In fact, I don't believe that you can. Until the BIOS has had a chance to actually perform its POST processing, there is no opportunity to disable power delivery. Bottom line, in the very special case of the period of time between when you plug in power and when you press the power button for the first time (which initiates POST), I don't believe that you can avoid having power at the USB ports.
The one way to avoid this situation is to set the After Power Failure parameter to Power On. In this case, it will execute POST and boot into the default O/S when power is connected.
Hope this helps,
Also, I would update the bios. But, that is just me:
Doc (not an Intel employee or contractor)
Thank you for your reply. Sadly, the NUC7i3 (7th generation) kept the USB ports off until button press (After Power Failure parameter set to Power Off). Having to go to NUC8 (8th generation) appears to have changed that. Also updated the BIOS, but to no avail.
I'm must say that I'm confused. If I understand your question correctly you want to have USB ports without power when NUC is OFF (S5 state)? If this is your request, it can be done by disabling (unchecking) the USB S4/S5 Power in BIOS Advanced > Power > Secondary Power Settings (this will not affect the USB charging port - the front USB port with yellow color).
Thank you for your reply. That is the mode I have the BIOS set to: S4/S5 unchecked. When power is applied, CPU is off, but each of the four USB-A connectors are powered. Interestingly, 5VDC also remains available at the two internal USB 2.0 headers.
For the NUC8, the S4/S5 setting is under Power> Secondary Power Setting.
The NUC7 provides all four USB-A connectors to be disabled when power is first applied.
Yes, the power state of the USB ports can be controlled on entry to S5. Unfortunately, there is a window of time between power being applied and the BIOS having a chance to POST and set the state of these ports (which can only occur when the user presses the power button).
Ok, at I typed that, I came to the realization that this is still a BIOS (or F/W) defect. The BIOS *does* have a chance to run when power is first applied and it could set this state at that time. It is unclear why the operation of this NUC design differs from the others and whether this is a H/W or F/W flaw. This needs to be investigated.
Intel Customer Support: please assign resource to follow up on this with design team.