Rewiring USB device to use USB2.0 instead of USB1.1, need a little help

[Stalin828]

I have a keyboard that has a built in USB hub. However, that hub is basically useless because it is USB 1.1. I also have a USB 2.0 hub that I pulled from the scrap heap at work. My intention is to use that to modify my keyboard hub (or replace it) to make it USB 2.0.

My main question is this, I know that USB cables are shielded and the data lines are twisted pair to prevent crosstalk and noise; however, when rewiring inside the keyboard I will need to run some 8 inch 18-22 gauge stranded wires to get from the current USB controller to the new USB controller. (Should I go the replacement route, I was also tinkering with the idea of just seeing if I could replace the USB controller chip with the new USB 2.0 controller chip, but my gut instinct is that wouldn't work, also I tried that and it didn't work ).) Currently, inside the keyboard it is running the data lines from the USB controller to the USB ports which are about 8 inches away without using twisted pair or shielding the cables (they appear to be 18 gauge stranded cables). However, I am under the impression that for USB 2.0 to work as "high speed" that 8 inches of unshielded, non twisted line wouldn't work. Granted my understanding is from one class I took that discussed this (when I was a freshman, a few years ago) on the way to earning my computer science degree (I just graduated no job yet though ).

So, any of you gurus know how much of an impact 8 inches of unshielded, non-twisted cable will have on a USB ports ability to run as a full high speed port?

My other consideration was I could chop up an old USB cable and jack the shielding from it and then use that and some heat shrink tubing to make my own shielded twisted pair cable for just the data lines.

 

[Pommie]

You have a computer science degree? You're kidding me, right? Which university did you graduate from?

Mike.

 

[blueroomelectronics]

Wow. Wonder which University?
USB uses the same cable and pinouts for all versions. It would be the USB IC that determines the speed.

 

[Blueteeth]

There are many differences between USB1.1 and 2.0. Most think just speed (1.5/12, vs 12/480) but its mostly the higher level protocolthat is different. I have used a USB 1.1 cable for high speed 2.0...thats not the problem. The differences are in hard-coded hardware. As billy said, its the USB device chip which determines the version.

 

[Stalin828]

Wow. Wonder which University?
USB uses the same cable and pinouts for all versions. It would be the USB IC that determines the speed.

Yeah I know, I even said that in my original post. I tried just swapping the USB chip and it didn't work. So I am rewiring the thing.

I also don't appreciate the implication that my university sucks. I only elected to take one class that discussed things like crosstalk and interference and how it effects data transmission. Instead I opted to take extra assembly classes and classes on reverse engineering and classes on OpenGL programming, etc. My consideration was they would be more useful for me as I would be programming not running wire.

I just play around with electronics in my spare time for fun. And I don't know crap about them. I learn by accidentally breaking my crap then trying to fix it.

My original question was, I know that in order to run at USB high speed your USB cables need to be properly shielded and the data lines twisted pair, how long of a run of unshielded, non-twisted cable will have an effect on data transmission? The shielding does matter though. Old USB 1.1 cables are unshielded most of the time whereas new USB 2.0 cables are almost always shielded. Try doing a data transfer on a old unshielded USB 1.1 cable, the speed will be atrocious. I have a 3 foot cable here and I get about 1/3 of the speed through it as I do through my shielded 2.0 cable.

Thanks for the help anyway. I will just experiment with it and cannibalize shielding from a spare USB cable if I have any problems. I don't think I will, but figured there was no harm in asking people that know more about it then me.

 

[Sceadwian]

Umm.. I hate to state the obvious but, why don't you just twist the wire by hand and use foil to shield it? Also what you're doing will not make the keyboard work with USB 2.0 and because the keyboard itself is USB 1.0 any USB 2.0 device that needs high speed will NOT get it because the USB 1.0 device will be taking up a lot of time slots. You'd be surprised how much bandwidth something like a keyboard or mouse can use on USB 1.0 connections because every charactor is sent within an entire USB frame, same with mouse position requests and some mice send these several hundreds of times a second. You should probably relgate ALL of your USB 1.0 devices to one root hub and only use USB 2.0 devices on a seperate root hub to avoid traffic congestion. What you're doign is simply put a bad idea, if you even get it to work.

 

[Sceadwian]

I wouldn't laugh too hard about Stalkin having a computer science degree either, they don't typically concentrate too much no the hardcore hardware side of things.

 

[Stalin828]

Umm.. I hate to state the obvious but, why don't you just twist the wire by hand and use foil to shield it? Also what you're doing will not make the keyboard work with USB 2.0 and because the keyboard itself is USB 1.0 any USB 2.0 device that needs high speed will NOT get it because the USB 1.0 device will be taking up a lot of time slots. You'd be surprised how much bandwidth something like a keyboard or mouse can use on USB 1.0 connections because every charactor is sent within an entire USB frame, same with mouse position requests and some mice send these several hundreds of times a second. You should probably relgate ALL of your USB 1.0 devices to one root hub and only use USB 2.0 devices on a seperate root hub to avoid traffic congestion. What you're doign is simply put a bad idea, if you even get it to work.

The keyboard has two USB ends on it. And they go through different USB controllers. One dedicated to the keyboard, one dedicated to the hub. However, it appears that they ground both ground wires together into the same trace on the PCB. Also, they appear to put both 5V lines together into the same trace. Interestingly enough the D+ and D- lines for the keyboard are both run into the same PCB trace as well (it appeared I was interrupted by a customer at this point and haven't had a chance to look again). To avoid messing with the keyboards power, my plan was to grab the two data lines and reroute them to my hub and pull the ground and power from where they exit the keyboard PCB on the way to the hub and wire those to my hub.

Aluminum foil around the wires is adequate shielding? I guess that is basically what they use in coax cables, can't believe I didn't think of that. Thanks for the tip.

 

[smanches]

Computer Science degrees almost never talk about EE design. They rarely go into PC architecture from a hardware standpoint most of the time, and then it's just glossed over. Uni's these days are only in it for the throughput, not the quality.

Go figure, just like everything else.

 

[Stalin828]

Computer Science degrees almost never talk about EE design. They rarely go into PC architecture from a hardware standpoint most of the time, and then it's just glossed over. Uni's these days are only in it for the throughput, not the quality.

Go figure, just like everything else.

They offered courses in hardware architecture. They only required one though. Once you got to junior and senior level courses you could choose more advanced programming concepts or more advanced hardware concepts. I chose the programming stuff.

 

[Sceadwian]

However, it appears that they ground both ground wires together into the same trace on the PCB. Also, they appear to put both 5V lines together into the same trace.

The D+ and the D- lines are common traces as well? Doesn't matter then that they're separate connectors the devices are effectively on the same USB bus. If you have heavy duty tinfoil it's probably thicker than the shielding on a coax cable...

What you have is effectively the USB hub and keyboard itself connected via a passive USB hub (the common traces) The reason for the active hub (the rest of the circuit and the USB hub chip) is for nothing more than buffering. Basically it reads the signals and copys it very simply to every other connector, but prevents degradation from too many passive USB connections. So the hub chip itself isn't very sophisticated, it's just a signal conditioner.