Power supply design on Arty-Z7

[asmi]

Hi,

I'm currently designing a board with the same TPS65400 that is used on Arty-Z7 board. Looking at the schematic, I see that all ENSW pins of this IC are connected to +5V rail (VCC5V0 on schematic) via 20K resistor. However datasheet says that these pins are supposed to be referenced to VDDD pin and maximum allowed voltage is 3.6V, so having +5V there is beyond this restriction. According to datasheet these pins already have weak internal pullups, so they can simply be left unconnected if not used for power rail sequencing, but feeding 5V in them would force these resistors to sink current instead of sourcing it, which is probably not what they were designed to do.

I would really appreciate if someone would explain what I'm missing here as I'm really confused.

[jpeyron]

Hi @asmi,

Unfortunately, I do not have experience with the TPS65400. I reached out to the engineer that worked on this for the Arty-Z7 board to see if they had anything further to add to your forum question. Also have you reached out to the TI forums here?

cheers,

Jon

[asmi]

15 hours ago, jpeyron said:

Hi @asmi,

Unfortunately, I do not have experience with the TPS65400. I reached out to the engineer that worked on this for the Arty-Z7 board to see if they had anything further to add to your forum question. Also have you reached out to the TI forums here?

cheers,

Jon

I didn't contact TI yet. Their evaluation board (schematic of which I used as primary reference for PMU sub-circuit) agrees with datasheet and references these signals to VDDD. I generally don't use chips in out-of-spec modes if I can help it, so I'm wondering if your engineer who worked on this knows something I don't, or it is just a mistake that happen to not affect chip functionality because manufacturer was over-conservative when he declared "Absolute Maximum Ratings".

[jpeyron]

Hi @asmi,

First off good catch. We will need to prepare a minor revision to fix this issue. Our engineer that design this part of the board wanted to let you know that your assumption is correct.

thank you,

Jon

[asmi]

Thanks for the update! It looks like the problem can be solved by removing resistor R168. Unfortunately I don't have this board (yet - gotta pool "hobby cash" for a couple of months to buy it), so I can't test if this works or not.