differently branded 8.2uH inductor, but it is rated similarly (tolerance, series resistance, saturation current, current rating, etc) to the TI recommended part. 

2) feedback is using a 10K ohm and 750ish ohm resistor, but this should only have a minor effect on the setpoint.

Neither of these explains the behavior of the circuit. 

I have fabricated two test boards, both of which exhibit the same behavior.  On one of them, I also tried a different webbench-calculated compensation network of 31.6k ohm resistor, 1nf, and 1pf, which did not fix the problem either.

Schematics and layout attached.  Thanks :)



  • For the FB resistors you mention, 11.5 V is the proper output.  If the Voltage is falling it could be due to layout.  It is hard to tell form the blurry bitmap image.  Can you send it in PDF format with individual layers?  Or possibly even the original PCB file (I have viewers for many types)?  Also can you post some waveforms including the switching node?  I notice that WB only called for 2 x 4.7uF output capacitors.  That is not very much.  That fan probably has rather high start up current.  You may need 100uF or so.

  • Thanks for the quick response :)

    I have captured the switching node (NetC7_1) with and without the load.  With no load, the regulator goes to a sporadic operation probably for higher efficiency.  With the fan connected, the regulator is on and regular each cycle.  Neither looks like proper operation for a buck regulator.

    During testing, I tried with and without a 470uF electrolytic output cap.  This made no difference, but as this is steady state behavior, I don' think startup has much to do with it.


    Attached are also higher resolution screenshots of the top and bottom layers. 

  • Any thoughts on this?
  • I'll take a closer look on Monday.

  • The waveforms look normal to me. In no load condition you are operating is eco-mode. With load, you are just in DCM, not CCM so you will see that charcateristic ripple. The layout is far from ideal, especially the location of the output inductor and positioning of the catch diode and input capacitors. It looks like your final 12V output is a pretty thin etch. How far does it go to your fan?
  • I will review the layout guidelines.

    The fan (really just for testing) connects to the board through a pin header about 6cm away from what you see there.  A separate decoupling capacitor (ceramic 4.7uF) is located by the connector.

    The intent of this power supply was to drive another board (with multiple dc/dc regulators on it, roughly 10-15 watt load) through a ~20cm cable.  Using the board results in very undesirable behavior as its regulators switch off and on repeatedly with their own UVLO settings.

    Thanks :)

  • Most likely it is IR losses that are causing your issues.  If you have to re-pin your PCB, you might as well optimize the layout.