USB hubs don't work on windows 10

[Nominal Animal]

It's definitely a windows or motherboard issue with hubs.

Have you considered running another OS from a Live USB stick (or other external drive), to find out if it happens in other OSes also?  These USB peripherals (including any storage formatted for Windows NTFS) should be fully supported in Linux and FreeBSD, and the only thing to watch is to see if the devices drop out there also.

My experience with Gigabyte motherboards has been generally good, although a voltage regulator on my GA-MA78GM-S2H did explode (the IC cratered).

I have heard from others that mixing USB 3.x and USB 1.x devices on the same port or hub in Windows is "problematic".  (The problems I've heard were related to uploading firmware to microcontrollers with native USB 1.1 (full speed, i.e. 12 MBit/s) interfaces, when connected to a host USB 3.x port or a USB 3.x hub.)  The workaround that I've seen reported to work is to use an USB 2.0 hub between the USB 1.1 devices and the USB 3.x trunk.  While this is purely anecdotal stuff, it is something to check.

[Bassman59]

It's definitely a windows or motherboard issue with hubs.

Have you considered running another OS from a Live USB stick (or other external drive), to find out if it happens in other OSes also?  These USB peripherals (including any storage formatted for Windows NTFS) should be fully supported in Linux and FreeBSD, and the only thing to watch is to see if the devices drop out there also.

My experience with Gigabyte motherboards has been generally good, although a voltage regulator on my GA-MA78GM-S2H did explode (the IC cratered).

I have heard from others that mixing USB 3.x and USB 1.x devices on the same port or hub in Windows is "problematic".  (The problems I've heard were related to uploading firmware to microcontrollers with native USB 1.1 (full speed, i.e. 12 MBit/s) interfaces, when connected to a host USB 3.x port or a USB 3.x hub.)  The workaround that I've seen reported to work is to use an USB 2.0 hub between the USB 1.1 devices and the USB 3.x trunk.  While this is purely anecdotal stuff, it is something to check.

Maybe this is all related or not, but the boss recently gave me a new Windows 10 box (Dell Xeon thing) and it does something fucking annoying. I have three monitors connected to the computer, and I use the USB 2.0 hubs in the monitors for, among other things, to connect my keyboard (Unicomp) and mouse (Kensington Expert Mouse) to the computer. Each Type B port on the monitor goes to a host port on the computer, which I assume are all USB 3.0 (there are couple of USB-C ports on the computer's front, too).

Before I leave the office, I log off the computer and turn off the monitors with their front-panel buttons. And every morning when I get in, I turn on the monitors and wiggle the mouse or hit the Escape key. Sometimes the computer wakes up, sometimes not. So I tap the power button to wake it up. Then I can move the mouse to the password box, enter the password, and I'm in. After this, the mouse moves but the computer won't respond to mouse clicks. It is fucking weird. Sometimes reconnecting the mouse solves the problem. Sometimes not. So I have to reboot. It's stupid.

[JDubU]

Maybe this is all related or not, but the boss recently gave me a new Windows 10 box (Dell Xeon thing) and it does something fucking annoying. I have three monitors connected to the computer, and I use the USB 2.0 hubs in the monitors for, among other things, to connect my keyboard (Unicomp) and mouse (Kensington Expert Mouse) to the computer. Each Type B port on the monitor goes to a host port on the computer, which I assume are all USB 3.0 (there are couple of USB-C ports on the computer's front, too).

Before I leave the office, I log off the computer and turn off the monitors with their front-panel buttons. And every morning when I get in, I turn on the monitors and wiggle the mouse or hit the Escape key. Sometimes the computer wakes up, sometimes not. So I tap the power button to wake it up. Then I can move the mouse to the password box, enter the password, and I'm in. After this, the mouse moves but the computer won't respond to mouse clicks. It is fucking weird. Sometimes reconnecting the mouse solves the problem. Sometimes not. So I have to reboot. It's stupid.

I've had similar problems with wired mice and keyboards not working at all before Windows boots up (for example, with BIOS settings).  Turned out that they were having problem with USB 3 ports.  When plugged into USB 2 ports, they worked fine.

[Nominal Animal]

Turned out that they were having problem with USB 3 ports.  When plugged into USB 2 ports, they worked fine.

This (and the other anecdotes above) sound very similar to what I've heard.

The key thing to note here is that most HID devices - keyboards, mice, joysticks - are actually USB 1.1 devices, and USB 2.0 compatible; they only transfer data at low speed (LS, 1.5 Mbit/s) or full speed (FS, 12 Mbit/s), and not at USB 2.0 high speed (HS, 480 Mbit/s).  It is exactly the low/full speed devices, 1.5 Mbit/s or 12 Mbit/s, that some USB 3.x host ports and hubs have issues with.  (And often they can be worked around by adding an USB 2.0 hub in between.)

I am not an USB expert, and know very little about the actual hardware implementations, but it seems to me there are some (common?) USB 3.0 chips that do have issues with USB 1.1 (or USB 2.0 LS/FS) devices without help from a software driver running on the host OS.  Or perhaps they need some specific hardware bootup sequence that some BIOSes don't get exactly right.  Or perhaps the chips' datasheets have a note that a certain resistor or capacitor has to be present for initial bootup USB 1.1 compatibility, and those are omitted from some implementations because the software driver "already does it"; cost-cutting.

Because of this, I do recommend using an USB 2.0 host port and hubs dedicated for HID (and USB 2.0) devices, separate from the USB 3.0 ones.

If you don't have enough ports, then splitting off the HID (and USB 2.0) devices from the USB 3 "trunk" using an explicitly USB 2.0 hub (as opposed to USB 3), may also work.  (That is, it has worked for those having issues with USB 1.1/USB 2.0 LS/FS native-USB microcontrollers; I myself have not had any issues with this.)

When I am working on a microcontroller circuit connected to my computer and also externally powered, I use a cheap ADuM3160/4160 USB isolator – LS/FS only; selectable via switch, so only 1.5Mbit/s or 12Mbit/s, no 480Mbit/s support! – from fleabay, followed by a powered USB hub (one using an isolated switchmode supply, so I can "ground" it to whatever ground potential I choose).  This protects the host computer from my own idiocy (like shorting the USB bus accidentally, or even backfeeding a higher voltage).  These cheap USB isolators all use the recommended schematic shown in the ADuM datasheet, only differing in the DC-DC converter they use which in this particular case is basically irrelevant and can even be removed since the downstream USB hub will provide the (isolated) power; but, if you want a good one, the Olimex USB-ISO is quite nice in my opinion.

That kind of "isolated slow USB branch" can be useful for other situations too, for example when fixing/checking suspicious USB devices.

[Simon]

I've been using a USB2 hub.

[Nominal Animal]

I've been using a USB2 hub.

Yes; that is what differs from the other reports I've heard, and makes me interested to find out the underlying issue.  (Although I don't use Windows, I do want my HID widgets to work there too.)

Do you have a "pure" USB 2 tree down to the problematic devices, starting from an USB 2.0 port on the host?  Could you set one up for temporary testing?

[Simon]

I'll try the USB2 hub on a USB2 port but I did this before, maybe now with the new drivers it will work.

[mikerj]

There are dozens of windows 10 machines at work (all Dell desktops and laptops) that use hubs to connect multiple debuggers, serial adapters, measuring instruments etc. and I've not heard of anything like this.  Occasionally something will stop working and will need to be unplugged and re-plugged, or very rarely a bad USB cable will cause issues but generally it's pretty solid.

[Bassman59]

I've been using a USB2 hub.

Yes; that is what differs from the other reports I've heard, and makes me interested to find out the underlying issue.  (Although I don't use Windows, I do want my HID widgets to work there too.)

Do you have a "pure" USB 2 tree down to the problematic devices, starting from an USB 2.0 port on the host?  Could you set one up for temporary testing?

The monitors I mentioned above all have USB 2.0 hubs, and I tried using a standalone powered USB hub as well, and I keep having this problem! The computer (again, a Dell Xeon) doesn't have USB 2.0 ports, only the USB 3 (blue inserts in the jacks) and USB-C ports.

I have all sorts of USB devices connected to the computer, mostly the usual array of JTAG programming things and USB-to-serial adapters. There is one USB-to-serial adapter that occasionally gives me fits and requires a re-attach to get the computer to recognize it.  I don't recall having any trouble with anything else.

Oh, as an aside, the computer also has an Intel gigabit Ethernet PCIe card installed. For awhile after going to sleep, the computer wouldn't recognize that interface. It wouldn't even show up in the Device Manager. A reboot solved the problem. Maybe a recent Win10 update fixed that, because it hasn't happened in a few weeks.

I will also point out that my 2012 Mac mini and my 2020 iMac both have only USB 3.0 (and Thunderbolt 2 for the former and TB 3 for the latter) ports, and I've never had any problem with keyboards and mice not working properly, either with direct connection or through a hub.

I just wanna know why one of the largest and most reputable manufacturer of PCs can't get its products to properly handle keyboards and mice.

[Simon]

It's the first time I have had so much trouble and it all worked initially so I am suspecting a change in Windows 10 or that it replaced a driver.

[Halcyon]

Speaking of Windows 10 USB shenanigans: My work PC has decided that it wants to continually disconnect and reconnect USB drives randomly now. Hooray!

[msuffidy]

I have an external hard disc that I found out was reconnecting every time the cable was jiggled. I thought it was the drive's socket. I do have some cables here that I figured out fit so I was going to try another cable to see if the problem still existed. I dual boot win10 and linux, so in terms of the hub it could rule out the hardware from the software if it worked in linux.

[Simon]

Yes I have had a few susceptible cables like that. Just don't touch the damn things