The switch on the enclosure probably does not affect the power
feeding the USB part of the interface. The enclosure might look
like this. The "left side" of the USB bridge chip is bus powered.
The HDD is powered by the adapter. The switch is basically only
affecting the HDD.
+5V ----- USB to ribbon pwr /
D+ ----- IDE adapter ------- HDD -----x x--- Enclosure +5/+12
D- ----- chip inside power source goes
GND ----- enclosure to the drive
But the chip in the enclosure could potentially notice what is going
on, on the HDD side, and alert the computer. Just one of those
tiny design details.
A USB bus reset should also be a suitable event, to cause the
bridge to restart its sequence with the HDD.
I tried looking at the documentation for a TUSB6250 on
www.ti.com,
and I don't see how the USB bridge detects that the drive is powered.
(I selected TI because I thought I'd get a good datasheet. Other
companies are more likely to be providing the chip used in your
enclosure, but aren't likely to hand out a datasheet on the web.)
What you got is the luck of the draw. Maybe another enclosure design
does it right.
If your motherboard has a USBPWxx jumper block, you could try setting
it to the +5V position, rather than +5VSB, for the USB stack you are
using. When the computer sleeps, that will remove power to the USB
side of the interface. And at least then, the USB bridge chip will
go through its initialization sequence, when it receives bus power.
For whatever that is worth. Note that really modern motherboards,
have removed that jumper block, and the user is stuck with a constant
supply of +5VSB on the USB plug. Treat this change in the jumper setting,
as an experiment, because there is no guarantee it'll work any better.
It is just something to try.