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.

TPS53667EVM-769: Fusion Digital Power Designer - Writing Commands/Storing in NVM

Part Number: TPS53667EVM-769
Other Parts Discussed in Thread: TPS53667

I am designing a power supply for a core rail.  I am looking at using the TPS53667.  I got the dev board and downloaded the fusion digital power designer.  I got connected and I am able to talk to it.  I have a separate bias supply connected to 5V, so I can talk without turning on my main power.  The only thing that I am doing is changing the factory default 1V output to 0.8V under the Vout Command on the GUI.  It shows that I changed the value and then I write it to the device.  It successfully writes it to the RAM.  Then I turn on my main supply (12V) and it starts regulating 1V not 0.8V.  The GUI says that the setpoint is 0.8V but its not on my load and the monitors.  

However, when I write the 0.8V while the 12V supply is on, it will change the voltage dynamically while running.  Looking at the schematic, the 12V does not touch the TPS53667 at all, so I am not sure what is going on.  I have not changed any of the settings or pins on the board, so I am not sure what is going on.

Another note as well, under the monitor on the GUI, the output current is always 0 no matter how much I load the board.


  • Hello Christopher,

    I have some questions for you.

    1. How are you interfacing with the hardware? Are you using the TI USB-to-GPIO device?

    2. Are you only changing Vout Command, or are you also changing VBOOT?

    3. Do you get a valid input voltage reading from the GUI?



  • 1. Yes I am using the TI USB to GPIO

    2.  I did not change the VBOOT command, when I did it started at 0.8V, just missed it in the GUI.

    3.  Yes I get valid input voltage.

    The output current and input current are incorrect.  For example I was at 70A out, the Iout monitor was around 20A (I had 4 phases on).  The IIN was at 1.5A but on my power supply it was 3.2A.

    Another question, how do I write this to NVM so I can keep my settings the same?  I see the Store Config to NVM, and when I do it, it says successful, but when I power cycle it does not store it.  I feel like I am missing a step, but not sure what it is?


  • Hi Christopher,

    How are you performing your current measurements? Going from 12V in to 0.8V out, you would expect about a 14x current multiplication assuming 95% efficiency. 

    The output current and input current are incorrect.  For example I was at 70A out, the Iout monitor was around 20A (I had 4 phases on).  The IIN was at 1.5A but on my power supply it was 3.2A.

    With this test, the 70A/3.2A measurements doesn't make sense. 3.2A x 12V = 38W in and 70A x 0.8V = 56W out. That is 147% efficient, which can't be correct.

    The chip's data of 20A/1.5A are fairly close to expected values. 1.5A x 12V = 18W in and 20A * 0.8V = 16W out. That is 89% efficient, which is reasonable for having all four phases going at a low current.

    Also, try writing to VBOOT, writing to hardware, then store config to NVM and see if that gets the device to boot up to the correct output voltage after a power cycle. 

    If you have more questions or if this doesn't reveal your issue, I have an EVM on hand and I can try to replicate the issues you're having in the lab tomorrow.



  • Travis,

    So what was happening was that the Eload I was using was hitting the current threshold at what my voltage was at.  I just had to swap to a bigger load, At 70A out I am getting 90% efficiency.  I also went back through and was able to get it to flash to NVM, so I think I am good here.  

    Only thing is the current monitor was still not matching what I was measuring with the O-scope.  I had 13.44VIN 4.61AIN and 0.8VOUT 70AOUT, but the monitor was around 30Aout and 2.5A in, its not a big deal, there might be a setting I was missing, but it didn't really correlate to anything.  I had 4 phases on, and those were not multiples of what the monitor was reading, but its just an FYI.  On phase was 3C hotter than the others but it was due to a bad solder job on my part (swapped the inductors for what I want to use).