Western Digital starts selling a 314GB PiDrive

[tllim]

TLLIM   So what is the maximum current you recommend that a single USB port can carry, the std 500ma or is the board designed  for something greater?. I see the point where you are given the option to change the current source from either the on board PWM or select (through the jumper) to source from the external DC supply.

Looks like you use 100mil trace in internal pwr config feeding a 150mil trace to USB conn  = This looks like the  500ma USB std.

and

For external pwr current is feed through the 150mil trace which would be about 750ma per USB conn = This is a non std configuration.

Adding the two USB connectors would be about 1.5A max current draw from ext pwr (if configured to external pwr)

Even though you increased trace width for external pwr, to allow an extra 250ma to be available to a single USB conn it
would seem a little more common sense to to just use a powered hub and not stress the board (is not fused in ext pwr conf that I can see),
This is not all inclusive, with a proper pwr supply, good tracking of current draws, lots of things can be done, but all it takes is one brief short on the USB side and the trace may open up.

Also, is there a paticular reason (or application) you had in mind when designing this configuration (the external pwr one) Off loading the USB current perhaps?

Next time consider shooting for a 3A output and protect it with resettable fuses (maybe USB 3 will come down in cost and/or USB-C) for ext pwr. The SBC wars have just begun, I'm sure with all the input folks have been feeding into this forum have given rise to the next idea/approach, might affect cost/size, but nowadays product life is getting a little shorter each cycle.

The first sentence assumption is correct.

As mentioned, this implementation is reduce PMIC loading when connect to high peak current USB device such as hard disk. The USB current limited (we set at 680mA, slightly more than 500mA) setting still the same on both selection.