powering a usb cable

**evandude** · 08-02-2006, 01:19 PM

Originally Posted by turbocad6

see, now I thought the oposite was true, & that only special hubs can share both without reducing the 2.0 to 1.X.... I didn't think just any 2.0 hub can support both speeds simultaniously???

The book I've read this from (USB Complete by Jan Axelson) doesn't make any distinctions about any different 'types' of usb 2.0 hubs, it speaks in terms that imply they're referring to ALL usb 2.0 hubs. since this book is my primary reference, if there's something they're leaving out then I'd be ignorant of it as well :P

but, at http://www.everythingusb.com/usb2/faq.htm it says:
"Hi-Speed and Full/Low-Speed USB devices can coexist nicely on USB 2.0 hubs. Connecting such a hub to a USB 2.0 port is recommended." which seems to support the idea.

Reading the actual usb 2.0 specification, it sounds like 2.0 devices attached to the hub along with the 1.x device might have to wait while the low speed transaction is taking place, but it still sounds like they can operate at 2.0 speeds when the 1.x device is not being communicated with. if that is the case, the 1.x device would be slowing the 2.0 devices down when it is communicating, I don't think it would be nearly as severe as actually forcing them to 1.x speeds...

with a usb 2.0 hub, an external usb 2.0 storage device, and any usb 1.x device, it should be pretty easy to test. transfer a large file to/from the storage device without the 1.x device plugged in, and then test it again with it plugged in... i'll try giving it a shot when i'm not at work and have some free time (which is rare

)

**turbocad6** · 08-02-2006, 01:43 PM

I know I've seen a few hubs that specificaly advertized that they can do both simultainiously, & I think I've seen plenty of references to the speed dropping to the lower if any one device is 1.X... I just assumed this was the case, which is why I'm using 2 hubs... maybe that's not necissary..... cool... I gotta test this..

**evandude** · 08-02-2006, 02:17 PM

Okay, I found some info that sorta relates to your mention of different types of hubs, in the wikipedia entry for USB.

hubs have "transaction translators" which handle communication between a low/full speed (1.x) device and a high speed (2.0) bus; cheaper ones have one for the whole hub, better ones have one per port. The difference is that with only one, all the 1.x devices are sharing the same 1.x bandwidth, ie - only one 1.x speed device can communicate with the hub at a time; with one TT per port, all the 1.x devices can communicate with the hub at the same time, and then from there each TT is only competing for high-speed (2.0) bandwidth, which is much greater. basically with one TT per hub, all the 1.x devices would work like they were plugged into a single hub; with one TT per port, they would act like they each had their own dedicated hub, so your 1.x devices would run faster on a TT-per-port hub...

the article still doesnt give much detail about how this pertains to mixing 2.0 and 1.x devices on one hub; since it does mention TT's competing for 2.0 bandwidth in the TT-per-port hubs, that leads me to believe that any 2.0 devices attached to a hub would be competing with the 1 or more TT's for 2.0 bandwidth as well, and thus still run at 2.0 speeds.

the article says:
"The Transaction Translator in a Hi-Speed hub (or possibly each port depending on the electrical design) will function as a completely separate Full Speed bus to Full Speed and Low Speed devices attached to it."
since that says "to full speed and low speed devices attached to it" and does not include high speed devices, that also leads me to believe that high speed devices function independently of the low/full-speed devices running through the transaction translator.

I also just turned up this diagram from a tom's hardware guide article that someone was using to illustrate the opposing view, even though it seems to support my thought:
http://images.tomshardware.com/2003/.../single-tt.gif
as you can see, the usb 2.0 device is NOT passing through the TT, so it is running at full speed and thus only competing for usb 2.0 bandwidth with the output of the TT, which even with 3 or more usb 1.x devices plugged in, would only be a small fraction of the full usb 2.0 bandwidth.
the full article:
http://www.tomshardware.com/2003/09/...ogy/page7.html
in the article they run speed tests with a usb 2.0 hard drive and usb 1.1 webcam plugged into the two different hubs (one with one TT per hub, one with one TT for each port) and the transfer speeds for the usb 2.0 hard drive do not change very much (16.2 MB/s vs 16.6 MB/s)

I think that kind of proves it (although i'd still like to test it out on my own), because 16.2 MB/s is well above the maximum speed of usb 1.x, but that's the speed they were getting with a usb 2.0 HD and usb 1.1 webcam in the same, single TT usb 2.0 hub.