Can U3S6 be used on front part of case

Discussion in 'Asus' started by Jim, Apr 25, 2012.

  1. Jim

    Jim Guest

    Jim, Apr 25, 2012
    #1
    1. Advertisements

  2. Jim

    Paul Guest

    There's a cable assembly from Fractal Design, for sale on Newegg. The
    picture here, shows the interface standards involved in their
    2xUSB2/2xUSB3 assembly.

    http://www.newegg.com/Product/Product.aspx?Item=N82E16811994040

    The U3S6 has faceplate mounted, individual USB3 connectors. So
    I don't see a way to go from 9 pin USB3 metal connectors, to
    an IDT 2x10 (1 blocked keying pin) type connector.

    You could try this SilverStone, "internal USB3" adapter card with
    NEC chipset. $30.

    SilverStone SST-EC01-P
    http://www.newegg.com/Product/Product.aspx?Item=N82E16815256001

    The manual is abysmal. It shows a 19 pin pattern in the "layout photo"
    on page 5. So we're supposed to assume it has the correct pinout. What
    I'd want to do, is get a proper pinout from SilverStone, and compare
    it to any Asus or Gigabyte manuals with 2x10 USB3 headers, just to be
    sure there is an actual defacto standard.

    http://www.silverstonetek.com/downloads/Manual/accessories/Multi-EC01-P-Manual.pdf

    (As far as I know, PCI Express slots only have +12V and +3.3V. There is
    no +5V. Either the PCI Express card would need a +12V to +5V regulator,
    or they would need a power cable, and that's why there is a 4 pin power
    connector on the board. And it doesn't look like the SilverStone kit
    provides a power cable to use. You might need a Molex 1x4 Y cable for example.
    The manual also refers to "unlimited" current, meaning there is no fuse in
    the path, like a PolyFuse ??? Perhaps try it without the power cable first,
    and if it works, don't bother with cabling it up! If it really is unlimited,
    a short could burn something. Again, too bad the manual sucks.)

    Intel sometimes defines motherboard header patterns. They did patterns
    for AC'97 and HDAudio headers for example. But in the USB3 field,
    Intel has been way behind, and the Intel site is now such a "jungle",
    I wouldn't think of looking there for a new version of their "FP standard".
    At one time, things were easy to find on the Intel site. Now the
    site is defined by narcissist web designers, and Intel doesn't
    give a rats behind.

    This is an Asrock manual, supposed to be one of their first with
    the 2x10 in it. P55 Extreme4. I'd want to see multiple brands
    of motherboards with the same pinout, before deciding it's a
    defacto standard. (See page 31.) I have no idea what "ID" stands
    for, whether it is a no connect pin or not. If a connector
    has extra pins like that, they should really be used for things
    like an extra shield GND. The "ID" thing doesn't tell me anything.
    (What it could mean, is the motherboard senses a ground return from
    a cable assembly on that pin, so the motherboard could tell the
    front panel is wired up, but to what purpose ? It's not like this
    alone, would cause a driver to be installed or anything.)

    ftp://174.142.97.10/manual/P55%20Extreme4.pdf

    ID (pin 1) X X D+
    D+ X X D-
    D- X X GND
    GND X X SSTX+
    SSTX+ X X SSTX-
    SSTX- X X GND
    GND X X SSRX+
    SSRX+ X X SSRX-
    SSRX- X X VBUS (+5?)
    VBUS X (Pin missing = key = pin 20)

    When I'm unsure about a pinout like that, I use a multimeter
    to "buzz out" the wiring. Presumably, you can find a pinout
    for the nine contacts inside a USB3 connector somewhere, and
    then check for continuity from the appropriate 2x9 section,
    to the two USB3 connectors.

    I've found several of my Antec cases, mis-wired when I checked them
    (back in the USB2 era), so at least for my home builds, the
    "buzz out" is mandatory for anything front panel related. I like
    to be absolutely sure I got the wiring right.

    Here is a picture of a blue USB3 connector, with the pins labeled.
    Presumably there are other pictures like this to compare against.

    http://usb3expresscard.com/images/usb/usb3_pinout.jpg

    Good luck,
    Paul
     
    Paul, Apr 25, 2012
    #2
    1. Advertisements

  3. Jim

    Jim Guest

    Hi Paul and thanks for the detailed reply, I was worried the answer was
    going to be no (unless i jumped through tons of hoops) I guess it may be
    time to upgrade the motherboard/cpu/ram I have right now I only have
    USB2 motherboard support, I know USB3 native support for Intel is new
    but it's been around for a while on older boards so i'll hunt out when
    Asus started to do them, once again thx for the reply buddy.

    Jim
     
    Jim, Apr 26, 2012
    #3
  4. Jim

    Jim Guest

    Paul been doing some more digging and was wondering if this would do the
    job?
    http://www.kustompcs.co.uk/acatalog/info_3499.html

    Jim
     
    Jim, Apr 27, 2012
    #4
  5. Jim

    Paul Guest

    Yes, cosmetically, it solves the problem. Agreed.

    My concern is, with USB3, how many extensions can you
    make, without upsetting the gigabit per second signals
    being used ?

    Personally, if using the USB3 signals, I'd take them right
    off the faceplate of my U3S6, even if it meant getting
    behind the PC to do it.

    You can buy the kustompcs.co.uk and test it out, chaining 3499.html
    cable to fit the Fractal Design cable assembly inside (3499.html
    is male, Fractal design 2x10 is female). It'll mean removing a
    PCI slot cover on the PC, and routing the cable inside. And then,
    using a USB3 device plugged to the front to test.

    The kustompcs.co.uk looks to have good attention to RF design
    issues - it looks like it has a proper shielding solution, no
    gaps visible.

    The signal transmission should be protected by CRC, for each USB3
    packet sent. I presume a re-transmission is attempted, if there
    is an error. What you don't want though, is a significant error
    rate, because even with CRC protection, eventually an errored
    data pattern leaks through such a scheme. CRC is "armor plating",
    but is not "bulletproof". You could transfer some really big files
    (DVD sized) and use MD5sum or SHA1sum to check whether any bits
    in the files are getting flipped or not.

    If your testing shows this chain of cables is working for you,
    then keep using it :)

    I'd much rather pick off such a high speed interface, via
    a back faceplate connector, just for data safety reasons.
    I'm a conservative guy, having fought in the lab with
    high speed signals for so many years, and losing in the
    process :) YMMV.

    http://en.wikipedia.org/wiki/USB_3.0

    "The "SuperSpeed" bus provides a transfer mode at 5.0 Gbit/s

    All data is sent as a stream of eight bits which are scrambled
    and then converted into 10-bit format. This helps to reduce
    electromagnetic interference (EMI). [Drops usable datarate to 4Gbit/sec,
    minus packet overhead or the like.]

    It is still going to be tethered to 16 feet (maximum) cables"

    Considering the reach limits of SATA, by comparison, 16 feet is
    a hell of a long way. They mention "receive equalization", as
    a means for USB3 to work at those long lengths.

    http://www2.electronicproducts.com/...links-article-FAJH_TexasIns_Apr2011-html.aspx

    "The writers of the USB 3.0 standard added provisions for a
    receiver continuous time linear equalizer (CTLE) function
    to be applied to the received signal in order to open the eye.
    However, there are cases (such as long channels like that
    encountered when placing USB 3.0 connectors at the front panel
    in desktop systems, Fig. 4) that even with CTLE, the high-frequency
    loss and jitter are so severe that it is not possible to get an
    open eye after a three meter cable."

    Now, I located that article, *after* writing this post. You can
    see the challenges to doing a good job there. This is why I worry...

    Paul
     
    Paul, Apr 27, 2012
    #5
  6. Jim

    Jim Guest

    Hi Paul, the last thing I want to do is sort the "problem" out and end
    up with corrupted data, I think i'll take a punt on it purely because
    it's a easy fix and do some testing with a mix of file size I have some
    images that are 20GB+ in size so i'll shift them about a few times and
    see if i get any errors, if i do then i'll just bin the idea and upgrade
    the mainboard etc, many thx for your detailed answers, it's been a great
    help.

    Jim
     
    Jim, Apr 27, 2012
    #6
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.