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.

TPS7A4501 load transient response vs. temperature

Other Parts Discussed in Thread: TPS54620

Good morning.  I have the TPS7A4501 configured as a 1.2V power supply.  The feedback is 0 ohms to the output with nothing to ground.  The input voltage is 1.8V, supplied from a plane layer, and there is 10uF ceramic on the input.  The output capacitor is 2x 10uF X7R (plus a bunch of high frequency decoupling).  The load is the serdes power supply for this ASIC.  The data sheet says that the worst case current is something like 0.96A in the configuration I have.  It works fine at room temperature.  However, I have to get it to work at -40C.  The problem is that when reset is released for the ASIC, there is a load transient that pulls the 1.2V low enough that it triggers the reset controller, and then I'm stuck in a loop.  In the stock configuration, this starts happening at about 0C.  I can't observe the load transient at room temperature.  It's probably there but with the board configuration I have it is too hard to measure.  So, therefore, at least one of these things is true:

  • The ASIC has a load transient when reset is released that gets bigger as the temperature goes down.  (I'm working with the vendor on that.  It is not TI.)
  • The TPS7A4501 has a load transient response that gets worse when cold.  It has to be the transient response, because the current limit on this part goes UP when cold.

There is another guy who has the same ASIC on a different board.  But his power supply is a switcher based on TPS54620, with 2x 330 uF aluminum-polymer.  There are other loads on 1.2V, but he still has about 400-500% margin.  His board works at -40C.

I tried increasing the output capacitance.  Currently I have 4x 100uF ceramic plus 68 uF tantalum on the output.  That got me to -23C.  The brute force solution improves the response, but we aren't going to make it to -40C at this rate.

So, the questions are:

  • What is the expectation for load-transient response vs. temperature?  Should it get better or worse?  (There is data on the output current limit.  It goes up when the temperature goes down.)
  • Is there a way to change the design to improve the load transient response?

Thanks,

Boris