Increasing the speed of dial-up to that of cable -- possible?

[CBFalconer]

Why can't telephone wires carry those high frequencies?

They can. But they have heavy losses. Almost 40 years ago we were
resolving 8 picoseconds down standard telephone pair cables,
unshielded. But we had to put a repeater in every 200 or so feet
of cable.

 

[nobody >]

The answer is no, it can't. Plain old twisted pair bell
wire cannot retain the analog signal at the sufficiently
high frequencies necessary to come anywhere near ISDN, ADSL,
let alone *cable* broadband speeds.

That's funny.. I'd swear that it's plain old twisted pair bell wire that
runs about 6,500 feet from the remote DSLAM to my house. I wonder why
I'm getting 2900/670 down/up. Must be that fancy CAT5 cable that runs
between my computer and the modem that does it. (I always use an
external modem)

Reality is that you can push Gigabit down a two-strand barbed-wire fence
if you have the bucks to spend on the equipment at the ends.

If you had bothered to do some research you would know this.

I recommmend you do some basic, prerequisite learning of
basic core technologies before trying to bite off more than
you can chew. The building blocks of education are not
trying to invent some new thing when you don't even know how
the old things worked.

If you are trying to smarten up poor Radium, you'd better take your own
advice.

 

[Radium]

Because they were optimized for low cost not high
performance, a long time ago.

Is the low speed of analog dial-up internet also due to the
insufficient density, insufficient thickness, and/or excessive
electrical-resistance of the copper telephone wire?

If the wires had their copper atoms more tightly packed, thicker, and
protected against oxidation as well as EMI/RFI, would this assist in
speeding up analog dial-up connections?

Could protection against EMI/RFI be provided by cover the wire with
insulator, covering this insulator with 2nd layer of copper -- to
absorb EMI/RFI - and then covering this 2nd layer of copper with a 2nd
layer of insulation?

 

[kony]

That's funny.. I'd swear that it's plain old twisted pair bell wire that
runs about 6,500 feet from the remote DSLAM to my house.

Reread what I wrote. _Analog_ signal. I think we're all
well aware that ADSL uses telephone wire, and we're also
aware it can't meet cable broadband speed which was one of
the stipulations.

I wonder why
I'm getting 2900/670 down/up. Must be that fancy CAT5 cable that runs
between my computer and the modem that does it. (I always use an
external modem)

Reality is that you can push Gigabit down a two-strand barbed-wire fence
if you have the bucks to spend on the equipment at the ends.

A nice idea but no, you can't. The SNR drops so low there
would be insufficient differentiation to reach high speeds,
no matter how good the equipment. Maybe if you meant a few
inches of wire instead of a telephone line length run, but I
don't see many people stringing only a few inches of barbed
wire as it wouldn't make for much of a "fence".

I don't ask you to agree though, show anyone doing it,
pushing a cable broadband speed signal reliably down a steel
barbed wire run long enough to connect your modem to a
DSLAM. Considering even your TP copper wire can't, unless a
device is specifically made to do this you have no
reasonable expectation it can be done, nor that anyone would
ever even try to create such a device for "barbed wire
broadband".

If you are trying to smarten up poor Radium, you'd better take your own
advice.

Start by reading more carefully. People don't generally
write things without reason (except of course Radium).

 

[kony]

If the wires had their copper atoms more tightly packed, thicker, and
protected against oxidation as well as EMI/RFI, would this assist in
speeding up analog dial-up connections?

Analog dialup is a dead-end tech already, nothing is
fruitful about your train of thought.. If you'd like to go
pack some copper atoms, have fun.

 

[Aaron Leonard]

~
~ > Because they were optimized for low cost not high
~ > performance, a long time ago.
~
~ Is the low speed of analog dial-up internet also due to the
~ insufficient density, insufficient thickness, and/or excessive
~ electrical-resistance of the copper telephone wire?
~
~ If the wires had their copper atoms more tightly packed, thicker, and
~ protected against oxidation as well as EMI/RFI, would this assist in
~ speeding up analog dial-up connections?
~
~ Could protection against EMI/RFI be provided by cover the wire with
~ insulator, covering this insulator with 2nd layer of copper -- to
~ absorb EMI/RFI - and then covering this 2nd layer of copper with a 2nd
~ layer of insulation?

I admit it, you made me laugh.

You can pack your POTS line's copper atoms till they resemble neutron star
material, but you'll never get beyond 64kbps (ignoring data compression effects)
for a single channel call switched thru the public telephone network, because
that network doesn't carry anything but 64kbps.

Aaron

 

[CBFalconer]

.... snip ...

You can pack your POTS line's copper atoms till they resemble
neutron star material, but you'll never get beyond 64kbps
(ignoring data compression effects) for a single channel call
switched thru the public telephone network, because that network
doesn't carry anything but 64kbps.

Actually the POTS network is designed to carry roughly 300-4000 Hz
bandwidth, with certain delay specifications over the range. In
practice, today, that is modulated onto a 56k digital signal, and
transmitted over a 64 k bandwidth channel (the last 8 khz is for
signalling). V90/V92 modems have the ability to grasp that digital
bandwidth and use it. If it isn't there, they are limited to about
28 kHz or less.

That is why distance from the local station is important. POTS is
a mature, well understood technology.

 

[Floyd L. Davidson]

Actually the POTS network is designed to carry roughly 300-4000 Hz

More like 400-2800 Hz, if we assume you mean the PSTN as
a whole. The local loop portion (outside plant cable)
is designed to handle at least 750 KHz. (Both portions
can handle higher bandwidths under common circumstances,
but that is what they are _designed_ for.)

bandwidth, with certain delay specifications over the range. In

There are no delay specifications for the PSTN.

practice, today, that is modulated onto a 56k digital signal, and

In practice there are no 56K digitial signals in the
PSTN, and certainly nothing that is "modulated" at that
rate.

"Modulation" is a term that applies to analog signals,
not to digital signals. Digital signals are "encoded".

Commonly an analog signal is digitized so that it can be
transported over a digital facility. Likewise digital
signals are used to modulate an analog carrier to
provide a means of transport over analog facilities.

That _is_ confusing. ;-)

And of course the PSTN is made up of many combinations
of analog and digital. For example, T1 lines are
digital, as are all channels through virtually all
switching systems. But radio signals, and common
telephone sets, are analog.

A modem is a device to transport a digital signal via an
analog channel, and that is true for the RS-232
connection to a dialup modem exactly the same way it is
for a "digital" radio system (just that radio modems
start at about 1000 times the cost of a dialup modem).

Aaron Leonard is quite correct about 64Kb/s "single
channels". The 64Kb/s level referred to as a "DS0" is
the lowest level in the digital hierarchy that is now
universally used for everything in the PSTN except the
local loop. Aaron is precisely correct that regardless
of whatever bandwith there may or may not be on the
local loop, no single connection through the switching
systems used in the PSTN is going to be greater than
64Kb/s.

transmitted over a 64 k bandwidth channel (the last 8 khz is for
signalling).

Not bandwidth. Bitrate. It is *not* 8 KHz for
signaling, it isn't 8 Kb/s though either. It is up to 8
Kb/s, but might be less. (In most configurations used
today the signaling channels are on separate circuits,
so the voice channels are always fully 64Kb/s.)

V90/V92 modems have the ability to grasp that digital
bandwidth and use it. If it isn't there, they are limited to about
28 kHz or less.

I can't tell what you mean by that, so
I'm not sure...

Whatever, it certainly is *not* 28 KHz. I assume you
meant digital bitrate, and 28 Kb/s. The minimum bitrate
that should ever be seen is 14.4Kb/s.

That is why distance from the local station is important. POTS is
a mature, well understood technology.

Agreed! But obviously not by many...

 

[CBFalconer]

More like 400-2800 Hz, if we assume you mean the PSTN as
a whole. The local loop portion (outside plant cable)
is designed to handle at least 750 KHz. (Both portions
can handle higher bandwidths under common circumstances,
but that is what they are _designed_ for.)


There are no delay specifications for the PSTN.

There are, in one form or another. Any time you control phase, you
are imposing delay vs frequency restrictions.

In general you are quibbling over my choice of verbiage, which,
while not accurate, was intended to carry the general idea to the
masses. The existing cross-links included Microsoft groups, which
allows for a lot of ignorance.

 

[Floyd L. Davidson]

There are, in one form or another. Any time you control phase, you
are imposing delay vs frequency restrictions.

There are *NONE* for the PSTN. None.

Private lines can be conditioned, but that is never done
for anything on the switched network. It is never even
measured.

In general you are quibbling over my choice of verbiage, which,

There was a lot more inaccuracy that just a choice of
words, it was the technical specifics that were wrong.
Examples are the above statements regarding "delay
specifications" and the error in claiming the PSTN is
designed to "carry roughly 300-4000Hz" is significant
when in fact it is designed for 400-2800 Hz (and at the
absolute best will *never* pass a 4000Hz tone).

That sort of error was through out the article, plus
there was the fact that you highly insinuated the Aaron
Leonard's article was incorrect, when in fact he was
precisely correct. The point of course is that I know
from past articles that Aaron is technically quite well
versed about telecom. You have your areas of expertize
too, but telecom isn't one of them.

 

[CBFalconer]

There are *NONE* for the PSTN. None.

Private lines can be conditioned, but that is never done for
anything on the switched network. It is never even measured.

I don't know where you are getting your info, but mine comes from
30+ years ago when developing PABX systems for general use. There
were very definite delay/phase specifications involved. I believe
they originated with Ma Bell, and were adhered to by GTE at the
time. We were a private firm, and (successfully) developed
hardware to meet those specs. I was Chief Engineer.

Your 'no 4000 Hz' no doubt refers to the signalling reservation.

Please translate "PSTN" for me. I don't think we agree.

 

[Floyd L. Davidson]

I don't know where you are getting your info, but mine comes from
30+ years ago when developing PABX systems for general use. There

40 years of working on the PSTN, not PBX's.

were very definite delay/phase specifications involved. I believe

Cite one. (They are *all* specifications for private line
service.)

they originated with Ma Bell, and were adhered to by GTE at the
time. We were a private firm, and (successfully) developed
hardware to meet those specs. I was Chief Engineer.

Your 'no 4000 Hz' no doubt refers to the signalling reservation.

No, 4000 Hz has nothing to do with signaling. It is
*IMPOSSIBLE*, both in theory and in practice, to pass
4000 Hz through a single channel over the switched PSTN.

Please translate "PSTN" for me. I don't think we agree.

It is the common terminology: Public Switched Telephone Network.

 

[CBFalconer]

.... snip ...


No, 4000 Hz has nothing to do with signaling. It is
*IMPOSSIBLE*, both in theory and in practice, to pass
4000 Hz through a single channel over the switched PSTN.

In the good old days 4000 hz was reserved for signalling blasts
over the analog lines, and there were dire retributions due for
anyone who had the temerity to introduce 4kHz onto the lines.
There is no such thing as an analog filter with infinite cut-off.

 

[Ken]

More like 400-2800 Hz,

In Europe 300-3400Hz is the standard.

 

[Floyd L. Davidson]

In the good old days 4000 hz was reserved for signalling blasts
over the analog lines,

That isn't true.

Nobody has *ever* used 4000 Hz for signaling. Let me
give you a clue: You cannot, and have *never* been able
to, pass 4000Hz over any carrier system used for the
PSTN.

The closest thing to that were the Lenkurt analog FDM
systems (such as 45BX carrier) that use "out of band
signaling", at 3700 Hz (the US standard, while CCITT
specified 3825 Hz). In fact, any signaling tones above
3400 Hz are considered "out of band".

and there were dire retributions due for
anyone who had the temerity to introduce 4kHz onto the lines.

Indeed, there *are* (as in there are still the exact same
reasons, and it is forbidden). It has nothing to do with
signaling.

There is no such thing as an analog filter with infinite cut-off.

What's your point though?

There *are* some rather good analog filters. In fact
the way that ADSL works, putting data on top of you
voice line, is by using frequencies above 4000 Hz and
filters. It is the _customer_ that is not allowed to put
anything above 4000 Hz on a POTS loop.

 

[Floyd L. Davidson]

In Europe 300-3400Hz is the standard.

If you check that out, you'll find it isn't. A POTS
connection is only guaranteed to pass 400-2800 Hz. That
is the minimum specification.

Not that individual channels (and even circuits) are
never specified at 300-3400 Hz, because they are.

But if you measure a POTS line and find that nothing above
3000 Hz gets through, it will not mean the line is out of
specs.