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.

TPS53915: DC-DC outputs seeing 2% difference and Vfb is out of spec range.

Part Number: TPS53915

I am using repeating TPS53915 with 12V input and 3.3V output. The output is rather different from the 4 DC-DC converter.

DC-DC1 = 3.418V ;Vfb = 0.618V

DC-DC2 = 3.268V ;Vfb = 0.592V

DC-DC3 = 3.288V ;Vfb = 0.596V

DC-DC4 = 3.416V ;Vfb = 0.618V

I have add 0.1uF feedforward capacitor in parallel with upper FB resistor.

I am doing a remote sensing by connecting a thin copper from the load and route it back to the upper FB resistor.

I also have a 10ohm resistor connecting from the local 3.3V to the FB as well.

I would like to know why the output of the DC-DC1 and DC-DC4 are with 3.58% error. DC-DC2 and DC-DC3 are around 1% error which is acceptable.

Also, the Vfb reading are different as well. I thought it should be in a tight range of 597mV to 603mV.

Could you help to advise why I am seeing this problem?

Default Fsw is 400kHz (no change of Fsw from PMBus)

DCDC1 Cff 220pF

No Load

8A load

DCDC1 No Cff

No Load

8A Load

DCDC2 220pF Cff

No Load

8A Load

DCDC2 No Cff

No load

8A Load

DCDC3 220pF Cff

No Load

8A Load

DCDC4 220pF Cff

No Load

8A Load

  • I am assuming the default switching frequency of 400kHz and vout voltage has not been changed by a pmbus command. 

    There are a couple of items to look into.

    Is DCDC1 and DCDC4 operating with a load current less than 1A. 

    The converter is configured for DCM operation at light loads I would expect the output to be a little higher, see figure 5 in datasheet.

    I would not expect 3%, maybe 1% increase for load dcm operation. 

    A feedforward capacitor of 0.1uF is quite large value to be paralleled with a 45.3kOhm. 

    What are the output voltages when not using the feedforward capacitors (Cff). 

    Try 220pF across the 45.3k

    Can you share a scope plots of the SW nodes for the DCDC1 and DCDC2 with and without Cff.

    If the SW node is not periodic and stable, changing the R14 from 150k to 0Ohm will improve stability. 

    Can you share a pcb layout of a DCDC1 and a DCDC2. 

  • Hi David, 

    Removing the Cff and reducing the Cff to 220pF did not help. The above scope shots in the main post is tried with MODE shorted to GND.

    As I can't share further any confidential information to you, please share with me your direct email so that we can discuss better.


    Kay Seah

  • Hi David,

    I have also noticed that the default Switching Frequency is not 400kHz.I did not use PMBUS to change any Fsw and I read the PMBus register D3 with readback value as 0xB2 which is 400kHz.

    DCDC1 – 590kHz

    DCDC2 – 385kHz

    DCDC3 – 521kHz

    DCDC4 – 326kHz

    Is there anyway that Fsw will be different from the default? The VCO in the controller should have locked to the desired 400kHz.

    I have checked my 12V input and the ripple is low.

    I can't find a way to upload scope shots photo without needed to upload to some website. Let's take this in the email.


    Kay Seah

  • BTW, the layout for the 4x DCDC TPS53915 are copied and the same. Could this be any possibility that the TPS53915 is faulty? How to tell?

  • my email is   ddaniels  (remove)   @ (remove)   We can correspond there.   

    The device does not have a VCO,  it uses an adaptive on time control some fsw variation is typical.    The schematic shows feedback network has two paths, one that is local  "ac path" and the other remote "dc" through the 0 ohm.   

    To rule out the devices, it is possible to swap DCDC1 and DCDC2 devices with each other.   We need to determine if the behavior follows the device or the location on board.     If the board location is suspected, we should review the "ac" and "dc" feedback paths.