Colin wrote:
> I'm wondering if anyone else has had any problems using a powered usb hub
> with asus motherboards. I notice when I plug in the hub that windows XP SP2
> constantly loads and unloads the hub with the accompanying beep every few
> minutes. I'm wondering whether the voltage from the hub is backing up the
> hubs usb connection. The hub is a Belkin top of the line 7 port high speed
> usb 2 hub. Thanks for any info.
>
>
I can show you a schematic for a hub. The first is the schematic, while the
other two describe a couple of the major parts.
http://www.smsc.com/main/tools/usb/usb20h04evb.pdf
http://web.archive.org/web/20070103060933/http://www.smsc.com/main/dat...eets/20
http://www.micrel.com/page.do?page=/product-info/products/mic2026.shtml
The above schematic, shows a USB hub with both bus powered and self powered
operating modes. In self powered mode, a wall adapter is plugged into jack J1.
Next to the jack, they've located a relay. The relay changes the power source,
depending on whether the wall adapter is powering the unit or not.
The ports on the hub are distinguished. One port is the "upstream" port. It
comes from the computer, and should be using a different shaped jack. Or
perhaps a captive cord with USB plug on the end, is permanently fastened
to the hub.
The downstream ports go towards the peripherals. The downstream ports in this
case, are protected by the Micrel MIC2026 overcurrent detection and power
switching devices. The Micrel checks how much current a peripheral draws, and
presumably the main USB hub chip shuts off a port which is overcurrent.
So the key to your question, is one involving power. If you follow the logic
of that design, when the wall wart is being used, there is no path between
the wall wart +5V and the upstream USB connector's +5V. Only when the wall
wart is disconnected, is the upstream +5V rail connected (eventually) to
the downstream peripheral connectors.
The 5V power used in USB, is a regulated voltage, and is not suitable for
"diode steering". On Firewire, or perhaps TERMPWR on SCSI, diodes are used
to prevent power problems. But as far as I know, USB doesn't use diodes.
That schematic above, shows the use of a relay, to prevent two voltages
sources from fighting one another. (Somehow, I can't imagine a hub
manufacturer, actually using a relay, but the purpose of that schematic
is to "telegraph" design concepts to the reader - schematics like that
are not always the most practical way to build things.)
If it was my hub, I'd probably be plugging a cable into the upstream port,
and checking it to see if I can see +5V on the 5V pin of the connector. It
might be easier to check that, at the end of a USB cable plugged into the
port. Based on the schematic of a 4 port hub above, there shouldn't be
any voltage present, if the hub is self powered. (So to do the test, you'd
want the adapter for the hub plugged into it.)
A Polyfuse (sometimes green in color, and located near where the port is),
is a protection device used on the motherboard. If the Belkin tried to push +5V
up into the motherboard, the difference in voltages, across the low resistance
of the wiring, could result in a relatively large current flow. That could
cause the Polyfuse on the motherboard to open. The Polyfuse will recover
automatically, once it has had a chance to cool off. Which is why I'd suggest
checking the hub, to see if it is emitting +5V on the upstream port.
If probing for +5V on the Belkin upstream port, be careful not to short the
+5V to ground. The shell of the connector is grounded.
http://www.riccibitti.com/pc_therm/USB_A_pinout.jpg
Another possibility is a protocol problem. Perhaps the motherboard discovers
the hub, enumerates it, is quite happy, sends the usual regular polling packet,
gets no response for some reason, declares the port dead, discovers it again
etc. I don't know why it would be doing that. Especially if USB devices plugged
straight into the P4PE port otherwise worked OK.
The P4PE USB2 has a "USB 2.0 HS Reference Voltage" setting in the BIOS. That
is also something you could play with. That setting is not something typically
seen in the BIOS of modern boards, and very few boards have such an adjustment.
Presumably it has something to do with the physical interface to the
D+ and D- data signals. That setting is actually a register in the ICH4
Southbridge - Intel says only one value is valid, and others are "reserved".
So I guess Asus decided to allow the users to play with the value. I
wouldn't count on that fixing anything though.
You might also look for reviews of the Belkin hub, and see if others
have noticed the same problem. There is probably a part number stamped
on it somewhere.
Paul
>> Stay informed about: Asus P4PE motherboard and USB hub 
FAQ
Profile
Private Messages
Log in
