wtm@neoucom.UUCP (Bill Mayhew) (10/11/87)
Hello,
My house is wired with that skinny gray 4 wire stuff (I think Ma
Bell calls it D Station Wire or some such thing). I also have two
phone lines. One line was installed in 1930 (you should see the
arncane thing it has for a protector block!), the other lead was
installed in 1977. All the wiring in the house carries both lines.
I often have modems going on both lines, and I don't see any
interaction. About the only problem is when somebody picks up an
extension on the same line as an active modem.
In many places where there aren't many free pairs, the telco uses
duplexer coils. What they do is use a balanced transmission line
to place two conversations on one pair and use the ground as the
return path. All is well as long as the mess stays in balance. In
my case, since one of my lines was ancient, the telco opted to make
the second line completely new. If you bug the telco enough, you
might be able to get them to give you dedicated lines, but you'll
probably have to talk to quite a few telco personell who insist it
isn't their problem first. The telco, in general, isn't terribly
sympathetic to residential lines used for modems. They only
guarantee least common denominator service of being able to hear
and talk to a human on the other end.
In some places, i think, the term "phantom circuit" is used to
refer to duplexed lines. (Anybody know if the terms are
equivaltent?) Also, sometimes loading coils are used for line
equalization and can produce abberant modem behavior if things get
out of adjustment.
I just got through running some tests with a trellis modulated high
speed modem, a Telebit Trailblazer in fact. I was surprised at how
much better it did on crummy lines than regular Hayes modems. Its
performance was superior even when talking to a standard 212 1200
baud modem on the other end. There is a line at the office that I
simply can not use from home via a Hayes because the continuous
stream of ~r{_i stuff prevents even logging in. Even in 1200 bps
mode without nmp, I used that line with the Telebit and never got
any noise. Perhaps the more sophisticated Viterbi algorithm used
in the Telebit's detector was able to sort out the apparent phase
jitter on that line. Naturally, you pays for what ya get. The
suggested price for the Telebit is >$1K. (There is currently a
price incentive for registered UUCP sites on the Telebit, but I am
unsure of the details. If you're interested, you should talk to
Telebit-- I don't have any info.)
Hope this is useful,
Bill
(wtm@neoucom.UUCP)larry@kitty.UUCP (Larry Lippman) (10/13/87)
In article <728@neoucom.UUCP>, wtm@neoucom.UUCP (Bill Mayhew) writes: > My house is wired with that skinny gray 4 wire stuff (I think Ma > Bell calls it D Station Wire or some such thing). I also have two > phone lines. One line was installed in 1930 (you should see the > arcane thing it has for a protector block!), the other lead was > installed in 1977. A helpful hint: some older telephone protector blocks may be the cause of noise on telephone lines. As an example, the 98-type protector block has a circular cap in its center; unscrewing the cap reveals carbon blocks which slide out for replacement. Over the years carbon may flake from these slide-out carbons and cause high resistance faults to ground; these faults are generally not low enough in resistance to interfere with dialing or ringing, but they can cause intermittent noise. In addition, protector blocks containing long, tubular fuses can introduce noise as corrosion occurs on the fuse contacts and on the internal solder joints which connect to the fuse wire. If you have noise on your telephone line, and the telephone company fails to repair the problem, take a look at the protector block. If you have one of these older protector blocks, and a noise problem, suggest that the telephone company replace it; they probably won't object, since it paid for itself a long time ago, anyhow. :-) > In many places where there aren't many free pairs, the telco uses > duplexer coils. What they do is use a balanced transmission line > to place two conversations on one pair and use the ground as the > return path. All is well as long as the mess stays in balance. In > my case, since one of my lines was ancient, the telco opted to make > the second line completely new. No "duplexer" coils are used to create more pairs for subscriber telephone lines. Creating a "simplex" circuit to ground is used for SX signaling in some leased line applications, but is never used for talking due to the excessive amount of noise and crosstalk to which such an arrangement is susceptible. Simply stated, there is no way in which the loop signaling and ringing requirements of a dial subscriber line can be met through the use of center-tapped repeating coils (i.e., transformers) to create additional "phantom" pairs. TWO "physical" pairs equipped with suitably connected repeating coils can create a THIRD "phantom" pair, but there can be no DC signaling passed on the physical pairs. The phantom pair can have DC signaling, however. But none of this is practicable for normal dial subscriber lines. The only remaining use of a phantom pair is to handle the DX signaling/E&M signaling or loop signaling requirements of 4-wire circuits used as interoffice trunks, PABX tielines, or foreign exchange lines. What IS used to derive a second dial subscriber line from a given dial subscriber line is a technique called AML (Added Main Line) carrier. AML creates an additional line by means of a transmitter and receiver operating at frequencies between 26 kHz and 100 kHz. This transmitter and receiver use separate frequencies and operate full-duplex. At the central office end, a bandpass filter is used to prevent RF energy from radiating back into the CO apparatus for the physical line; the necessary transmitter and receiver fits on a single printed circuit card. The CO transmitter is idle unless the AML subscriber line is in use. At the "physical" pair subscriber location, a passive lowpass filter is used to prevent attenuation of the RF frequencies. At the AML subscriber location, a companion receiver and transmitter is used, with the transmitter operating only when the AML telephone set is off-hook. A small DC current is drained from the "physical" subscriber line pair when it is on-hook; this DC power is fed to a DC-DC converter and is used to trickle-charge a nicad battery. This battery provides power for the AML apparatus and telephone set. Sometimes the telephone company will provide an auxiliary AC-line transformer to assist in charging the nicad battery, since AML works on the assumption that neither the AML subscriber not the physical pair subscriber will talk excessively (thereby causing battery drain and/or lack of charging). In AML operation, the net result is that the physical subscriber line operates independently of the AML subscriber line, and both can be talking at the same time without interference. Multichannel AML is available which can provide up to six AML subscriber lines from one physical pair. With a multichannel arrangement, there is no physical pair subscriber, but instead the physical pair is used to supply a higher amount of DC operating power, in addition to carrying the RF frequencies. The major manufacturer of AML apparatus is Continental Telephone & Electronics. AML works fine for voice communication using a standard telephone set. Transmission of data over an AML circuit is awful at best. AML circuits are not designed with the precision that goes into interoffice FDM carrier circuits. AML is basically a cheap and dirty product which competes with the cost of having a telephone company run new cables. AML has its place - but it ain't for data transmission! Also, many electronic telephone sets, dialers, modems and other telephone accessories will not operate on an AML circuit because the on-hook voltage is between 6 and 12 volts (rather than the conventional 48 volts), and the off-hook voltage is even less. There is also no momentary DC current open on an AML line to indicate that a called-party has disconnected. > If you bug the telco enough, you > might be able to get them to give you dedicated lines, but you'll > probably have to talk to quite a few telco personell who insist it > isn't their problem first. While an operating telephone company is not obligated to give you a physical pair, they will do so if you: (1) bug 'em enough with complaints; or (2) pay for a business line. The potential usage of a business line rules out AML. > The telco, in general, isn't terribly > sympathetic to residential lines used for modems. They only > guarantee least common denominator service of being able to hear > and talk to a human on the other end. This is quite true. The biggest problem with data communication over dial subscriber lines is noise. Such noise is usually the result of wet cable, splices or terminals in the telephone company outside plant. Locating the cause of noise on a specific line is a very time-consuming process, and generally requires two telephone company craftspersons. Bear in mind that I am referring to noise which can interfere with data communications; such noise is usually more subtle and is more difficult to locate that noise of a magnitiude which interferes with voice communication. The effective cost to an operating telephone company of tracking down and repairing a noise condition on a residential line could be a hundred dollars or more. This cost may well exceed the yearly revenue from a residential line. So, this is why operating telephone companies are not thrilled about expending this effort on residential lines. Whether you agree with them or not, the operating telephone company does have a point. An operating telephone company has only a finite amount of outside repair force, and they try to set priorities for repair of trouble - with the residential modem subscriber last. Part of the problem plaguing operating telephone companies are the number of modem noise complaints - which even when diligently investigated - turn out NTF (No Trouble Found). > In some places, i think, the term "phantom circuit" is used to > refer to duplexed lines. (Anybody know if the terms are > equivalent?) It is a correct term, but it has become somewhat archaic since for all intents and purposes, phantom circuits are no longer used for voice - only for signaling. So it is more common to refer to a phantom circuit on a 4-wire line as simply the DX circuit - which refers to the type of signaling used. > Also, sometimes loading coils are used for line > equalization and can produce abberant modem behavior if things get > out of adjustment. Loading is used to compensate for the effects of distributed capcitance in telephone cables. There are no "adjustments" for loading; loading coils, which come in various inductances, are installed at a known interval which is determined by the capacitance and gauge of cable. As an example, H-88 loading using 88 mH coils is installed every 6,000 feet on cable with a capacitance of 0.083 uf/mile. Loading has the effect of creating a distributed low-pass filter on a telephone pair. The above example of H-88 loading results in a maximum frequency of 3,500 Hz; in a long loop, the "knee" is pretty sharp on a frequency-attenuation plot. Loading coils are not usually a cause of noise or other trouble on a telephone line. Loading coils are toroidal in construction, so stray magnetic radiation is minimum, resulting in little likelihood of crosstalk being caused by adjacant loading coils in a loading coil case. WATER, in one form or another, is the single most prevalent cause of noise on telephone lines. <> Larry Lippman @ Recognition Research Corp., Clarence, New York <> UUCP: {allegra|ames|boulder|decvax|rutgers|watmath}!sunybcs!kitty!larry <> VOICE: 716/688-1231 {hplabs|ihnp4|mtune|seismo|utzoo}!/ <> FAX: 716/741-9635 {G1,G2,G3 modes} "Have you hugged your cat today?"