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.

TPS65400 Current Sharing Mode on EVM

Other Parts Discussed in Thread: TPS65400

TI,

I have a TPS65400 EVM, and I was trying to share SW3 and SW4.  In the EVM, I tied VOUT3 to VOUT4, and pulled COMP4 to VDDA.  The input voltage was 10V.

I modified the feedback resistors of the EVM so that I have a total of three outputs of 3.3V each driving a load of 1 ohm each, from SW1, SW2, and SW34 combined.  It was working fine for about 5 minutes and then the shared output of SW3 and SW4 failed.

When I went back to no share mode and removed all the loads, VOUT4 was putting out a sine wave of -5V to +5V, and the SW4 output is ringing from 0 to 10V every 200us. But the SW3 was also working in a funny mode.  It put out a correct 3.3V for a few second just after power up, and then it went to 1.6V and was trying to go back to 3.3V every half a second or so.

But then if I went back to sharing SW3 and SW4, the output behaved similar to VOUT3 above, but I can still see SW3 and SW4 putting out a square wave between 0 to 10V, but of course the duty cycle of the square wave is not correct.

VOUT1 and VOUT2 were all working fine all the time.

Obviously the SW3 and SW4 was destroyed in some manner.  But I do not understand how they can be destroyed just by going to current sharing mode, and the funny thing is that it can function correctly for a few seconds after I cool it down, so it imight somehow be related to temperature or charged capacitor stored somewhere.  I already used the software to set the registers back to the default states.

  • This morning, I use the software and wrote different bytes to the USER_DATA_BYTE_00 and 01, and it magically fixed the problem.  The VOUT3/4 then worked normally and output 3.3V to a 1 ohm load.  

    I was happy and so I repeated the test again, with three outputs of 3.3V each, from VOUT1, VOUT2, and VOUT3/4 combined.  Each was driving a 1 ohm load.  Then after about 5 minutes of run, the same problem came back.  But now I cannot fix this problem by using the same method.  After I did several STORE_DEFAULT_ALL, it even damaged the VOUT1, which is now putting out about 1.0V.

  • The problem is now shifted to always at VOUT1. The error status from the software shows it is overvoltage error while the actual output voltage is at only 1.1V. If I use software to turn off the output for a minute and turn it back on, then the output is good at 3.3V for about 10 seconds and then it goes back to this error.

    I don't know if it could be solder flux residual left over under the chip as my experience tell me that this kind of strange behavior is most likely due to bad soldering. We might need to order another EVM to compare.
  • Okay, I finally use a soldering iron to put a lot of heat at the back of the board behind the chip, and I was able to suck up a lot of solder from the thermal vias under the chip. The EVM is now working as expected. I am able to drive three loads of 1 ohm at 3.3V each. The outputs are all normal now. It really amazed me such an expensive EVM is not fully tested and customers need to fix the EVM ourselves. I am still always getting CML errors from the GUI software, but I am able to adjust all the parameters from the GUI menu.
  • Thanksa lot for your sharing. Actually, we have wrote an documents to our assembly vendor to test every board before shipping it to customer.
    will send email to them to take care next time.
    thanks!