USB curosity

[Tim]

Allegedly, on or about 15 December 2015, Fred Smith sent:
> there is definitely a limit on USB cable length, and for some 
> devices its very strict. I have a Canon LIDE 210 USB scanner here.
> I had originally put together cables long enough to reach from the
> shelf by my desk to the computer, and it totally failed to work when
> using that cabling. it works only if I hold it on my lap (no room on
> the desk) so that the USB cable is no more than 4-5 feet long.
> 
> I've used much longer USB cables with a Brother laser printer, but
> I was probably pushing my luck there, since the standard does
> specify rigid cable length limits. 

There's specifications for 1.5 megabit/sec (low-speed) of 3 metres, and
12 megabit/sec (full-speed) of 5 metres for USB 1.1.  And USB 2.0
specifies 480 megabits/sec (hi-speed) of 5 metres.  Apparently USB 3.0
doesn't specify a maximum cable length, but the practicality of
signalling (*) parameters places it at 3 metres.

* It's not just data going down a wire, there's handshaking, as well.
The timing of data and handshaking response is critical.  Introduce too
much delay (due to cabling length, or its electrical characteristics),
and every transmission becomes mistimed.

Those are the full end-to-end cabling lengths, including the wiring
inside the device from one device's electronic circuits to the other
device's circuitry.  Hence a 5 metre connecting cable actually exceeds
the specifications.

There are ways to correctly exceed these lengths between equipment
that's placed too far apart, and that's to have cables with electronics
inside them (probably in the connectors).  It may be possible to simply
use a hub in the middle, but I haven't researched that.

It's worth noting that some cabling will be horribly inferior.  If you
split apart cheap cabling (perhaps the one supplied with your
equipment), you may find very thin wiring that simply runs in parallel
to each other inside the sheath.  The data pair of wires should be
twisted around each other, to minimise the effects of outside
interference to the data signals.  So replacing cabling *may* be enough
to resolve a problem.

You may find that the header cabling in some desktop PCs (from
motherboard to front-panel connectors) is inferior.  I had one that just
used ribbon cable.

-- 
[tim at localhost ~]$ uname -rsvp
Linux 3.9.10-100.fc17.x86_64 #1 SMP Sun Jul 14 01:31:27 UTC 2013 x86_64

All mail to my mailbox is automatically deleted, there is no point
trying to privately email me, I will only read messages posted to the
public lists.

Lucky for you I typed this, you'd never be able to read my handwriting.