TPS65217D Battery Free Configuration

Randy,

The first configuration turns on the TPS65217 because the USB supply is connected, which is a wake-up event for the device; there are three: USB detection, AC detection, and push button input is pulled low. (See Modes of Operation p.15). The power path does not switch to the AC input because there is no delta between the battery input and the AC input. There must be at least 190mV difference between these two inputs for the power path to prioritize the AC over the battery input. (See Power Path USB/AC detection limits p.8). Does your design allow there to be a 190mV delta between these two inputs?

Janice

Hi Janice,

It turns out that I have tried my configuration with a diode added between my 5V Bias Power Supply (connected to the PMIC USB input) and the PMIC VBAT terminal. This is successful at allowing the PMIC to switch to the System 5V Power Supply (connected to the PMIC AC input), but essentially behaves the same as my original configuration with the Battery terminal left open - the system powers up when AC power is applied to the system.

So, the solution that I think will work will be to add a transistor switch to apply 5V Bias to the Battery terminal when my system first gets AC Power. In this configuration, the PMIC appears to work the way I would like (does not automatically enable system power). When the system power switch is pressed, the bias for my transistor switch will disable the switch to the battery terminal such that the PMIC will switch to use the AC Power input. What I want from TI is to confirm that the behavior I am relying on is supported - can you assure me that the PMIC will not automatically turn ON in the condition where my 5V Bias is first applied to both USB input and VBAT terminals of the PMIC when the AC power is not present?

Sincerely,

Randy holmberg

Randy,

 

TI cannot guarantee that the PMIC will not automatically turn ON when power is applied to the USB input, because it goes against the data sheet.

 

Janice

Janice,
Rather than tell me that the actual behavior of the part that I am reporting goes against the data sheet, can you please find out why the PMIC actually does work the way I want it to in this configuration? If there is a reason why the PMIC can not be expected to reliably behave the way I am observing, then would TI be willing to make me a version of the PMIC (REV E) that is same as REV D, but does not automatically turn ON when 5V is first applied to the USB input.

If none of these options are viable, then I think I have my answer. You are telling me the PMIC is really not suitable for this application and I should replace it with something else - disagree???


Randy Holmberg

Randy,

I've asked design to look into this configuration, to get some more insight, and fine the assurance you are looking for; however, I believe this boiles down to a timing game: what node will receive the power first? Depending on that it will cause it to act differently (USB gets power first it will wake up the part, battery gets power first the part will not wake up without PB being presssed). Since there is no delta between the two nodes we are completely relying on timing. Have you tested this configuration multiple times, and it's worked every time?

Janice

Hi Janice,

That is very helpful - I appreciate your looking into this further for me. For my design, I can modify the configuration such that VBAT is guaranteed to power up before the USB input. If TI can provide some guidance that this will cause the PMIC to not power up with out a PB event, then that would be a big help. It would also be helpful if TI can provide guidance on how much delay I should plan on between VBAT power application and USB power application to have some reasonable assurance that the PMIC will wait for the PB input to turn on.

Here is what I think I would make my design look like for power up sequencing.

1) Provide a power switch to apply 5V power to VBAT initially.

2) A short time later a second switch will activate to apply 5V power to the USB input.

3) When the PB input is activated, the system 5V power will drive the AC input of the PMIC. This will also cause the switch to the VBAT terminal to turn OFF so that the PMIC switches over to draw most of its current from the AC input.

Any further guidance would be much appreciated.

Randy Holmberg

Randy,

Your set up will work provided that the Voltage rise time is fast enough. There is a power detected deglitch  time of 22.5ms (not tested in production) on the AC/USB pins, so there should be a sufficient enough delay that the Bat input will always be chosen. I would account for these delays in your testing procedures. From what you are describing it sounds that you are adhering to the spec outlined on page 8 "power path USB/AC detection limits"

 

Janice

Hi Janice,

That is all very helpful to me. Please also let me know if you are willing to review my schematic sheet with the PMIC connections. 

Thanks,

Randy Holmberg

425-770-6405