USB-C power delivery standard incompatibility

[nightswimmer]

While doing some tests connecting the behavior board to USB-C ports with power delivery feature, I noticed there is an issue when the behavior board is also connected to the external 12V power supply.

The issue happens when the board starts connected to the external 12V but NOT to the usb hub or pc. All of the issues descrived below only happen with the 12V connected.

In these conditions the board has around 1.7V on the USB power rail (that I am assuming is leaking backwards from the 3.3V voltage regulador). So then when we plug the usb connector it does not work, since the PD control system detects there is 1.7V where there should be 0V, so it will not turn on. If we connect the USB first, and then the 12V afterwards it works fine.

But the results get more confusing if we use a USB-C to USB-A adapter and then a USB-A cable to the behavior board. If we connect the adapter to the USB-C port, and then the cable to the adapter, it works. But if we have the cable already connected to the adapter and the board, and then connect the adapter to the USB-C port, then it doesn't work. This is because the adapters usually negotiate the power and request 5V from the port. So when we connect just the adapter the power is negotiated and we get 5V at the output before the board is connected. But if we try and connect all at the same time, the adapter requests the 5V but since the port detects the 1.7V on the line it will never turn on. From other tests I've done using different boards, the PD system would only turn on the power if we had around 0.5V or less on the line.

This can be a bit confusing specially since most users won't likely understant what's causing the problem. And considering USB-C is becoming more used every day, it would be interesting to account for this in future versions of the board (and most likely any other boards that use an external power supply and likely will have similar issues).