180 phase ambiguity

Eric Jacobsen <eric.jacobsen@ieee.org> wrote:
 
> Usually there's a framing system (or other similar mechanism) with
> Unique Words that will reveal the proper data polarity.
 
> Many FEC codes are linear, so the inverse of a codeword is also a
> codeword, and a Viterbi decoder will properly decode an encoded data
> stream or its inverse.  So the Unique Words (or whatever mechanism is
> used) can be either coded or uncoded, it's a system/architecture
> decision.

10baseT ethernet, like coaxial ethernet uses polarity sensitive 
Manchester coding. It is, however, easy for transceivers to detect
the polarity and correct it, with the added advantage of another
LED and marketing point. (Or maybe without the LED, but still with
the marketing point.)
 
> Using differential coding gets around all of that.

As well as I understand it, ethernet could have done that for 
10baseT (UTP ethernet), but keeping the same coding as coaxial
seemed to be an advantage.

-- glen

On Mon, 12 Nov 2012 07:44:43 +0000 (UTC), glen herrmannsfeldt
<gah@ugcs.caltech.edu> wrote:

>Eric Jacobsen <eric.jacobsen@ieee.org> wrote:
> 
>> Usually there's a framing system (or other similar mechanism) with
>> Unique Words that will reveal the proper data polarity.
> 
>> Many FEC codes are linear, so the inverse of a codeword is also a
>> codeword, and a Viterbi decoder will properly decode an encoded data
>> stream or its inverse.  So the Unique Words (or whatever mechanism is
>> used) can be either coded or uncoded, it's a system/architecture
>> decision.
>
>10baseT ethernet, like coaxial ethernet uses polarity sensitive 
>Manchester coding. It is, however, easy for transceivers to detect
>the polarity and correct it, with the added advantage of another
>LED and marketing point. (Or maybe without the LED, but still with
>the marketing point.)
> 
>> Using differential coding gets around all of that.
>
>As well as I understand it, ethernet could have done that for 
>10baseT (UTP ethernet), but keeping the same coding as coaxial
>seemed to be an advantage.
>
>-- glen

There's a downside to differential coding in that it costs a little
bit of link margin in AWGN, so you give up a little bit of potential
performance to use it.   The performance loss is due to the "error
doubling" effect from a symbol error creating an error for the current
symbol as well as making the reference for the next symbol erroneous.

So it's generally not used in a lot of systems for that reason, and
many systems need a framing structure for other reasons, so the UWs
for polarity ambiguity resolution kind of come for free in that case.


Eric Jacobsen
Anchor Hill Communications
http://www.anchorhill.com