Re: USB won't recognize sophisticated devices.

Gary Sams wrote:


Sorry, but this is way over my head. Are we any closer to knowing why ny mouse doesn't work?

Thanks

Gary


When the computer is not able to set up the necessary "endpoints",
for communications, that is a communications problem. It doesn't
get any more fine grained than that. Either the motherboard is
at fault, or the device is at fault, or some cabling is bad
and so on.

About the only other thing I can suggest, is the use of a USB
hub as a "cleaner" of USB signals. There were isolated cases in
the past, where a USB peripheral had high "jitter" in data signals.
(A certain music player was involved I think.) That means the 1's
and 0's didn't arrive at exactly regular intervals. Some motherboard
USB ports (Nforce2 Southbridge) couldn't "see" those devices with
high jitter. The workaround was to use a *USB2* hub, because USB2
hubs reclock and regenerate the data on the way through. As long
as the hub can deal with the jitter, the signal passed back to the
computer is clean. And if the hub is self-powered, with a wall wart,
then you know it has a good source of power for the peripheral.

That test is worthwhile, if you already have a USB2 hub on hand,
to try it. It might turn out to be a waste of money, if you
have to go out and buy one for the test.

While I have seen one USB device, intended for testing USB ports,
I don't know if owning such a thing would leave a user further
ahead in the long run (say, test passes, but real device still
doesn't work). You would almost need a USB protocol
analyzer, to get to the bottom of an unknown communications
problem.

http://www.lecroy.com/tm/products/ProtocolAnalyzers/images/PTP_Session_Decode.jpg
http://www.lecroy.com/tm/products/ProtocolAnalyzers/PDF/USB_high.pdf

One test I like to use on my computers, is booting an alternate
OS when things have "gone South" on me. For example, lspci,
lsusb, and dmesg in Linux, can give some enumeration info on
various bus interfaces. If you can see and use the hardware in
an alternate environment, that tells you the OS or drivers are
implicated, back in Windows. The distros I have on hand, are
Knoppix (knopper.net) and Ubuntu (ubuntu.com). The distro is a
700MB download, and is a ISO9660 file. A program like Nero, can
be used to burn a 700MB CD, based on the download (converts ISO9660
into a CD). There are smaller distros (puppylinux ~100MB), if you
aren't on broadband, and needed something smaller. But I haven't
tested that one, so cannot comment on quality. A lot of the
700MB is crap, and doesn't really help make a good test environment,
but I cannot say how small a distro could get, and still have
useful functionality. The very smallest Linux environments, have
been booted from a floppy.

Paul
.

Relevant Pages

  • Re: USB wont recognize sophisticated devices.
    ... Either the motherboard is ... About the only other thing I can suggest, is the use of a USB ... The workaround was to use a *USB2* hub, ... The distros I have on hand, ...
    (microsoft.public.windowsxp.hardware)
  • HIGHMEM slows down 2.6.11-rc3-bk7 machine
    ... disabled in BIOS, mainly USB/NET/FIREWIRE/SATA stuff. ... Linux agpgart interface v0.100 Dave Jones ... usbcore: registered new driver hub ... ehci_hcd 0000:00:1d.7: new USB bus registered, ...
    (Linux-Kernel)
  • Re: 2.6.5-rc2-mm4 (and 3) IRQ problem
    ... ACPI: Power Resource ... ata2: SATA port disabled. ... ehci_hcd 0000:00:1d.7: new USB bus registered, ... ehci_hcd 0000:00:1d.7: root hub device address 1 ...
    (Linux-Kernel)
  • Re: why can i see a floppy?
    ... Ignoring ACPI timer override. ... PCI Interrupt Link, ... usbcore: registered new driver hub ... ohci_hcd 0000:00:02.0: new USB bus registered, ...
    (Linux-Kernel)
  • USB in 2.6.X kernels
    ... it has a NEC USB 2.0 OHCI based controller. ... has anyone experienced USB OHCI host controller problems with new kernels? ... PCI Interrupt Link ... drivers/usb/core/usb.c: registered new driver hub ...
    (Linux-Kernel)