|
Posted by Dan Lanciani on July 11, 2008, 6:54 pm
If you were Registered and logged in, you could reply and use other advanced thread options
kludge@panix.com (Scott Dorsey) wrote:
|A T-1 circuit transmits 1.54 Mbps, which in a perfect world with zero
|channel noise requires 1.54 MHz bandwidth to transmit. In reality it's
|happy with a bit more bandwidth.
I would have thought that as the noise approaches zero (and the S/N ratio
approaches infinity) the required bandwidth approaches zero. 1.54Mb/s
shouldn't require anything like 1.54MHz of bandwidth until the S/N ratio
approaches unity, and that's an incredibly noisy channel.
Dan Lanciani
ddl@danlan.*com
|
 
| |
Posted by Rob Warnock on July 13, 2008, 11:46 am
If you were Registered and logged in, you could reply and use other advanced thread options
+---------------
| kludge@panix.com (Scott Dorsey) wrote:
| |A T-1 circuit transmits 1.54 Mbps, which in a perfect world with zero
| |channel noise requires 1.54 MHz bandwidth to transmit. In reality it's
| |happy with a bit more bandwidth.
|
| I would have thought that as the noise approaches zero (and the S/N
| ratio approaches infinity) the required bandwidth approaches zero.
+---------------
Yes, but... You have to change your encoding to use more distinct signal
levels per Baud, and almost all modems were (and almost all still are)
designed with *fixed* D/A encoding (and AA/D decoding) schemes. One notable
exception was the Telebit trailblazer, which chopped the ~3 KHz analog
phone line up into many narrow subcarriers, and could adaptively choose
each sub carrier's modulation among 1, 2, 4, or 6 bits/Baud, depending
on current S/N and IM distortion. It was the exception to the rule.
+---------------
| 1.54Mb/s shouldn't require anything like 1.54MHz of bandwidth until
| the S/N ratio approaches unity, and that's an incredibly noisy channel.
+---------------
But T-1 didn't use an adaptive modulation system -- it used a single,
*fixed* transmission method, AMI (Alternate Mark Invert), which is
effectively no more efficient than NRZ binary.[1] Thus the only thing
that happens when the S/N ratio gets better/worse is that the *error rate*
gets better/worse -- the spectral efficiency doesn't change.
-Rob
[1] Well, except that AMI allows an average D.C. disparity of zero,
which makes it nice for transformer coupling. But to keep the
clock density high enough for good sync you have to either use
"bit-robbing" on the LSB, or use B8ZS encoding [which came later
in the evolution of T-1, when "clear-channel 64"-kb was introduced].
(And, yes, I know that bit-robbing was also used for E&M until
CCIS came in.)
-----
627 26th Avenue <URL:http://rpw3.org/>
San Mateo, CA 94403 (650)572-2607
***** Moderator's Note *****
<Minor Quible>
The ones density requirement for AMI is 1 in 13 bits, and the "stress
test" for such a line is 1-in-13 to this day. Although robbed-bit
signalling _sometimes_ provided greater ones density, it could fall
short on a busy carrier with most of the trunks in service, since the
low-order bit was robbed only during idle state, and when the trunk
was seized the low-order bit was part of the voiceband enconding.
This probably seems like a difference that makes no difference,
because a trunk that's in use would be expected to produce a fair
number of ones just from modulation. However, pauses between sentences
are surprisingly long, and of course only one side of a conversation
is talking at once anyway, so two adjacent channels with a fairly
"quiet" conversation going on could produce more than 13 zero signals
in a row. It was also very common to have E&M channel units which were
carrying SF signalling, and that could mean a "all zero" state during
pauses between outpulsing, even when the connection wasn't completed
yet. More to the point, it was common practice to "busy out" trunks by
forcing them into seizure, which also sometimes resulted in an all zero
state.
ESF changed the signalling away from robbed-bit, which made B8ZS
necessary, and that solved the problem. The only remaining issue is
one of semantics: even experienced techs and engineers are prone to
refer to a DS1 span as "1.536" Mb/s, even though B8ZS obviated the
need for the 1-in-193 framing pulse that was used on the original D-1
banks, so a "data pipe" DS1 is really running at the full 1.544 Mb/s
rate, and throughput can actually exceed that by a very small amount,
due to the shortening of the words which B8ZS entails.
</Minor Quible>
Bill Horne
Temporary Moderator
Please put [Telecom] at the end of your subject line, or I may never
see your post! Thanks!
We have a new address for email submissions: telecomdigestmoderator
atsign telecom-digest.org. This is only for those who submit posts via
email: if you use a newsreader or a web interface to contribute to the
digest, you don't need to change anything.
|
| Similar Threads | Posted | | Touch Tone at 1964 World's Fair [telecom] | July 7, 2008, 6:19 pm |
| Re: Touch Tone at 1964 World's Fair [telecom] | July 11, 2008, 11:15 pm |
| Re: Touch Tone at 1964 World's Fair [telecom] | July 16, 2008, 2:53 pm |
| Panel (was Re: Touch Tone at 1964 World's Fair) [Telecom] | July 11, 2008, 9:54 am |
| Auditing telephone bills [Re: PicturePhone at 1964 World's Fair] [Telecom] | July 15, 2008, 6:59 pm |
| Point-to-point microwave links (was 1964 World's Fair) [Telecom] | July 24, 2008, 4:48 pm |
| RE: Point-to-point microwave links (was 1964 World's Fair) [Telecom] | July 26, 2008, 6:49 am |
| Re: Point-to-point microwave links (was 1964 World's Fair) [Telecom] | July 28, 2008, 8:19 pm |
| History -- introduction of Touch Tone in Independent Companies? [telecom] | August 6, 2008, 8:17 pm |
| Re: History -- introduction of Touch Tone in Independent Companies? [telecom] | August 7, 2008, 4:01 pm |
|
|
![Re: Touch Tone at 1964 World's Fair [telecom]](/images/header.jpg)
Yahoo! 
Google 
Windows Live 
del.icio.us 
digg 
Netscape 
XML site map