I've got somewhat of a bizarre problem with my other halves computer. Her system used to be mine (details below), and I never had any problems with it at all, completely stable. Now whenever the usb subsystem is loaded up (lots of data transfer over the net (usb dongle), syncing an ipod or trying to use a webcam) the pc just completely locks up randomly. It's a hard lock and can only be solved through a restart. I've tried both 32 and 64 bit vista with the same result. I've swapped out the power supply and can't seem to find the problem. It doesn't matter if I use the front usb header or the rear io panel ports the same crash happens as soon as they're loaded up lots. I've also tried using the default windows driver and the ones off the nvidia site to the same effect. I've not got a clue what it could be! My last thinking is that it is because I didn't have enough motherboard standoffs to put screws in all the holes, and that one of the ones missing is the grounding for the USB. Spec is: AMD X2 3800 thermalright XP120 Abit AN8-SLI 4GB Corsair PC4000 160GB Seagate HD
You know what, this is how I broke my ASUS motherboard (A8N-32 SLI)...but it was the Ethernet plug (both of them) that was broken, and I NEVER touched them while transfer... Luckily I was under my last year warranty at the time, and it took a full month for ASUS to give me the worlds most dusty motherboard.. probably from their own system.. and too cheap to attempt to remove the dust (no even blowing). I had to wait 1 hours on the RMA phone service, which was long distance (yea their flag-ship product don't have a 1-800 number, but their LCD department does... makes sense). Abysmal... just abysmal. I'll never touch ASUS motherboards again... luckily each department is like different companies. Oh and if you think it was just me... think again, I can give you 3 more stories. Anyway, when your computer freezes, does animation plays wile your computer lock-ups? Like if you play a video, do the USB thing.. does the video still play and all USB devices are not working anymore despite still being powered. I had once this weried issue... it was Nvidia Ethernet controller or circuit that was faulty and didn't last (before I touched my system).Did you install Nvidia ActiveArmor or Nvidia Firewall?
Goodbytes, it's a complete hard system lockup, i.e everything stops, including any video playback etc. I've not got the activearmor or firewall installed either. Think I'll get some more mobo standoffs and try putting all of them in, which will mean a complete rebuild, so i can check I've not got anything else wrong either! Looks like a weekend of building pc's as the parts for my parents new rig should be arriving today too!
I would say it's an OS problem and the drivers are configured wrong. But it could be an IRQ routing issue if the OS is using interrupts to service the controller. But this is based purely on the information provided. So, here are some questions that might help solve the issue. Can you tell which of the USB controllers the devices are connected to? Under XP, if you go to the device manager, clich on the view item and select devices by connection. Expand ACPI multiprocessor PC, then Microsoft ACPI-compliant system, then PCI bus. You should see the USB trident () on a number of devices. I believe you should have at least 4 USB open host controllers (OHCI) and at least one USB2 enhanced host controller (EHCI). If you do not have at least 1 EHCI, then check your BIOS to make sure it is not disabled. The item in the setup utility may refer to it as an EHCI, or as USB 2.0 support. If you have an EHCI, and a trouble causing device is not connected to it, try connecting it to a USB 2.0 hub. Then in Windows, make sure the hub is connected to the EHCI and the deivce is connected to the hub. This should provide more bandwidth for the device as it is not ont he faster USB 2.0 connection. If the device wa salready attached to the EHCI, you might try the hub anyway if you have one on hand. But I would suspect the drivers. So long as your cable runs for USB are not terribly long, and you are not using passive extension cables, then grounding would be low on the my list of suspects.
He said that he did not touch the BIOS and that the problem was not there before, also BIOS option don't change by themselves. You COULD try to uninstall from the device manager, everything that is related to USB, and let the OS re-install the drivers. But I think that the board bended and broke/damaged some wires, like I had. Also his BIOS does not have any option to enable/disable EHCI. I have a board from the same series (higher-end actually), therefore we use the same BIOS. All you can do (pretty much) is enable/disable USB legacy support and set the speed of USB (FullSpeed or HighSpeed). You could try playing with those, you never know.
If you really want, I can provide all the technical details about my thinking. I don't think readers want to read posts where I cut and paste from various industry specs and explain the responsibility of software when interfacing with that hardware. The short answer is that the USB controllers are PCI devices, and I assume there are PCI cards that were in use beforethe transfer, but may not be now. This couls affect the IRQ programming for PCI. The reason I would suspect IRQs is because I highly doubt Windows is polling the host controllers, that is just wasteful when they can generate an interrupt. Again, without going into the tech details, part of operating the controllers will be waiting for things to happen and this is where interrupts come into play. It is possible that Windows is waiting for an interrupt from the HCs, but does not get it due to a problem with the IRQ system or even be a reentrance problem with the ISR for the HC given that this happens under heavy traffic. Yes, that is exactly the question I was fishing for. High speed is USB 2.0 which is the EHCI(s). In contrast, OHCIs are fullspeed, which is USB 1.1. So, if set to fullspeed and not high speed, the the EHCIs will be disabled. These are the easy things to fix and explain away. I seriously doubt that the unused standoff hole is the problem. The ground plane is common for the entire mainboard. The reason for multiple points outside of the mechanical one is that, at RF frequencies, it is possible to set up a standing wave in the plane due to a potential difference and have it start acting as an antenna. But they are not critical, and all the scratch builds here on Bit using wood, plexi, fiberglass, and other non conductive materials should demonstrate that this usually does not cause a problem. As for physical damage to the ports, it would take a large amount of damage spread out all over the board for all the USB ports to be affected. If it was only one or two ports, then physical damage might be possible, but all of them? What happened, mult-layer PCB damage at the MCP? Even if it were physical damage to the mainboard, the result would be nonfunctional USB, not USB that works only when traffic is light. It could be a noise problem in the external cabling. USB has a limit of 15ft/3m IIRC. I have seen some devices that are more sensitive than others to cable length though. It could also be that the power demand by the device is not being sufficiently met by the host controller, and maybe that is related to the way the new owner uses the computer, or maybe (but not always) it is a sign of hardware failure. In both cases, cabling and power, the quick fix is to use a self powered hub.
Thanks for all the suggestions. I'll check for EHCI(s) when i get to the mrs house. I know that there are less non powered usb device connected compared to when i ran the board, as I used to use a usb keyboard, mouse, charger my iphone and use an unpowered external hd. In terms of cable runs, there are two devices that use long cables (1~1m, the other ~2m) which both run past some power cables. I'll try re-routing them away from the power cables and see if that makes any difference. My mrs also mentioned that the front ports on her case used to be dodgy too, so i'll try unplugging them from the mobo to make sure something in the case isn't shorting/bringing the whole lot to a freeze.
Ok just checked out the bios and it's set to usb V1.1 & V2, so should have high speed enabled. Checked in device manager and there are two usb root hubs listed, one running at high speed, the other at full speed. The network dongle is attached to the high speed one whilst the mouse, printer and webcam are attached to the full speed one. Last night i unplugged the wireless dongle and plugged her ipod in and managed to get it to fully sync all her music (something which reliably caused it to crash before). So from that I'm assuming it's a problem with the usb dongle. I've got a wireless pci card in my media pc which i'll swap for it and see if that solves the problem! Another interesting thing I just noticed is that the ipod shows up in device manager with the yellow triangle with an ! in it. According to windows i have the latest drivers for it, and I have the latest iTunes installed so it should be ok, however it doesn't appear to like it for some reason.
indeed, that is the case, there are current limiters that will trip when a device exceeds that, and you should get an error message. There is something called Powered USB, but it is not part of the USB spec at all, and not really accepted much outside of point of sale systems AFAIK. If the device is self powered, it does not matter what it draws from the external power supply as that has nothing to do with the bus. A quick note on the USB spec, it is freely, and publicly available from usb.org as a searchable PDF so it should be easy to verify claims about the basics of USB. Other stuff might take an engineer to understand though. Glad to hear you found the problem, and have a solution.
Hey guys, I'm not going to be able to check out the fix anytime soon as I just badly dislocated my shoulder so won't be driving anywhere in the near future. Thanks for the pointers and I'll let you know if it works when i get round to trying it out.
