why does fcc limit 56k to that limit with a law?

0 views
Skip to first unread message

LovingPerson

unread,
Feb 2, 2004, 1:34:51 PM2/2/04
to
Hi all:

I have been around since the 300 baud days. Slowly, modems went
up in speed. Eventually, to 9600, 14.4 . Then the speed seemed to go
up very fast. It seemed promising when it hit 28.8 quickly followed
by 56k. I was very hopeful that it would continue to go up. However,
it has been many years and that has not happened.

Why is the FCC in collision with the big communications companies
in trying to limit the poor man's high speed? Why would fcc make such
an unjust agency law?

Someone who knows about this, please enlighten me. Am I missing
something? Why isn't the American poor outraged? Are you as outraged
as I am?

sincerley, Moser.

Kirk Strauser

unread,
Feb 2, 2004, 2:15:06 PM2/2/04
to
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

At 2004-02-02T18:34:51Z, sayl...@aol.com (LovingPerson) writes:

> Why is the FCC in collision with the big communications companies in
> trying to limit the poor man's high speed? Why would fcc make such an
> unjust agency law?

Telephone equipment is designed to accept certain power levels. If someone
exceeds those power levels, it could disrupt the phone equipment. If the
phone equipment were disrupted, it could cause service outages. Part of the
FCC's regulatory mission is to keep people from doing things that will knock
out public services, and they do this by passing laws making those acts
illegal.

OK, then. That means that a modem can only transmit at a certain power
level, and no more. The speed limit you're referring to reflects the
maximum amount of data that can be transmitted over telephone lines via an
analog modem at the legal power levels so as to avoid disrupting telephone
service for everyone else.

Does that clarify things, or were you hoping for a more paranoia-fueled
answer?
- --
Kirk Strauser
The Strauser Group
Open. Solutions. Simple.
http://www.strausergroup.com/
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFAHqGv5sRg+Y0CpvERAn7PAJ92GVk42orG3Bv4jJo1lrWB0WWevgCfVSkt
STOuydM0TY8mF6YOFi+/pRY=
=ogoQ
-----END PGP SIGNATURE-----

Robert Redelmeier

unread,
Feb 2, 2004, 2:18:38 PM2/2/04
to
In comp.dcom.xdsl LovingPerson <sayl...@aol.com> wrote:
> I have been around since the 300 baud days. Slowly, modems went

Ditto here, and I've used some 110 too.

> up in speed. Eventually, to 9600, 14.4 . Then the speed seemed to go
> up very fast. It seemed promising when it hit 28.8 quickly followed
> by 56k. I was very hopeful that it would continue to go up. However,
> it has been many years and that has not happened.

Because the phone lines were designed for voice (remember
acoustic couplers?) which requires fairly limited bandwidth,
64 kbit/s max. The Signal-to-Noise ratio combined with a
3750 baud (max) rate doesn't allow more info to pass.

The chief bottleneck is using audible frequencies and
multiplexing for voice transmission trunking. xDSL gets
around this by using higher frequencies on the short local
loops and separate data transmission trunking.

> Why is the FCC in collision with the big communications companies
> in trying to limit the poor man's high speed? Why would fcc make such
> an unjust agency law?

They didn't. It's simply physics. Do you believe that the
FCC or even big comms are competent enough to surpress a market?
You give them _far_ more credit than they deserve and have shown.

-- Robert

Bev A. Kupf

unread,
Feb 2, 2004, 2:28:28 PM2/2/04
to
On Mon, 02 Feb 2004 19:15:06 GMT,
Kirk Strauser (ki...@strauser.com) wrote:
>
> At 2004-02-02T18:34:51Z, sayl...@aol.com (LovingPerson) writes:
>
>> Why is the FCC in collision with the big communications companies in
>> trying to limit the poor man's high speed? Why would fcc make such an
>> unjust agency law?
>
> OK, then. That means that a modem can only transmit at a certain power
> level, and no more. The speed limit you're referring to reflects the
> maximum amount of data that can be transmitted over telephone lines via an
> analog modem at the legal power levels so as to avoid disrupting telephone
> service for everyone else.
>
> Does that clarify things, or were you hoping for a more paranoia-fueled
> answer?

And even 56k works because one end of the connection (telco to service
provider) is all digital, meaning that a single analog to digital signal
conversion is done.

However, if you cannot get a digital line, you can consider using inverse
multiplexing modems (or bonding modems) to achieve 112k or beyond using
analog lines.

--
Bev A. Kupf
"The lyfe so short, the craft so long to lerne" -- Chaucer
Tintin turns 75 <http://www.tintin.com>

B'ichela

unread,
Feb 2, 2004, 3:02:10 PM2/2/04
to
In article <fed77941.04020...@posting.google.com>, LovingPerson wrote:
> Why is the FCC in collision with the big communications companies
> in trying to limit the poor man's high speed? Why would fcc make such
> an unjust agency law?
It has to do with the amount of bandwith needed to go beyound
56K. Yes a law is a dumb idea but.. its done. Even if you had a modem
that COULD go faster than 56K do you realize that most American
Telephone systems Can't even handle 33600bps well. Thats due to the
wiring of the telephone system itself.
As for Poor mans High speed access. Have you checked out
Planet Maca's Shell acounts yet? you get 33600bps but a real Unix
shell where you can download to/from. Plus Email, Free Usenet
Newsgroups and the account does not allow easy spread of trojans and
Virii? No Popups or time limits, you get a 20MB of disk space on the
free access level. Planet Maca's Opus is listed as a Freenet (Free Shell
account ISP) not Peer to Peer. and is dialup accessable with any standard
terminal program? See my signature for access info.

--
From the Desk of the Sysop of:
Planet Maca's Opus, a Free open BBS system.
Telephone 860-738-7176 300-33.6kbps Telnet://pinkrose.net.dhis.org
The New Cnews maintainer
B'ichela

J. Yazel

unread,
Feb 2, 2004, 4:25:12 PM2/2/04
to
On Mon, 02 Feb 2004 19:15:06 GMT, Kirk Strauser <ki...@strauser.com> wrote:

>-----BEGIN PGP SIGNED MESSAGE-----
>Hash: SHA1
>
>At 2004-02-02T18:34:51Z, sayl...@aol.com (LovingPerson) writes:
>
>> Why is the FCC in collision with the big communications companies in
>> trying to limit the poor man's high speed? Why would fcc make such an
>> unjust agency law?
>
>Telephone equipment is designed to accept certain power levels. If someone
>exceeds those power levels, it could disrupt the phone equipment. If the
>phone equipment were disrupted, it could cause service outages. Part of the
>FCC's regulatory mission is to keep people from doing things that will knock
>out public services, and they do this by passing laws making those acts
>illegal.

> ==== SNIP ====

He may be complaining about being mislead by having the product
labelled as 56k capable without any warning that the customer will not
be able to use the "capability".

As an example, most modern automobiles are capable of driving over
100 mph but you see very little advertising that says "buy our 100 mph
automobile!".

He may agree with me that the manufacturers should be required to
mark the product with the useable speed.

Jack

Floyd Davidson

unread,
Feb 2, 2004, 4:31:49 PM2/2/04
to
B'ichela <mda...@pinkrose.net.dhis.org> wrote:
>In article <fed77941.04020...@posting.google.com>, LovingPerson wrote:
>> Why is the FCC in collision with the big communications companies
>> in trying to limit the poor man's high speed? Why would fcc make such
>> an unjust agency law?

The rule is neither unjust, nor is it "in collision with the" whatever.

> It has to do with the amount of bandwith needed to go beyound
>56K. Yes a law is a dumb idea but.. its done. Even if you had a modem

No it is neither a dumb idea, nor does have anything directly to do
with bandwidth.

>that COULD go faster than 56K do you realize that most American
>Telephone systems Can't even handle 33600bps well. Thats due to the
>wiring of the telephone system itself.

Again, that just isn't true. Consider that the standard T1 lines,
running 1.544 Mbps (with a bandwidth of 772 KHz) uses standard
telephone lines.

The limitation that prevents 56K modems from using the same
technology to go faster is a result of /two/ things. One is
power, the other is bandwidth.

As Bev A. Kupf mentioned in her excellent article, the high
speed down link side of a 56 K modem is digital all the way, and
involves exactly one CODEC in the telephone company equipment
(and another one in your modem, for a single pair). Normally
that telco CODEC is used to convert PCM (digitally encoded
voice) back to (quasi-) analog, but with a v.90 modem it is fed
digital data and the Pulse Amplitude Modulated output is used as
a digitial signal rather than as an analog signal. The
/bandwidth/ limitation is that the sampling rate for the CODEC
is 8000 Hz, meaning it cannot use more than 4Khz of bandwidth
for its output.

Hence, the only way to go faster, using that equipment, is to
increase the /power/ level. What the FCC has done (70 some
years ago) is determine what the *maximum power* that any one
telephone line can used without interfering with other telephone
lines using the same cable facility.

Hence it is a very smart specification, and quite just. It
prevents your neighbors' modems from interfering with *your*
voice calls (not to mention your modem connection too)!

It is definitely *not* something you want changed! The
appropriate way to get higher speeds over telephone cables is to
use different equipment at the telco. That is what DSL, ISDN,
and T1's (as well as cable modems, which use different kinds of
cable) do, and that equipment allows higher *bandwidth* (not
more power) to be used over a given cable.

--
Floyd L. Davidson <http://web.newsguy.com/floyd_davidson>
Ukpeagvik (Barrow, Alaska) fl...@barrow.com

Kirk Strauser

unread,
Feb 2, 2004, 5:00:05 PM2/2/04
to
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

At 2004-02-02T21:25:12Z, J. Yazel <jya...@ai5.net> writes:

> He may be complaining about being mislead by having the product
> labelled as 56k capable without any warning that the customer will not
> be able to use the "capability".

I'm pretty sure he was grousing that we don't have 122Kbps modems, not that
he can't fully use his 56K modem.

> As an example, most modern automobiles are capable of driving over 100
> mph but you see very little advertising that says "buy our 100 mph
> automobile!".

I don't think that's a good analogy. Modems are limited to 56K by the
design of the telephone system, not by some arbitrary legal restriction.
You're perfectly free to run your modem as fast as the "road" between your
computer and your ISP will allow.

> He may agree with me that the manufacturers should be required to
> mark the product with the useable speed.

Define "usable speed". When I was a sysadmin at an ISP, some of our
customers routinely connected at 51K+, while others were limited to 26400
(oh, how we hated pair gain). How would you require the manufacturer to
label that?


- --
Kirk Strauser
The Strauser Group
Open. Solutions. Simple.
http://www.strausergroup.com/
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFAHsf65sRg+Y0CpvERAmDxAKCaYYkoXlFhQV9/quk0h2ko1OmkQwCfdu3v
WPXvuGy+ir0cPodGcW7qNg0=
=5xMw
-----END PGP SIGNATURE-----

Bill Williams

unread,
Feb 2, 2004, 6:38:01 PM2/2/04
to
The problem is, we are asking the wrong question. The real question is
not 'What it the most data we can transmit over a wire between two
points?' It is 'What is the most data we can transmit over the dial up
telephone network?'

We can transmit a lot of data between two points. Certainly a few Mbs
over a high speed line. Much more if we are prepared to pay for a
dedicated line. But the ability to dial up anyone in the world, or even
anyone in North America is something else.

The real reason is very simple: the dial up system ( Called the Public
Switched Telephone System or PSTS) is designed to make voice
connections, person to person, on demand. Voice is analog and
historically about 4 kHz bandwidth was enough. As the world went
digital, only the connection from your house to the nearest telephone
office remained analog, all the conneections between offices, and all
the switching which routed your connection to the right destination
became digital, was almost univerally standardized on 8000 samples per
second (to allow nearly 4 kHz analog signals to be represented) and 8
bit samples to provide sufficient signal to noise. So the connection
between telephone offices is basically 64 kbs. In fact some of the
transmission equipment uses one bit per sample for other purposes, so
the most you can rely on is 56 kbs of clear channel. Thats all the bits
there are, you can't do better unless you redesign the whole PSTS.

Technically, your ISP can send a 56 kb signal to you. He sends it as a
digital stream to the telephone company, they forward it to an exchange
near you, they convert it to analog there and send the result to you
just as if it were voice, and your 56 kbs modem can turn it back into
data stream for you. Turns out it doesnt quite work, because the analog
signal created at the last leg is powerful enough at certain frequencies
to create interference with neighboring phones, so it had to be reduced
from 56kbs to 53kbs to eliminate this interference.

Even that you can only do if the sending end (your ISP) has a digital
link to the phone company. For ordinary consumer connections, there is
an analog link at each end, which introduces enough extra distortion
that you can't reliably transmit better than 33 kbs. Its actually
amazing you can do that well.

But, bottom line, there is a hard limit of 56 kbs over the public
switched telephone network, and practical limits rather lower. Its not a
regulatory decision, and its not basically about signal to noise ratios
over the line to your house, its fixed because that is how the PSTS
works.

JM

unread,
Feb 2, 2004, 8:57:43 PM2/2/04
to
quoting:
>It is definitely *not* something you want changed! The
>appropriate way to get higher speeds over telephone cables is to
>use different equipment at the telco. That is what DSL, ISDN,
>and T1's (as well as cable modems, which use different kinds of
>cable) do, and that equipment allows higher *bandwidth* (not
>more power) to be used over a given cable.


Ah yes, ISDN. ISDN is a digital phone line. With an ISDN line, you get a
direct digital connection all the way. You get the full 64kbps, as long as
you're dialing into the ISP's ISDN lines. Duel channel ISDN is 128kbps
becuase you're getting two phone lines, and using the bandwidth of those two
phone lines.

J. Yazel

unread,
Feb 2, 2004, 10:33:41 PM2/2/04
to
>
>> As an example, most modern automobiles are capable of driving over 100
>> mph but you see very little advertising that says "buy our 100 mph
>> automobile!".
>
>I don't think that's a good analogy. Modems are limited to 56K by the
>design of the telephone system, not by some arbitrary legal restriction.
>You're perfectly free to run your modem as fast as the "road" between your
>computer and your ISP will allow.

The analogy still fits. Automobile speeds are limited by the design of
the highway system (and of the automobiles), also not by some arbitrary
legal restriction.
>


>> He may agree with me that the manufacturers should be required to
>> mark the product with the useable speed.
>
>Define "usable speed". When I was a sysadmin at an ISP, some of our
>customers routinely connected at 51K+, while others were limited to 26400
>(oh, how we hated pair gain). How would you require the manufacturer to
>label that?

I was referring to the government established maximum speed. Just as
in road speeds, that it the fastest that you are allowed to go.

Jack

Bev A. Kupf

unread,
Feb 2, 2004, 10:45:32 PM2/2/04
to
On Mon, 02 Feb 2004 22:33:41 -0500,
J Yazel (jya...@ai5.net) wrote:
>>
>>> As an example, most modern automobiles are capable of driving over 100
>>> mph but you see very little advertising that says "buy our 100 mph
>>> automobile!".
>>
>>I don't think that's a good analogy. Modems are limited to 56K by the
>>design of the telephone system, not by some arbitrary legal restriction.
>>You're perfectly free to run your modem as fast as the "road" between your
>>computer and your ISP will allow.
>
> The analogy still fits. Automobile speeds are limited by the design of
> the highway system (and of the automobiles), also not by some arbitrary
> legal restriction.

This discussion is like a broken pencil - pointless. Nevertheless - your
analogy is incorrect. As Kirk and Floyd Davidson have pointed out in their
excellent posts, 56 kbps is _not_ a legislated limit. It is a limit
dictated by telco hardware.

On the other hand, 65 mph (on US highways) is a legislated limit. Existing
highways can tolerate higher speeds, and existing cars can operate at
higher speeds, and I'll wager that several of us here have proven that.

> I was referring to the government established maximum speed. Just as
> in road speeds, that it the fastest that you are allowed to go.

That then is the difference. 56 kbps is the fastest you _can_ go. You
cannot break that barrier with an existing analog telephone line. 65 mph
is the fastest you are _allowed_ to go on an existing highway. You can
break that barrier with existing conditions.

Beverly

J. Yazel

unread,
Feb 3, 2004, 2:49:10 PM2/3/04
to
On Tue, 03 Feb 2004 03:45:32 GMT, "Bev A. Kupf" <beva...@myhome.net> wrote:

>On Mon, 02 Feb 2004 22:33:41 -0500,
> J Yazel (jya...@ai5.net) wrote:
>>>
>>>> As an example, most modern automobiles are capable of driving over 100
>>>> mph but you see very little advertising that says "buy our 100 mph
>>>> automobile!".
>>>
>>>I don't think that's a good analogy. Modems are limited to 56K by the
>>>design of the telephone system, not by some arbitrary legal restriction.
>>>You're perfectly free to run your modem as fast as the "road" between your
>>>computer and your ISP will allow.
>>
>> The analogy still fits. Automobile speeds are limited by the design of
>> the highway system (and of the automobiles), also not by some arbitrary
>> legal restriction.
>
>This discussion is like a broken pencil - pointless. Nevertheless - your
>analogy is incorrect. As Kirk and Floyd Davidson have pointed out in their
>excellent posts, 56 kbps is _not_ a legislated limit. It is a limit
>dictated by telco hardware.

I don't think you understand the subject here. 56k is not a legislated
limit and it is also NOT dictated by telco hardware. I have seen hundreds
of messages stating that the legislated limit on telco lines is less than 56k
because of hardware power limitations.



>
>On the other hand, 65 mph (on US highways) is a legislated limit. Existing
>highways can tolerate higher speeds, and existing cars can operate at
>higher speeds, and I'll wager that several of us here have proven that.
>
>> I was referring to the government established maximum speed. Just as
>> in road speeds, that it the fastest that you are allowed to go.
>
>That then is the difference. 56 kbps is the fastest you _can_ go. You
>cannot break that barrier with an existing analog telephone line. 65 mph
>is the fastest you are _allowed_ to go on an existing highway. You can
>break that barrier with existing conditions.

As far as I know, there is no 56kbps tel. line barrier. The only barrier is
a lower legislated one. Just as on a highway you can drive faster than the
safe legislated speed, you can also send 56k illegally on a tel. line.

Jack


Bev A. Kupf

unread,
Feb 3, 2004, 3:02:17 PM2/3/04
to
On Tue, 03 Feb 2004 14:49:10 -0500,
J Yazel (jya...@ai5.net) wrote:
>>This discussion is like a broken pencil - pointless. Nevertheless - your
>>analogy is incorrect. As Kirk and Floyd Davidson have pointed out in their
>>excellent posts, 56 kbps is _not_ a legislated limit. It is a limit
>>dictated by telco hardware.
>
> I don't think you understand the subject here. 56k is not a legislated
> limit and it is also NOT dictated by telco hardware. I have seen hundreds
> of messages stating that the legislated limit on telco lines is less than 56k
> because of hardware power limitations.

Replace 56 kbps with 53 kbps - mandated limit. The original poster
wondered why analog modems supporting bps rates faster than 56 kbps
didn't exist. That _is_ a limit of existing telco hardware. Ergo
the following still holds true.

>>That then is the difference. 56 kbps is the fastest you _can_ go. You
>>cannot break that barrier with an existing analog telephone line. 65 mph
>>is the fastest you are _allowed_ to go on an existing highway. You can
>>break that barrier with existing conditions.
>
> As far as I know, there is no 56kbps tel. line barrier. The only barrier is
> a lower legislated one. Just as on a highway you can drive faster than the
> safe legislated speed, you can also send 56k illegally on a tel. line.

But no rate higher than that; that which the original poster desired.
Nolens volens, you cannot exceed 56 kbps on existing telco lines with an
analog modem. But you can drive faster than 65 mph (or 75 mph) on a
highway.

Steve Baron - KB3MM

unread,
Feb 3, 2004, 4:49:27 PM2/3/04
to
The really interesting part is that many. many 56K modems can't even
break the 30 K barrier and that ties in with the 53 K limit.

"Bev A. Kupf" <beva...@myhome.net> wrote in message
news:slrnc1vvlo....@myhome.net...

Floyd Davidson

unread,
Feb 3, 2004, 5:22:24 PM2/3/04
to
J. Yazel <jya...@ai5.net> wrote:
>
> I don't think you understand the subject here. 56k is not a legislated
>limit and it is also NOT dictated by telco hardware. I have seen hundreds
>of messages stating that the legislated limit on telco lines is less than 56k
>because of hardware power limitations.

...

> As far as I know, there is no 56kbps tel. line barrier. The only barrier is
>a lower legislated one. Just as on a highway you can drive faster than the
>safe legislated speed, you can also send 56k illegally on a tel. line.

If you will go back and carefully read the entire thread, there
are at least two or three different articles which explain that
in detail.

The absolute limit per line due to the theoretical design of the
Public Switched Telephone Network (PSTN) as it exists today, is
64Kbps. The practical limit, due to hardware installed, is 56
Kbps. The actual limit, due to typical characteristics of
circuits and attempts at designing a modem to work the best it
can on the average line, is 53 Kbps.

The basic increment of bandwidth used for the PSTN is a DS0,
which is 64Kbps. Until that changes (when the PSTN moves from a
"circuit switched" network to a "packet switched" network), the
64 Kbps upper limit is solid. That is simply because telephone
switches switch DS0's, not DS1's or whatever.

As has previously been noted, there are _several_ reason that
equipment currently installed will not actually provide a clear
channel 64Kbps data rate to a customer. That is because there
are several equipment implementations that use up to 8Kbps for
various other purposes, such as testing, signaling, or
reliability. (Worse yet, there are *many* places where
equipment is in use that actually only provisions 32Kbps per
line, not 64!)

The 53 Kbps is actually a modem manufacturing compromise between
what could be done, and what will work best for most users. It
would be quite possible to assume 1) less cable noise and 2)
more precision from the telco CODEC, and on the one-half of one
percent of all telephone lines where those conditions exist, 64
Kbps would be possible! That is to say, that if a precision
CODEC (spelled "expensive") were used in a line card that is
connected to a short (say 1/2 mile or less) high quality cable
in an area with very little external noise and with no other
users on the cable, it is relatively easy to redesign almost
any v.90 modem to achieve 64 Kbps. (Or, forget the telco codec
and the cable, just use two modems back to back with a 6 foot
patch cord between them!)

That modem would not perform well for users with a typical line
though... So, instead of proving it can be done and going
broke, modem designers have provided modems which do as best
they can on the majority of lines, and in the process they've
simply tossed that miniscule number of users that could get
another few Kbps.

Now, where does anything you could call "legislated" fit into
the picture?

Kirk Strauser

unread,
Feb 3, 2004, 5:30:06 PM2/3/04
to
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

At 2004-02-03T21:49:27Z, "Steve Baron - KB3MM" <Steve...@StarLinX.com> writes:

> The really interesting part is that many. many 56K modems can't even break
> the 30 K barrier and that ties in with the 53 K limit.

Example?

I've never seen a modem that couldn't break 33600 *given decent phone lines*.


- --
Kirk Strauser
The Strauser Group
Open. Solutions. Simple.
http://www.strausergroup.com/
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFAICBU5sRg+Y0CpvERAhi2AJ98L7TkxNmN/tmS89yljIVzHmZDJACeOcm8
gufacsPqonS+b89jUFbdqSk=
=nmkd
-----END PGP SIGNATURE-----

Don Bruder

unread,
Feb 3, 2004, 5:53:43 PM2/3/04
to
In article <87r7xbg...@strauser.com>,
Kirk Strauser <ki...@strauser.com> wrote:

>
> At 2004-02-03T21:49:27Z, "Steve Baron - KB3MM" <Steve...@StarLinX.com>
> writes:
>
> > The really interesting part is that many. many 56K modems can't even break
> > the 30 K barrier and that ties in with the 53 K limit.
>
> Example?
>
> I've never seen a modem that couldn't break 33600 *given decent phone lines*.

You've never seen mine.

Oh, wait... you said "given decent phone lines"... Never mind.

(Sitting at the end of nearly 20 miles of mixed copper and glass, and
considering myself lucky when I get a connection above 26400 that holds
for more than 5 minutes at a stretch. People just a mile further up the
road - which translates to about 7 miles up the wire due to the
geography around here - consider it near-miraculous when they get a
connection faster than 9600. They call it a full-blown miracle when a
connection *AT ANY SPEED* holds for longer than about 8 minutes without
getting dropped.)

--
Don Bruder - dak...@sonic.net <--- Preferred Email - SpamAssassinated.
Hate SPAM? See <http://www.spamassassin.org> for some seriously great info.
I will choose a path that's clear: I will choose Free Will! - N. Peart
Fly trap info pages: <http://www.sonic.net/~dakidd/Horses/FlyTrap/index.html>

Floyd Davidson

unread,
Feb 3, 2004, 5:57:40 PM2/3/04
to
J. A. Mc. <jaS...@gbr.online.com> wrote:
>On Tue, 03 Feb 2004 03:45:32 GMT, "Bev A. Kupf" <beva...@myhome.net> found
>these unused words floating about:

>
>>65 mph
>>is the fastest you are _allowed_ to go on an existing highway. You can
>>break that barrier with existing conditions.
>>
>>Beverly
>
>Some areas ... others have up to 75 MPH posted, legal limits.

How does that relate a v.90 modem though? In fact, there is no
posted limit for the modem. Instead, the same thing happens
with the modem that happens on the road: if you go too fast for
the conditions you will seriously corrupt whatever you are
transporting that fast. On the road, your life and your body
are distorted, sometimes beyond recognition or recovery. Same
with data over a telephone line.

The power limits on telephone lines are not some arbitrary
limit. Originally the specifications were designed around the
effects of crosstalk between cable pairs and the intermod
distortion caused when high levels are introduced into Frequency
Domain Modulation (FDM) telephone carrier systems. Later on,
microwave radio systems were designed to match the exact same
limits that cables had, thus if a signal was too high for a
cable, it would also be too high for the carrier and for the
radio.

That design target was also used in the theoretical design of
the digital hierarchy, and implemented in the physical design
of Time Domain Modulation (TDM) digital carrier systems. The
dynamic range of a PCM encoded quasi-analog channel carried by
digital facilities is matched very closely to the dynamic range
1) required for human speech recognition, 2) available from
typical telephone cable designs, 3) available from typical
telephone carrier designs, and 4) available from typical
telephone microwave designs. The *same* upper power level per
channel applies to *all* of them! (Actually TDM has a slightly
lower limit, which is explained below.)

This power limit is simply a matter of if you put two signals
onto different pairs in a cable, they *will* show up on all pairs
in the cable. If the level is high enough, they *will*
interfere with each other. Likewise channels in a FDM carrier
system suffer from inter-modulation if the levels are too high.
Likewise with microwave radios, where FM modulation has what is
called a "knee", which is a point at which loading the system
more no longer results in a linear increase in distortion, but
instead there is an exponential increase in distortion.

The TDM digital carrier systems that are almost universal today
are slightly more sensitive to levels. There are only 255 codes
available with PCM to indicated a voltage level. The lower
levels are right down there with the typical noise on a cable.
The upper level is only 3.7 dB above testtone level! (The upper
level for cable is +5 dB.) If you crank your modem output level
up 10 dB, you *don't* get a 10 dB higher signal at the distant
modem... you only get 3.7 dB louder, and every voltage that
originally went over the limit is grossly distorted. (Your data
will not be recoverable!)

There is no posted, legislated, mandated or whatever speed
limit. It's just a simply a fact of the physical design of the
roadway that you can drive the car around that bend at 35 mph,
but if you try to do it a 75 mph, you'll be rolled up in a ball
somewhere off the road and in one big hurt.

Floyd Davidson

unread,
Feb 3, 2004, 6:38:52 PM2/3/04
to
"Steve Baron - KB3MM" <Steve...@StarLinX.com> wrote:
>The really interesting part is that many. many 56K modems can't even
>break the 30 K barrier and that ties in with the 53 K limit.

How?

Robert Redelmeier

unread,
Feb 3, 2004, 7:05:58 PM2/3/04
to
In comp.dcom.xdsl Kirk Strauser <ki...@strauser.com> wrote:
> I've never seen a modem that couldn't break 33600 *given decent phone lines*.

Mine. I'm 1.5 miles from my RT, and voice quality
is excellent. The phone line is better than decent.
It just won't hold v.90 because there's an additional
D/A and A/D at the CO.

-- Robert

Steve Baron - KB3MM

unread,
Feb 3, 2004, 7:01:47 PM2/3/04
to
Didn't say that very well.

I meant 56 K modems in typical installations with some what typical phone
lines at typical distances from the central office etc.

"Kirk Strauser" <ki...@strauser.com> wrote in message
news:87r7xbg...@strauser.com...

Bruce Tomlin

unread,
Feb 4, 2004, 12:13:17 PM2/4/04
to
In article <fed77941.04020...@posting.google.com>,
sayl...@aol.com (LovingPerson) wrote:

> Hi all:
>
> I have been around since the 300 baud days. Slowly, modems went
> up in speed. Eventually, to 9600, 14.4 . Then the speed seemed to go
> up very fast. It seemed promising when it hit 28.8 quickly followed
> by 56k.

Which is the maximum allowed by law. Law of physics that is. There's
only so much data you can cram into a 4KHz audio signal. Especially
when it gets converted to 56K or 64K bits at the phone company.

> I was very hopeful that it would continue to go up. However,
> it has been many years and that has not happened.

Yes it has. It's called DSL. I've had it for four years now.

I get 6 megabits down and 512K up over my phone line. And I can make
phone calls at the same time, too. Of course in order to get that
speed, the other end of that connection is in a beige box somewhere in
my neighborhood less than a mile away.

Bruce in Alaska

unread,
Feb 4, 2004, 1:06:43 PM2/4/04
to
In article <GHWTb.12791$w36....@newssvr22.news.prodigy.com>,
Robert Redelmeier <red...@ev1.net.invalid> wrote:

This happens a lot more than most folks think. I khow of a few
places where the Northen Telcom DS100 Switch predates the
Introduction of Digital Inputs, and has Analog Frontend inputs.
When these areconnected to Digital DS1 Incoming Trunks they must
be run through a D/A to inoput into the analog input to the switch, then
after switching go through another D/A on the outgoing local Line.
Each D/A and A/d conversion drops the bandwitth a bit and the connection
speed for analog modems suffer because of it.

Bruce in alaska
--
add a <2> before @

Floyd Davidson

unread,
Feb 4, 2004, 4:04:07 PM2/4/04
to
Bruce in Alaska <bru...@btpost.net> wrote:
>In article <GHWTb.12791$w36....@newssvr22.news.prodigy.com>,
> Robert Redelmeier <red...@ev1.net.invalid> wrote:
>
>> In comp.dcom.xdsl Kirk Strauser <ki...@strauser.com> wrote:
>> > I've never seen a modem that couldn't break 33600 *given decent phone
>> > lines*.
>>
>> Mine. I'm 1.5 miles from my RT, and voice quality
>> is excellent. The phone line is better than decent.
>> It just won't hold v.90 because there's an additional
>> D/A and A/D at the CO.
>>
>> -- Robert
>>
>
>This happens a lot more than most folks think. I khow of a few
>places where the Northen Telcom DS100 Switch predates the
>Introduction of Digital Inputs, and has Analog Frontend inputs.

Right idea, but the wrong reasons! The DMS-100 began life with
the ability to use digital trunking. It was a fairly primitive
module called a DSC or DSM or something like that (I haven't
actually seen one in well over a decade), which handled only one
DS1. Later, when they started using 68000 cpu's in the XPM's
(instead of 8085 cpus) they developed a DTC (Digital Trunk
Controller) that could handle something like 20 DS1's. (We're
talking late 70's and very early 80's here, right at 25 years
ago.)

>When these areconnected to Digital DS1 Incoming Trunks they must
>be run through a D/A to inoput into the analog input to the switch, then
>after switching go through another D/A on the outgoing local Line.
>Each D/A and A/d conversion drops the bandwitth a bit and the connection
>speed for analog modems suffer because of it.

The problem is that remote units, whether they are Nortel or
Lucent, or whatever, get installed with one of two types of
interface. The choice of which to use greatly affects how well
v.34 and v.90 modems will work, but that is rarely ever a
consideration when the engineering is done. What is considered
are the number of available line cards in the terminating
switch. If, as is common, the switch has a surplus (since
adding remote units might well free up hundreds or even
thousands of line cards in the switch as those lines are moved
to line cards in the remote units), the engineers (or their
Pointy Haired Bosses) do not want to spend extra money to buy
new digital interface equipment (relatively expensive devices)
when they can instead install significantly less expensive
channel banks filled with cards that convert channels analog
lines that utilize those hundreds or thousands of analog line
cards already installed in the switch.

That the "Universal Interface" (which uses the analog line
interface on the switch) is also independent of the type of
switch, and can be used with either a 5E from Lucent or a
DMS-100 from Nortel, for example.

The "Integrated Interface", which is what we all want to find
has been used on our own line, is a switch specific module and
also requires a switch specific software upgrade. Hence it can
be very expensive. On the other hand, for a new installation,
or where an old switch is physically crowding its available
space, the "Integrated" way takes up less space and is a much
cleaner, less maintenance prone, method.

That can make for some seemingly odd situations! Big switching
offices are more likely to have the room and to have it filled
with unused line cards... and therefore often end up with
inferior service as far as a modem user is concerned. Little
dinky places where someone tried to avoid spending much money to
begin with, may not have the physical space, or the hardware, to
use a Universal Interface! For all the wrong reasons (poor long
term planning), the big operation gets the cheesy lines and the
two-bit operation has the satisfied customers with a smile.

Floyd Davidson

unread,
Feb 4, 2004, 11:15:53 PM2/4/04
to
Jimmy <NoOn...@Nowhere.Net> wrote:
>On Wed, 04 Feb 2004 12:04:07 -0900, Floyd Davidson <fl...@barrow.com>
>wrote:
>
>I've got a better question about the FCC:
>
>How come they allow network TV to have people humping on screen,
>allow commercials that opening discuss erections during the Superbowl,
>and allow shows like Friends where every second line is a sexual
>innuendo - but got all bent out of shape because Janet Jackson
>showed off her new tit(s)?

I didn't watch the Superbowl. I don't own a TV. I haven't
watched 30 minutes of TV in the past month or so.

I've never seen people "humping on screen", commercials during
the Superbowl, and have no idea what "Friends" even is!

Perhaps that does explain what I think of the FCC and Television.
Newton Minow was the guy who got it right... "... a vast wasteland".
And geeze, that was nearly a quarter of a century ago! Nothing has
changed...

Bev A. Kupf

unread,
Feb 4, 2004, 11:33:00 PM2/4/04
to
On Thu, 05 Feb 2004 02:26:13 GMT,
Jimmy (NoOn...@Nowhere.Net) wrote:
> How come
> [ deleted ]

> got all bent out of shape because Janet Jackson
> showed off her new tit(s)?

I didn't watch the superbowl. BBC news has an interesting
article titled "US apoplexy over Jackson flash" that you might
care to read. Here's the URL:
<http://news.bbc.co.uk/2/hi/entertainment/3456497.stm>

It was a relief to know that the indignation has nothing to
do with Ms. Jackson being a bit of minger.

David C.

unread,
Feb 5, 2004, 12:49:42 AM2/5/04
to
sayl...@aol.com (LovingPerson) writes:
>
> I have been around since the 300 baud days. Slowly, modems
> went up in speed. Eventually, to 9600, 14.4 . Then the speed
> seemed to go up very fast. It seemed promising when it hit 28.8
> quickly followed by 56k. I was very hopeful that it would continue

> to go up. However, it has been many years and that has not
> happened.
>
> Why is the FCC in collision with the big communications
> companies in trying to limit the poor man's high speed? Why would
> fcc make such an unjust agency law?
>
> Someone who knows about this, please enlighten me. Am I
> missing something? Why isn't the American poor outraged? Are you
> as outraged as I am?

First off, there is no way an analog phone line will ever go to more
than 56K without all the phone companies replacing millions of
dollars worth of equipment. And that's not going to happen because
the existing equipment is perfectly good for carrying voice calls -
which is what the voice network is meant for.

As for why, the phone network digitizes audio at a sampling rate of 8K
bytes/sec, using 7-bit u-law encoding. This is 56K bits per second.
Any encoding that tries to pack more bits onto the analog wire will
simply see data loss as soon as the analog signal hits the telco's
central office - where it will be digitized at 56K for transmission
through the network.

As for why the FCC puts an artificial limit at 53K instead of
allowing a full 56K, that's because large parts of the country are
using old wires. In order to hit the full 56K, you have to transmit
at power levels that can interfere with other lines on those older
wire bundles.

The FCC couldn't care less about your modem speed. They care quite a
bit about if your voice calls (and that's what any analog modem
connection is) is going to interfere with your neighbor's voice calls.

If you think losing 3Kbps out of a theoretical maximum speed of 56K is
something to get outraged about, then you should look into anger
management classes.

-- David

Floyd Davidson

unread,
Feb 5, 2004, 4:18:14 AM2/5/04
to
sha...@techie.com (David C.) wrote:
>
>First off, there is no way an analog phone line will ever go to more
>than 56K without all the phone companies replacing millions of
>dollars worth of equipment. And that's not going to happen because
>the existing equipment is perfectly good for carrying voice calls -
>which is what the voice network is meant for.

Welllll... many, if not most, are offering xDSL service of one
kind or another. Hence it seems to be that indeed they *are*
replacing millions of dollars worth of equipment!

>As for why, the phone network digitizes audio at a sampling rate of 8K
>bytes/sec, using 7-bit u-law encoding. This is 56K bits per second.

Except, that isn't what they do. It is an 8KHz sampling rate,
but it is an 8 bit u-law encoding. That is 64 Kbps for a DS0
rate.

>Any encoding that tries to pack more bits onto the analog wire will
>simply see data loss as soon as the analog signal hits the telco's
>central office - where it will be digitized at 56K for transmission
>through the network.

Unless, of course, it is IP traffic that is routed via the
Internet. In which case the rate depends on what kind of xDSL
service was used on the local loop.

>As for why the FCC puts an artificial limit at 53K instead of

The FCC does *not* put any artificial limit at 53Kbps. Anyone
who wishes to encode the data so that it will pass 56Kbps is
perfectly able to do so.

>allowing a full 56K, that's because large parts of the country are
>using old wires. In order to hit the full 56K, you have to transmit

It has nothing to do with "old wires". The same specification
for power levels applies to the most recent cable designs. That
is *not* artificial, it is not arbitrary either. The cable is
literally designed to provide a specific level of crosstalk when
loaded with the specific power levels that the FCC has set as
the standard.

>at power levels that can interfere with other lines on those older
>wire bundles.

That is not really true. It *is* possible to encode the data
for not just 56Kbps, but for 64Kbps! But... it would work in so
few instances that it is not worth the compromises that are
necessary. The compromise involves using codes that require
less noise, and as a result of those compromises the modem would
get a slower connection on the average line. That is a trade
off that none of us who get less than maximum speed would like
to see.

>The FCC couldn't care less about your modem speed. They care quite a
>bit about if your voice calls (and that's what any analog modem
>connection is) is going to interfere with your neighbor's voice calls.

v.90 modems are not an "analog modem" in the downlink direction.
The power level design target for telco cable is *not* unique to
voice calls. It applies to all digital services (T1's, ISDN, xDSL,
etc. etc.)

It is simply wrong to say that the FCC could care less about your
modem speed. The power specifications are what prevents somebody
from interfering with *your modem call*! (Just imagine someone
connecting their Hi Fi amp to the telephone line...)

>If you think losing 3Kbps out of a theoretical maximum speed of 56K is
>something to get outraged about, then you should look into anger
>management classes.

That is an *exceedingly* good point!

Kirk Strauser

unread,
Feb 5, 2004, 10:15:11 AM2/5/04
to
At 2004-02-05T04:15:53Z, Floyd Davidson <fl...@barrow.com> writes:

> I've never seen people "humping on screen", commercials during the
> Superbowl, and have no idea what "Friends" even is!

Ahh, I think I read something about you recently:

http://www.theonion.com/onion3604/doesnt_own_television.html

(Smile - it's funny.)
--
Kirk Strauser

William Moss

unread,
Feb 5, 2004, 10:49:42 AM2/5/04
to
Flod,

Good post. I'm just correcting a few facts about early DMS-100 history
below.

"Floyd Davidson" <fl...@barrow.com> wrote in message
news:87smhqo...@barrow.com...


> Bruce in Alaska <bru...@btpost.net> wrote:
> >In article <GHWTb.12791$w36....@newssvr22.news.prodigy.com>,
> > Robert Redelmeier <red...@ev1.net.invalid> wrote:
> >
> >> In comp.dcom.xdsl Kirk Strauser <ki...@strauser.com> wrote:
> >> > I've never seen a modem that couldn't break 33600 *given decent phone
> >> > lines*.
> >>
> >> Mine. I'm 1.5 miles from my RT, and voice quality
> >> is excellent. The phone line is better than decent.
> >> It just won't hold v.90 because there's an additional
> >> D/A and A/D at the CO.
> >>
> >> -- Robert
> >>
> >
> >This happens a lot more than most folks think. I khow of a few
> >places where the Northen Telcom DS100 Switch predates the
> >Introduction of Digital Inputs, and has Analog Frontend inputs.
>
> Right idea, but the wrong reasons! The DMS-100 began life with
> the ability to use digital trunking. It was a fairly primitive
> module called a DSC or DSM or something like that (I haven't
> actually seen one in well over a decade), which handled only one
> DS1.

No, the original T1 interface to the DMS100 was called the DCM.
(Digital Carrier Module)
It handled 5 T1s. The processor was an 8085.

> Later, when they started using 68000 cpu's in the XPM's
> (instead of 8085 cpus) they developed a DTC (Digital Trunk
> Controller) that could handle something like 20 DS1's. (We're
> talking late 70's and very early 80's here, right at 25 years
> ago.)

The DTCs came out in the mid 80's, and used 6800 (not 68000) processors. The
68000 processors were used for the SuperNode core
(CM, MS, ENET, etc. processors).

There were early switched that installed analog Trunk Module units.
These handled up to 30 analog trunk interfaces, in addtiion to service
circuits. In some cases, the analog transmission carriers they connected
may have been replaced by digital T1s, and connected by channel banks.

I suppose if someone didn't want to but an RLM, they would extend analog
lines through a channel bank, but that would seem like an expensive way to
go for residential lines because there would be no concentration.

William
who used to design this stuff in a previous life

Bruce in Alaska

unread,
Feb 5, 2004, 1:21:12 PM2/5/04
to
In article <87smhp9...@barrow.com>,
Floyd Davidson <fl...@barrow.com> wrote:

> Welllll... many, if not most, are offering xDSL service of one
> kind or another. Hence it seems to be that indeed they *are*
> replacing millions of dollars worth of equipment!

Actually Floyd, the DSlames go in the copper before the switch frontend
and the telco doesn't replace anything, they just add the DSlames.

Floyd Davidson

unread,
Feb 5, 2004, 6:29:20 PM2/5/04
to
Bruce in Alaska <bru...@btpost.net> wrote:
> Floyd Davidson <fl...@barrow.com> wrote:
>
>> Welllll... many, if not most, are offering xDSL service of one
>> kind or another. Hence it seems to be that indeed they *are*
>> replacing millions of dollars worth of equipment!
>
>Actually Floyd, the DSlames go in the copper before the switch frontend
>and the telco doesn't replace anything, they just add the DSlames.

Actually, they replace the interface between the local loop and
the switch.

You can quibble that the interface replaced was often nothing but
a crossconnect on an MDF, but in many cases it is significantly
more than that if any kind of subscriber carrier or remote system
was originally in place.

The point is that telco's have to expend millions of dollars on
equipment to remain competitive (and the ones who do it wrong, go
broke and/or are gobbled up by those with better vision).

Floyd Davidson

unread,
Feb 5, 2004, 7:18:00 PM2/5/04
to
"William Moss" <willia...@hotmail.com> wrote:
>Flod,
>
>Good post. I'm just correcting a few facts about early DMS-100 history
>below.
>
>"Floyd Davidson" <fl...@barrow.com> wrote:
>> Bruce in Alaska <bru...@btpost.net> wrote:

...


>> >This happens a lot more than most folks think. I khow of a few
>> >places where the Northen Telcom DS100 Switch predates the
>> >Introduction of Digital Inputs, and has Analog Frontend inputs.
>>
>> Right idea, but the wrong reasons! The DMS-100 began life with
>> the ability to use digital trunking. It was a fairly primitive
>> module called a DSC or DSM or something like that (I haven't
>> actually seen one in well over a decade), which handled only one
>> DS1.
>
>No, the original T1 interface to the DMS100 was called the DCM.
>(Digital Carrier Module)
>It handled 5 T1s. The processor was an 8085.

Yes! That's the one! (Sorry, my memory is fogged on that one.)

The 5 DS1 interfaces are reformatted into 4 of the DS-30
interfaces used internally in the DMS. A DS1 has 24 channels,
and a DS-30 has 30 channels.

DTC's have 10 each DS1 interface cards, each handling 2 each
DS1's, for a total of 20 DS1's (480 trunks). They have 16 DS-30
interfaces to the internal side (connecting to the Network
Module).

>> Later, when they started using 68000 cpu's in the XPM's
>> (instead of 8085 cpus) they developed a DTC (Digital Trunk
>> Controller) that could handle something like 20 DS1's. (We're
>> talking late 70's and very early 80's here, right at 25 years
>> ago.)
>
>The DTCs came out in the mid 80's, and used 6800 (not 68000) processors. The
>68000 processors were used for the SuperNode core
>(CM, MS, ENET, etc. processors).

68000 processors in the XPMs. The SuperNode core used 68020's
and 68030's. The XPM's definitely had 68000's. And the XPM
software is written in an off the wall "XPM Pascal". DTC's
became available in 1984, and were one of the first XPM's
(eXperimental Peripheral Module). It is interesting to compare
the message passing DMS operating system, SOS, with the original
QNX that was developed almost next door in Ottawa, Ontario. It
is also interesting to speculate how much the Apple Lisa
influenced XPM software development... And while that is all
speculation, there is a much clearer connection between the
IBM format for tape files and the simply atrocious file system
that NTI put on their hard disks.

The use of 68000's in the XPM's was really interesting, because
while it provided the ability to do all kinds fancy things,
there was a huge price to pay to get to that functionality. The
original PM's with 8-bit 8085 processors could only address 64K
of RAM, and a software load over the original DMS bus took about
two minutes to accomplish and get the PM back on line. But a
68000 needed far more than 64K of RAM to do the fancy things it
could do. Initially they had to restrict them to something like
384K of RAM, simply because of the time it took to load the
software. With 2 68000's needing 6 times the memory, it took
excessive time to do a reload.

They found ways to increase the load speed (giving it more bus
time and doing it from a hard disk rather than a tape, for
example), and fairly soon were able to hike the RAM up to 768K
(I don't remember if that particular high water mark required a
hardware change or not). But it didn't really fly until the DMS
bus was made both wider and faster in a series of changes that
took place over about a decade of time. The last I knew, about
10 years ago, each cpu in a DTC was using 2Mb of RAM. By now
they may have twice as many cpu's and ten times the RAM in each!

During that evolution period NTI had to maintain support for a
multitude of hardware configurations. But they also had a
policy for just how far back they would support a software load,
which in practice forced customers to upgrade both software
and hardware on a regular basis. (That point became moot about
the time the equal access came along, because everyone had to
practically double the size of their software load to meet the
new FCC requirements.)

NTI also maintained entirely sepparate facilities (at Richardson
Texas, far from other NTI people) to support the military's
Autovon system. Dealing with Raleigh and Richardson were to
entirely different experiences!

>There were early switched that installed analog Trunk Module units.
>These handled up to 30 analog trunk interfaces, in addtiion to service
>circuits. In some cases, the analog transmission carriers they connected
>may have been replaced by digital T1s, and connected by channel banks.

TM-8's are the analog E&M trunk modules. They are regular PM's,
using 8085's.

Whether a switch was installed with DCM's or TM-8's was an
engineering decision. Both were available. Of course, in many
cases the existing trunking was not being changed, only the
switch. If analog trunking was currently in use, then analog
trunk interfaces were used. The alternative, which was not
uncommon either, was to install the switch with digital
interfaces and add analog to digital conversion equipment between
analog transport systems and the switch. There were units
available that converted, for example, two 12 channel L carrier
groups to a single DS1, which then crossconnected to the DTC
at DSX1 levels.

In 1980 I was involved in a situation where the company was
about to install a digital microwave and the USAF was about to
install a DMS-100/200, each replacing old analog equipment; but
they just couldn't or wouldn't trust the other to actually do
what they said they were about to do (historically, they had
good reason to mistrust each other). So both of them designed
for the other's existing analog equipment. The result was a
very nice digital microwave and a very nice digital switch,
using brand spanking new analog interfaces between them!

I found it infuriating!

>I suppose if someone didn't want to but an RLM, they would extend analog
>lines through a channel bank, but that would seem like an expensive way to
>go for residential lines because there would be no concentration.

I've seen that done in practice. Of course, the situation
wasn't quite what anyone would expect. At one time there were
telephones in every pump station on the Trans Alaska Pipeline
fitting that description...

Here are a couple of articles (the 2nd and 3rd of the three at
this URL) that I wrote on this topic back in the early 90's.

http://yarchive.net/phone/distributed_switching.html

David C.

unread,
Feb 8, 2004, 8:24:28 PM2/8/04
to
Floyd Davidson <fl...@barrow.com> writes:
> sha...@techie.com (David C.) wrote:
> >
> >First off, there is no way an analog phone line will ever go to more
> >than 56K without all the phone companies replacing millions of
> >dollars worth of equipment. And that's not going to happen because
> >the existing equipment is perfectly good for carrying voice calls -
> >which is what the voice network is meant for.
>
> Welllll... many, if not most, are offering xDSL service of one
> kind or another. Hence it seems to be that indeed they *are*
> replacing millions of dollars worth of equipment!

I should've been more specific here. I was talking about voice
lines.

I realize that the local loop of a DSL line is also analog, but it's
irrelevant to the discussion here.

>> As for why, the phone network digitizes audio at a sampling rate of
>> 8K bytes/sec, using 7-bit u-law encoding. This is 56K bits per
>> second.
>
> Except, that isn't what they do. It is an 8KHz sampling rate, but
> it is an 8 bit u-law encoding. That is 64 Kbps for a DS0 rate.

If a 64Kbps rate was being delivered all the way to your home, then
you would be able to receive data at that rate.

You must have something wrong here.

I realize that a DS0 is 64K, but the D-A conversion doesn't use all
the bits. The 8th bits were historically used for signaling. Even
though modern equipment doesn't do this anymore, they can't just
start using these bits, because not every telco around the world has
been upgraded.

>> Any encoding that tries to pack more bits onto the analog wire will
>> simply see data loss as soon as the analog signal hits the telco's
>> central office - where it will be digitized at 56K for transmission
>> through the network.
>
> Unless, of course, it is IP traffic that is routed via the Internet.
> In which case the rate depends on what kind of xDSL service was used
> on the local loop.

Again, irrelevant. I'm talking about a voice line.

>> As for why the FCC puts an artificial limit at 53K instead of
>
> The FCC does *not* put any artificial limit at 53Kbps. Anyone who
> wishes to encode the data so that it will pass 56Kbps is perfectly
> able to do so.

Then why hasn't anyone done it yet? Absolutely every modem
manufacturer says that it's illegal for them to do so. If they're
all lying, maybe you can explain why nobody want's to be the first
out the door with true 56K capacity.

>> The FCC couldn't care less about your modem speed. They care quite
>> a bit about if your voice calls (and that's what any analog modem
>> connection is) is going to interfere with your neighbor's voice
>> calls.
>
> v.90 modems are not an "analog modem" in the downlink direction.

Oh really? When did someone go and replace half of my voice line?

I am aware that the ISP-end of a modem call is a digital connection
to the phone company. Same as if I make a voice call to someone with
a voice phone on an ISDN line. But it doesn't mean that my call is
any less analog on my end.

Perhaps you should get out of "telco engineer buzzword" mode and try
to make your point without using language that does nothing but
confuse everybody who doesn't work for phone companies.

> It is simply wrong to say that the FCC could care less about your
> modem speed. The power specifications are what prevents somebody
> from interfering with *your modem call*! (Just imagine someone
> connecting their Hi Fi amp to the telephone line...)

Now you're contradicting yourself in one sentence.

First you say that the FCC does care about your data rates, and then
you say that their specs have nothing to do with data rates (because
they apply to lines at all kinds of different rates.)

So which is it?

You may know what you're talking about here, but from where I'm
sitting, I'm just seeing a whole mess of buzzwords with no
explanations that us mere mortals can understand. If you're trying
to explain something to this newsgroup, you aren't succeeding.

-- David

Floyd Davidson

unread,
Feb 8, 2004, 10:51:36 PM2/8/04
to
sha...@techie.com (David C.) wrote:
>Floyd Davidson <fl...@barrow.com> writes:
>> sha...@techie.com (David C.) wrote:
>> >
>> >First off, there is no way an analog phone line will ever go to more
>> >than 56K without all the phone companies replacing millions of
>> >dollars worth of equipment. And that's not going to happen because
>> >the existing equipment is perfectly good for carrying voice calls -
>> >which is what the voice network is meant for.
>>
>> Welllll... many, if not most, are offering xDSL service of one
>> kind or another. Hence it seems to be that indeed they *are*
>> replacing millions of dollars worth of equipment!
>
>I should've been more specific here. I was talking about voice
>lines.
>
>I realize that the local loop of a DSL line is also analog, but it's
>irrelevant to the discussion here.

It *is* relevant to what you stated. The telcos *are*
installing millions of dollars worth of equipment to facilitate
extra bandwidth for data. You said "that's not going to
happen".

Clearly it is happening, and has been for some time.

>>> As for why, the phone network digitizes audio at a sampling rate of
>>> 8K bytes/sec, using 7-bit u-law encoding. This is 56K bits per
>>> second.
>>
>> Except, that isn't what they do. It is an 8KHz sampling rate, but
>> it is an 8 bit u-law encoding. That is 64 Kbps for a DS0 rate.
>
>If a 64Kbps rate was being delivered all the way to your home, then
>you would be able to receive data at that rate.

ISDN delivers two DS0's to your door step over the same pair of
wires you say can't get even on DS0. Clearly you *are* able to
receive data at that rate. Just order up ISDN service and the
telco will exchange your v.90 modem for an ISDN modem. The same
cable pair will be used...

>You must have something wrong here.

Your lack of understanding is not evidence that I'm wrong.

>I realize that a DS0 is 64K, but the D-A conversion doesn't use all
>the bits.

The D-A conversion *does* use all 8 bits. (Actually, it uses at
least 11 bits (and maybe as many as 14) internally to produce
the u-law encoding, and outputs 8 bits on the digital side).

The A-D/D-A conversion that takes place in the CODEC on a line
card has no connection whatever to things like Robbed Bit
signaling.

> The 8th bits were historically used for signaling. Even
>though modern equipment doesn't do this anymore, they can't just
>start using these bits, because not every telco around the world has
>been upgraded.

That has no effect whatever on the local loop or the line card
CODEC. It only affects trunking. In fact virtually all telco's
converted to SS7 signaling ten years ago now, and today there
are relatively few instances where anyone encounters Robbed Bit
signaling. Unfortunately, those few instances could in fact
affect a typical connection to an ISP because those connections
would not be long distance but might be inter-office trunking
between adjacent telco CO's, and might not be using SS7 trunks.
Regardless, while Robbed Bit signaling is something to be aware
of, it doesn't affect the majority of calls to a local ISP.

Whatever, they *can* "just start using those bits", because they
have always been using them except on specific links that used
Robbed Bit signaling. You have to go back to equipment that is
more than 25 years old to find anything that cannot be run 8 bit
clean. There is very little to none of it left in service.

>>> Any encoding that tries to pack more bits onto the analog wire will
>>> simply see data loss as soon as the analog signal hits the telco's
>>> central office - where it will be digitized at 56K for transmission
>>> through the network.
>>
>> Unless, of course, it is IP traffic that is routed via the Internet.
>> In which case the rate depends on what kind of xDSL service was used
>> on the local loop.
>
>Again, irrelevant. I'm talking about a voice line.

It *is* relevant to what you have said. There is no reason
that we cannot "pack more bits onto the analog wire" *if* we
merely change the interface at the telco CO to route IP traffic
over higher speed packet switched data links rather than over
circuit switched data links. And as we see, that *is* what
the telco's are doing.

That is in fact exactly what xDSL service does.

Your statement "where it will be digitized at 56K" is just *not*
necessarily true. It is the same identical wire line loop used
for your telephone set, and it *does* "pack more bits onto the
analog wire".

>>> As for why the FCC puts an artificial limit at 53K instead of
>>
>> The FCC does *not* put any artificial limit at 53Kbps. Anyone who
>> wishes to encode the data so that it will pass 56Kbps is perfectly
>> able to do so.
>
>Then why hasn't anyone done it yet? Absolutely every modem
>manufacturer says that it's illegal for them to do so. If they're
>all lying, maybe you can explain why nobody want's to be the first
>out the door with true 56K capacity.

You aren't paying close enough attention to either what I've
said or to what the modem companies have said.

Rockwell ran tests on modems that could and did get 64 Kbps
before they initially came out with a "56K" modem. As noted,
that particular modem would get that particular data rate on a
*very* short loop. But the same modem doesn't do as well over
longer loops as the ones that are optimized for the higher noise
level of a longer loop.

It would not be good sense to go for optimization that gives
1/10th of 1 percent of all users a 56-64K connection, and
compromises the rest by getting 5-8k slower connections over
normal lines.

>>> The FCC couldn't care less about your modem speed. They care quite
>>> a bit about if your voice calls (and that's what any analog modem
>>> connection is) is going to interfere with your neighbor's voice
>>> calls.
>>
>> v.90 modems are not an "analog modem" in the downlink direction.
>
>Oh really? When did someone go and replace half of my voice line?

Yes really. Do a few searches with google. It has been
explained in various newsgroups any number of times in the past.

The telco line card CODEC D-A conversion outputs a *digital*
signal known as Pulse Amplitude Modulation (PAM) that a regular
telephone set decodes as a "quasi" analog signal which you can
hear quite well. But the data sent to your modem never was an
analog signal. It was digital data right from the start, and
the v.90 modem doesn't use it as a quasi analog signal. It
looks at it with a (14 bit ??) CODEC that converts each voltage
level to a specific 8 bit data byte. A classic *digital* link!

>I am aware that the ISP-end of a modem call is a digital connection
>to the phone company. Same as if I make a voice call to someone with
>a voice phone on an ISDN line. But it doesn't mean that my call is
>any less analog on my end.

We aren't talking about a voice call. The downlink side of
a v.90 connection is digital from the ISP all the way to your
modem.

>Perhaps you should get out of "telco engineer buzzword" mode and try
>to make your point without using language that does nothing but
>confuse everybody who doesn't work for phone companies.

Perhaps you should pay attention to what I'm saying and avoid
snotty responses. You clearly have mouthed off on a subject
that you do not understand, nor do are you even able understand
when it is explained.

If you don't understand part of what I said, all you have to do
is ask for further explanation and you'll get it. You certainly
do not expect me or anyone else to answer you by writing an entire
textbook on the subject, do you?

>> It is simply wrong to say that the FCC could care less about your
>> modem speed. The power specifications are what prevents somebody
>> from interfering with *your modem call*! (Just imagine someone
>> connecting their Hi Fi amp to the telephone line...)
>
>Now you're contradicting yourself in one sentence.
>
>First you say that the FCC does care about your data rates, and then
>you say that their specs have nothing to do with data rates (because

You seem to have read that backwards. I said that they *do*
care about your modem speed, and that is exactly why they have
regulations which prevent *me* from interfering with *your*
modem connection.

>they apply to lines at all kinds of different rates.)

Eh? That is correct, but that isn't what I said, nor did I
say that was the reason for anything. I said the reason is
so that I don't interfere with *your* modem.

>So which is it?

If you can't read simple English, no wonder you can't understand
a subject that is this technical.

Let me repeat, it is simply *wrong* to say that the FCC could
care less about your modem speed. You said that, and you were
wrong.

>You may know what you're talking about here, but from where I'm
>sitting, I'm just seeing a whole mess of buzzwords with no
>explanations that us mere mortals can understand. If you're trying
>to explain something to this newsgroup, you aren't succeeding.

Your inability to understand is neither evidence that others
can't, nor that my discussion is inadequate, much less wrong.

Kirk Strauser

unread,
Feb 9, 2004, 11:20:08 AM2/9/04
to
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

At 2004-02-09T01:24:28Z, sha...@techie.com (David C.) writes:

> You may know what you're talking about here, but from where I'm sitting,
> I'm just seeing a whole mess of buzzwords with no explanations that us
> mere mortals can understand. If you're trying to explain something to
> this newsgroup, you aren't succeeding.

Got Google? Start looking for definitions. It's not reasonable to expect
to engage in a technical conversation without being willing to learn the
language. Would you ask a mathematical question and get upset when someone
mentions an integral?

As for me, I'm going back to comp.sys.mac.hardware.misc, where I originally
found this thread. Thanks for the interesting conversation, folks - it was
enlightening.


- --
Kirk Strauser
The Strauser Group
Open. Solutions. Simple.
http://www.strausergroup.com/
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFAJ7Ki5sRg+Y0CpvERAoNjAJ4n4TtIqo1KuaMJHQ9Qf7aycKDjjACfZrzV
xqmHJQ+MPwRFAk+o4UZeAOs=
=mqrj
-----END PGP SIGNATURE-----

Stephen Satchell

unread,
Feb 15, 2004, 9:59:45 AM2/15/04
to
On Mon, 02 Feb 2004 10:34:51 -0800, LovingPerson wrote:

> I have been around since the 300 baud days. Slowly, modems went
> up in speed. Eventually, to 9600, 14.4 . Then the speed seemed to go
> up very fast. It seemed promising when it hit 28.8 quickly followed
> by 56k. I was very hopeful that it would continue to go up. However,
> it has been many years and that has not happened.
>
> Why is the FCC in collision with the big communications companies
> in trying to limit the poor man's high speed? Why would fcc make such
> an unjust agency law?
>
> Someone who knows about this, please enlighten me. Am I missing
> something? Why isn't the American poor outraged? Are you as outraged
> as I am?

Actually, the technical upper limit for voltage-based encoding is 64
kilobits/s, not 56 kilobits/s. The actual speed you get is lower for a
number of reasons.

1) There is a limit on the amount of power a component can place on the
line. This is to prevent interference with other services carried in the
same cable as yours. To stay under the power cap, established early in
the AT&T break-up when 47 CFR 68 was developed, modem designers had to use
some pretty convoluted techniques to keep within the mandated limits.

(N.B.: There is an IEEE Standard for wideband transmission testing -- I
have forgotten the number -- that uses 23 tones to measure a number of
characteristics very, very quickly. The people who developed the standard
had to jiggle the tone power and phase relationships to also stay under
the power cap, so that line testing wouldn't exceed the power limits.
That's sending 23 discreete tones, not data, so it's nothing against data.)

2) As the years go on, much of the nasty practices of the telcos to get
as much out of the transmission system with as low a cost as possible get
weeded out. Because of robbed-bit signaling, the guaranteed digital path
over long distances is only 56 kbits per second, not 64 kbits/s. Portions
of V.92 try to "guess" at which bits are being robbed and so increase the
channel capacity a bit, but this is a game that lives at the point of
diminishing returns.

3) You can get interference from other instruments that share your phone
line. Cordless telephones and certain answering machines can do all sorts
of nasty things to the signal, to the point that the maximum rate falls to
under 30 kilobits/s even on a "perfect" line. When I was doing modem
testing, I had one answering machine that would drive most Rockwell modems
crazy.

4) ADSL can do interesting things, too. For one modem, I needed two of
the microfilters in series before the modem would behave and get a decent
speed.

5) I once worked on a dialtone-over-coax system that couldn't maintain
*any* modem connection, including fax connection. I discovered in
researching the capabilities of the line (using software DSP, by the way)
that the timebase for the coax-based phone was so unstable that any glitch
in the time standard used would cause the error factors throughout the
coax network to wave like a big flag in a stiff breeze -- and the modems
HATED it. With long-term phase jitter exceeding 60 degrees, anything
faster than 300 bits per second was hopeless.

Boy, those were the days. I miss 'em.

Stephen Satchell
Former modem tester
Now a system administrator at a Web hosting facility.


--
Opinions in this posting not necessarily opinion of employer. Are
personal notes, posted using my computer, my account. Not legal opinion
or advice. Void where prohibited. For entertainment only. Your mileage
may vary. Not appropriate for children.

Reply all
Reply to author
Forward
0 new messages