I have no background in electronics, so these things aren't intuitive
for me. Please forgive me if this question is stupid, but if we're
saying that each of 10 USB port "draws one half amp," aren't we saying
that there's current flowing? Where is all that current going? Would
5 amps at 5 volts, generate 25 watts worth of heat?
Ron
As another poster pointed out, I was saying what would
happen in a pathological test case. A USB port can support up to
a 500 milliamp load. Drawing more current than that, causes the
Polyfuse in the circuit to open, shutting off bus power. By
default from the factory, normally the USBPWxx jumper blocks are
set so that the USB headers and connector stack get their power
from +5V. Since your power supply has 20 amps or more available
on that output rail, if the pathological test case arose, the
power supply could handle it.
If a user wanted a "Wake on USB" keyboard and mouse scenario,
they would change only the USBPWxx jumper block for the jacks
used for the keyboard and mouse. Usually the PS/2 connectors have
a similar option, if you wanted "Wake on PS/2" keyboard and mouse.
The rest of the jacks would remain powered from +5V in that case,
meaning there is little chance of exceeding the power supply's
limited amount of +5VSB.
In terms of power consumed when the computer sleeps, your PSU has
a limited output on +5VSB. The label on the side might list a
number like 2 amps or so. Only if you have a lot of devices that
remain powered when the computer sleeps, would you get close to
the limit. Since some power supplies are only 50% efficient in
the sleeping state, you could see up to 20 watts load at the
wall socket. Note that it is hard to measure this power accurately,
so if you want to do such a measurement, I recommend using a
"Kill-a-watt" type measuring device, combined with a PSU that
uses "Active PFC", as the waveform on an active PFC (power factor
corrected) power supply is close to sinusoidal, and makes getting
an accurate measurement much more likely. I've noticed my $400 meter
doesn't like the waveform from my non-PFC PSU, and gives ridiculous
readings on the AC side. Non-sinusoidal current waveforms tend to
exceed the bandwidth of the analog path in the metering device
(a typical good meter might only be capable of responding out to
50KHz or so - many meters will be worse than that).
Has anyone verified when the USB power is shut off ? Is there
a BIOS setting that works or not ? I can, for example, see
three Polyfuses next to the standby LED, and the proximity of
those parts suggests to me, that +5VSB is connected direct to
the Polyfuse. Normally, the powering path would look like this
on an Asus board:
USBPW12 "Typical Asus board"
(header)
+5VSB (always powered) ---X USB_jack_1
X---Polyfuse-+---+--X 5V
+5V (only when running) --X | | X D+
(filter --- | X D-
cap) --- | X GND
| |
GND | USB_jack_2
+--X 5V
X D+
X D-
X GND
I'm suspecting the A8N-SLI looks more like this, and it would
be nice to have someone confirm or deny that it works like this:
A8N-SLI Family hypothetical model (assumes no BIOS option)
USB_jack_1
+5VSB (always powered)---Polyfuse-+---+--X 5V
| | X D+
(filter --- | X D-
cap) --- | X GND
| |
GND | USB_jack_2
+--X 5V
X D+
X D-
X GND
A8N-SLI Family hypothetical model (assumes a working BIOS option).
(Here, either transistor #1 or #2 is turned on by BIOS code. I
don't see any chips or discrete transistors on the board, located
to do something like this. This would not be a good function to
integrate into the Southbridge either.)
+5VSB (always powered) ---+
|
Series-pass transistor #1
|
| USB_jack_1
X---Polyfuse-+---+--X 5V
| | | X D+
| (filter --- | X D-
Series-pass transistor #2 cap) --- | X GND
| | |
+5V (only when running) --+ GND | USB_jack_2
+--X 5V
X D+
X D-
X GND
In this Intel reference schematic, on PDF page 84, they use a small
8 pin device SI4501 to switch between +5V and +5VSB for a USB port.
Perhaps an Intel BIOS has an option to select one or the other...
ftp://download.intel.com/design/chipsets/schematics/25281202.pdf (9MB!)
Example of a small device suitable for switching USB power sources.
This is likely too expensive for Asus's tastes.
http://www.vishay.com/doc?70934 (SI4501)
Paul