View Single Post
  #2  
Old July 17th 03, 09:06 AM
Paul
external usenet poster
 
Posts: n/a
Default

In article , "da_xanadu"
wrote:

Hi,

I'm a first timer putting together my system., so excuse me if my question
is too elementary or my terminology is incorrect.

I'm in the final stage of connecting all the necessary cables but am at this
point stuck on one thing which is preventing me from going further.

Accoring to the P4P800 Deluxe specifications, the mobo features support for
8 USB connections. In addition to the 4 USB 2.0 ports at the rear panel, it
is also supposed to provide four more through internal headers.

Howerver, when I looked around I saw altogether two 9-pin connectors on the
board (USB_56 and USB_78). Does this mean that each of these is meant to
support two USB connections? (I guess USB_56 actually means this header is
meant for USB_5 and USB_6). If yes, how is that done--say, in terms of
cabling and connections?

(A bit of background info: My case (Kingwin KT-436WM) has four front panel
USB ports, two of which are meant to be connected to the motherboard. The
cables of these two ports have two 10-pin headers.)

Thank you very much in advance for any input!

da_xanadu


A single USB interface consists of two power connections (+5V and GND)
and two data signals (D+ and D-). So, that is what is required to make
it work. On the Asus 2x5 header, only nine pins are there, and the tenth
one is missing, to function as a "key" if the right kind of mating
connector is used. Of course, no case uses a USB cable with a key, but
that is besides the point.

A typical case has two sets of four (+5,GND,D+,D-) plus a wire called
alternately GND or SHIELD. The case GND/SHIELD wire has to be left
floating, because the Asus header only has eight working pins on it.
It seems to work without being connected.

If you have trouble getting USB2 rates to work, there are two
possibilities. One possibility is either cable or part of the case
assembly is not really USB2 compatible (high frequency characteristics
could be wrong). At least one case manufacturer will send a replacement
USB connector and cable to replace the USB 1.1 version they shipped by
mistake. The second possibility is the SHIELD wire is really needed.
You can pick up extra GND connections from the motherboard, by using
a GND pin from another header. This could be from another unused USB
header, or the center two pins of the AUX1 or CD1 1x4 audio connectors
are an alternate source of a GND connection.

BTW: The term +5 or VCC are used interchangably. There is no consistency
in the labelling of the connections, which is why for the uninitiated
this is so much of a puzzle. Since you don't want to blow up any
expensive USB peripherals, it is good to ask questions first.

Also, if you look at page 22 of this document:

http://www.formfactors.org/developer..._guideline.pdf

you can see that when Intel defined the header, eight of the pins
are as you would expect, while the ninth is something called
"overcurrent". That doesn't seem to exist and that is why the Asus
header has a NC (noconnect), because at the Asus end, they choose not
to listen to any device's equivalent of an overcurrent signal. However,
the stupid case manufacturers know that they should have a SHIELD signal
for some high speed cables, yet they cannot find a way of connecting the
SHIELD signal to one of the other GND signals in the cable assembly.
So, this leaves it to the user to leave the SHIELD connection dangling.
On some cable assemblies, the SHIELD is actually connected to GND at
the other end of the cable (someone in this group checked one with an
ohmmeter). Truly bizarre, and the result of standards makers not
understanding the practical aspects of mnaufacturing what they specify.
For most people, leaving the SHIELD unconnected seems to work fine.

HTH,
Paul