From: email@example.com (Floyd Davidson)
Subject: Re: ::::: IS MY ISP TELLING ME THE WHOLE STORY? :::::
Date: 23 Sep 1998 02:45:18 GMT
Steven Uhrig <firstname.lastname@example.org> wrote:
>John M. Hunt wrote:
>> "Lawrence L. Baldwin" <email@example.com> wrote:
>> >The problem is the 3db digital pad that is likely being introduced
>> >inter-office, just as you state.
>> Lawrence, could you tell us why the telephone companies find it
>> necessary to introduce digital pads? I presume that they are simple
>> ROM lookup tables to translate the 8-bit code created by a distant A
>> to D converter to a code that represents the 8-bit sequences that
>> would have been created had the analog input to said A to D converter
>> been a certain number of decibels (say perhaps 3) lower than the level
>> actually converted.
> Pads were part of the telephone network long
>before the invention of digital switching. They
>are there to make the echo that is present in any
>4 wire to 2 wire conversion less noticeable to the
There is a "trick" to that, which is not obvious. A pad placed
in the line will reduce the desired signal by the amount of the
pad's value, but an echo signal will pass through it twice, thus
reducing echo by twice the pad's value. Picking the right pad
values was a serious task prior to digital switching and
transport systems. What was called the Via Net Loss (VNL) plan
was used, and it was an attempt to come close to an optimal
value for most calls.
> The network has used standard pad
>values of 3 DB and 6 DB for decades. The longer
Well... for about 20 years actually. I'm not sure when the
Switched Digital Network (SDN) loss plan was formulated, but
it's implementation would be essentially at the same time as
implementation of digital switching systems. So two decades ago
it was rare, and today the VNL plan is rare.
The essential difference between the two is that SDN is a fixed
loss plan and VNL is a variable loss plan. Virtually all
trunking is the same for digital systems, and virtually every
trunk group has a different loss value for analog systems. The
echo delay times still change from longer to shorter calls with
both types, but there is a great difference between noise
characteristics for long haul analog trunks compared to long
haul digital facilities, which is why the two different loss
>the connection the more susceptible the human ear
>is to echo. So the longer the connection the
>larger the pad value. 6 dB for toll trunks and 3
>DB for short distance EAS and Interoffice toll
Originally there were several combinations of arrangements for
6, 3, and 0 dB padding of various types of trunks. However it
seems that in many instances (certainly we did this here in
Alaska with conscious intent) the plan was simplified, and all
connecting toll trunks have 6 dB of padding, and inter-toll
trunks have 0 dB.
However, Local Exchange Carriers have done strange things...
Many didn't seem to understand the change from VNL to SDN (I'm
not exactly up to speed on it either, so we form a large group!)
and a wide (wierd?) variety of pad values have been found in the
oddest of places for the oddest of reasons.
I do know, for example, of a LEC that had 2.8 dB pads in all of
their EAS trunks. Now just how they came up with that value I
would not know... but what really amazed me was that the very
first person to own a 56K PCM modem and try using it called them
up and asked if they had digital pads, and damned if they didn't
redesign their trunk interface levels to remove them within a
couple days just to be nice! (It made me look kind dumb, cause
the guy told me about it the next day, and I told him that
Hell would freeze over before he ever got a good connection.
And Hell froze the next day!)
> Not all digital switches use digital pads either.
>Some switches accomplish the padding by changing
>the gain of analog amplifiers in the 4 wire
>segment of the subscribers line unit, between the
>codec and the hybrid transformer.
Floyd L. Davidson firstname.lastname@example.org
Ukpeagvik (Barrow, Alaska) email@example.com