Two phones off hook: Hello...are you THERE?

That one happens to burn me up personally, because beginning in
the late 80's I was advocating ISDN for the Alaskan network.  It
was literally years later that *anyone* else understood.  We had
a VP of marketing in the early 90's that had caught on.  I knew
he was aware, but he certainly didn't indicate to me any
specifics of what he was discussing with others at his level.
But it is also true that I once literally chewed on another VP
in his presence about not working to get that plus Internet
connectivity.  The guy didn't say a word.

But six months later he resigned and went to work for Wilcom,
which at the time appeared to be one of the more progressive
American companies.  I took it that he just simply got fed up
with being part of a Dilbert Cartoon...

Because the cost of labor here is high, we were digitalized far
sooner than most of the US.  When the country as a whole was
about 30% digital, Alaska was virtually 100%.  And of course
that was the end of pulse dial phones.

But that was 20 years ago!  That isn't an example of "not
recently manufactured", that's an example so old the
instructions are only found on the walls of local pyramids,
using strange symbols that nobody understands anyway.

As long as ten years ago virtually all CPE was using hybrid
transformers, and that was probably still true even 5 years ago.
Just when it might have changed I'm not sure.  (I'm retired and
have only looked inside one telephone, which was an old one
anyway, in years.)

The telephone sitting next to me is about 12 year old...

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

f @barrow.com (Floyd L. Davidson) writes:

Hey! I resemble that remark...

The Touch-A-Matic 32 at hand has a 1982 datestamp,,, and the only
phones in the house without a hybrid transformer is a 7507....yep
the ISDN set designed, I'm told, for testing CO generics.
--
A host is a host from coast to coast.................wb8 @nrk.com
& no one will talk to a host that's close........[v].(301) 56-LINUX
Unless the host (that isn't close)........................­.pob 1433
is busy, hung or dead..........................­..........20915-1433

How would "sharp drop" indicate one or the other?  A 3 dB drop
can be correctly described as a "sharp drop", if it is either
noticeably fast or severe in effect.  I thought the defining
statement was that the "caller can hardly hear us", which
indicates that they can indeed be heard.

YMMV.

The same was true in the US until 1984.  But the specifications for
CPE were not exactly onerous! :-)

The graceful degradation tolerated by the analog network is not
tolerated by the digital network.  The move to digital switching
required a great deal of upgrading for typical outside plant
subscriber loops, and today the attention to the quality of the
loop is *much* greater than it was prior to digital switching.

(Heh, part of that can be credited to the use of v.90 dialup
access to the Internet.  To me that is a real hoot, because it
only came to be because the telecom industry refused to
implement ISDN with the claim that there was no customer base.
They refused to invest; and instead various modem companies did
and made millions...  which in the end forced the telco's to
invest in better subscriber loops without the potential for
profit that ISDN had.  The telecom industry is all just a
Dilbert Cartoon... which is no joke, because Scott Adams created
Dilbert while working as an ISDN Applications Engineer.)

The gain plans used today are in some ways more complex than
those used pre-digital.  I would imagine that is more true in
Europe than in the US too.  Europe used a fixed loss plan, while
(due to the typically greater distances involved) the US used a
variable loss plan, which admittedly cannot be called anything
other than overly complex!

But today a great deal more attention is given to subscriber
loop quality, which is somewhat balanced by a simplified circuit
loss plan, and then made more complex because longer distances
and international calls have become so much more common.
Overall, I think end to end it is more complex, but the part
referred to normally as a "Loss Plan" is indeed less complex.

Technology of course has made that complexity very easy to
handle.  Remember when echo suppression was both obnoxious and
expensive?  Today echo cancellation is commonly included even in
CPE devices (every modem!).  And it is not only not obnoxious,
it can make an otherwise obnoxious sounding circuit appear near
perfect to a customer.

Perhaps that is true of the most recently manufactured
equipment.  But the majority of equipment in use is not recently
manufactured.

Interesting.

At one time the Bell System only allowed an extension if it was
*paid* for!

Welllll...  I spent 4 decades getting it to actually work.  ;-)

I don't think anything you've said is "wrong".  But it is also
obvious that your concentration is on what it is supposed to do
once it gets to the field and mine is more on

"Floyd L. Davidson" <f @barrow.com> wrote in message

...which is what sometimes happens when an amplifier loses power. It depends
on the design, but it's quite likely that passive feedback components will
couple enough sound from input to output to still be heard faintly. Whereas
a simple impedance mismatch would maybe result in a loss of 10dB at most,
which is significantly less than the 30dB difference in levels that can
occur "normally".

Indeed. One of those great missed opportunities. Here it was actually
implemented, but too late and overpriced, which amounts to the same thing.

Quite! :-)

Indeed. Can still be a significant expense in VoIP terminals if a long delay
must be handled (needing lots of memory).

Is it not? I can't remember when I last saw a phone with a dial here, and I
would expect all electronic phones would meet the spec. Maybe it's different
in sunny Alaska - certainly you've had tone dialling for much longer than
us.

Likewise. The *differences* in ways to solve the same problem are always the
most interesting!

--
Phil McKerracher
www.mckerracher.org

"Floyd L. Davidson" <f @barrow.com> wrote in message



I was aware of the original description and I disagree with your
conclusion - the described "sharp drop" is more characteristic of a power
problem than an impedance mismatch, in my experience.


OK, I should have identified my location - I'm in Europe, where the
electrical characteristics of terminal equipment were regulated. Before
regulation there was a monopoly and internal specifications did the same
job. The analogue network was more forgiving in the sense that it was
(fairly) linear and so degraded "gracefully". However, horrendously complex
gain plans were needed to minimise echos and ensure stability, which have
been rendered almost obsolete by digitisation (because digital trunks are
essentially lossless, and you only need to worry about the "last mile").

Very rarely these days. In any case, the return loss of handsets in most
countries is specified to be within about 12dB of a defined impedance
similar to 600 ohms, which translates to an allowed difference of only a few
percent. See for example
http://www1.btwebworld.com/sin et/351v4p2.pdf


Again, maybe not in the US, but in the UK and many other countries there was
a specification (BS6317 from memory) that had to be met if you wanted to
allow this, and most phones did. It defined the DC characteristic to prevent
phones pulling the voltage too low. (It also addressed the problem of bell
tinkle when pulse dialling.)

Note that other countries prohibited the possibility and defined switching
schemes that allow only one phone to be connected at a time for "privacy"
reasons.

I do know what I'm talking about here - I used to design terminal interfaces
for international markets.

--
Phil McKerracher
www.mckerracher.org

That is a very likely possibility.

Also, in addition to variations between different products, any
given telephone set might exhibit changes in the impedance with
frequency in much the same way that a cable does (note the
previous mention of a "900 Ohm Loop" having an impedance that
actually varies from 2000 to 400 Ohms).

The effects of such variation can be strange, because while any
one telephone set will have been engineered to be within the
limits that would cause problems (such as echo on long distance
calls, or even difficulty for people with certain kinds of
hearing loss), when two phones are in parallel that may cause
some of those effects to be far out of specifications.

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

The phones with the "weak" volume are likely to be the self-contained
units, not the cordless units.

I think the "loss of volume" is more due to unequal division of the DC
power from the loop, rather than from impedance mismatch.  Technically
there will be a 6-7 dB level drop with two phones in parallel.  But
because of the way the human ear works, it takes a 10 dB drop in volume
to make one sound "1/2 as loud" as another.  So, this 6-7 dB drop can
be heard, but won't be interpreted as "the sound almost drops out
totally" by most people.

As another poster stated, a "basic" telephone set (no AC adaptor or
batteries) needs to draw some DC current from the "loop" to operate its
internal amplifier circuits.  The amount of DC voltage drop across the
line varies from one telephone set to another.  It is tricky to design
such a circuit in a way that it will be able to work well over a wide
range of currents and voltages.

In the case of the "weak" phones, the voltage between tip and ring is
probably too low to let the internal circuits work right.  I've seen
phones where you could still hear, but couldn't talk (or vice-versa)
when this happened.

There are a couple of ways to work around this:

If you can find some "traditional" 2500 sets that were made at least 25
years ago, those are electrically "passive" and will always work, even
with a very low voltage/current from the phone line.

If you have two IDENTICAL telephone sets (same brand, same model, same
age) you can use them both at the same time and the DC power will
divide equally between the two.

Many of today's higher grade cordless sets have "mulitple handset
capability."  For example, the excellent Panasonic KX-TG2740 system can
simultaneously bridge together up to 5 cordless handsets on one call!
There is no volume loss, and everyone's voice comes through at the
correct level on both the local and distant side of the conversation.
This Pansonic system actually has a real "conference bridge" built into
the base unit.  Note that in this case, there is only ONE SINGLE
connection to the external telephone line -- at the base unit.  Also,
all the electronics in the base unit is powered by an AC adaptor, so
whether the loop current from the telephone company is high or low
really doesn't matter.

It is possible.  But given the original description, it is
probably the impedance, not the DC power.

Actually it was not regulated in the old days, and the analog
network was *much* more forgiving in that respect that are the
digital switching systems used today.

And it is safe to say that many, probably most, telephone sets
do not exhibit a pure 600 Ohm impedance by any stretch of the
imagination!  If they use a strictly electronic interface, it
can be done.  But generally the interface uses a hybrid
transformer, and the cost of that transformer is directly
related to the impedance vs. frequency characteristics of the
telephone set.

None of them are!

It never was guaranteed.  It's just that with non-electronic
phones it has a better chance of working.

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

3 dB, exactly (if the two phones are identical).

  1 dB   --   just noticeable
  3 dB   --   significantly noticeable
  6 dB   --   customer will complain of some difficulty
 10 dB   --   customer will complain that line barely unusable
 20 dB   --   customer will complain that line is dead
 30 dB   --   line really is dead
 40 dB   --   legally non-existent signal

A 3 dB loss will make it sound "1/2 as loud", and a 6 dB drop can
make some calls difficult (if the distant party has a soft voice,
for example).

A 6 dB drop (1/4th the power) will make it merely uncomfortable.
It would probably require a 15-20 dB drop to get "drops out
totally" as the description.

However, the OP didn't describe his situation that way, but said
the "level of the caller's voice drops sharply and the caller
can hardly hear us until one phone is hung up".

That sounds to me like somewhere between 6 and 10 dB.  If the
level on that line is a few dB lower than the target of -9 dBm
for a testtone, it would only take 6 dB more loss to make it
that significant.  On a line with a couple dB higher than the
target, it might even be okay at 10 dB additional loss.

It usually means the phone is _dead_ though, not "the caller can
hardly hear us", which is more likely the variation in AC
impedance.  That is in fact a very common result of bridging
regular telephone sets (not even electronic ones) together.

Given the OP's description, it could be either the impedance or
a the DC power...  but I'd bet on the impedance as the most
likely of the two.

They might well "always work" as far as not going totally dead with the
other telset is offhook, but they will still suffer from the problems
of splitting the signal between two phones...

That is *really good* advice!

As near as I can tell virtually all of the current crop of
cordless phones with multiple handsets have that capability.
Some can have many handsets too.  I recently purchased a set
made by "Vtech" I think it was, as a Christmas present for some
friends.  I was impressed.  It appears the cordless designs have
benefited greatly from the popularity of cell phones and the
resultant research and economies that have resulted.  The
particular units that I played with were *far* better than
anything available as recently as perhaps 3-4 years ago.  And
the prices are attractive too.

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

If the telephone sets do not properly "load" the line, all Hell
breaks lose!

Essentially a telephone set is supposed to present an impedance
of 600 Ohms.  If you have a signal of 1 mW, and one telephone set
is offhook, that telephone set gets the entire 1 mW.  (It happens
that a 1 mW tone is rather loud too!)  But if two telephone sets
are offhook, and if they both have exactly 600 Ohms impedance, then
each of them is going to get half of the signal, or 0.5 mW.  That
isn't so loud.

But lots of cheap telephones don't present exactly 600 Ohm
impedance.  In fact, if is 300 Ohms (or 900 Ohms) it will work
quite well all by itself.  But if you happen to have one
telephone set that is 300 Ohms and another one that is 900 Ohms,
and use them both at once...  all the power goes into the lowest
impedance, and the other person can't hear a darn thing on the
high impedance set.

Unless some way is provided to crank up the volume when multiple
sets are offhook (which is exactly what a Conference Bridge
does), then there will be a problem.  And as long as the
telephone sets are pretty close to equal in impedance the losses
will be equal, but otherwise not.

Conference bridges are nice, and just using two phones on the same
line is not.

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

Whether 600 or 900 Ohms is specified, that is a /convention/.
The actual impedance of 26NL cable, for example, varies from
2000 Ohms at 200 Hz down to 400 Ohms at 4000 Hz.  26H88 loaded
cable is closer to 1200 Ohms at 1000 Hz, but loaded cable is
also less likely to be seen today than it was 2-3 decades ago.

Let me quote a section entitled "Subscriber Loop Design" from
"Telecommunication System Engineering", 3rd Edition, 1989, by
Roger L. Freeman.

   "2.4.6 Subscriber Loop Impedances.  For a conventional
   two-wire switch, the characteristic impedance is 900 Ohms.
   This is a /compromise/ /impedance/, and is the impedance
   looking into the line circuit."

   ...

   "Most equipment to be attached to a two-wire loop is
   considered to have a 600 Ohm impedance ... Both the 600 Ohm
   and 900 Ohm values are conventions and are compromises."

Hence you are correct that the loop impedance is 900 Ohms, but
what I had said was that the telephone set is 600 Ohms, which is
probably true, though not necessarily.  And, as noted, that is
exactly why some of them do not mix well with others!  (Some of
them have more variation in impedance than the cable does!)

Yes!  That involves having a different DC resistance, as opposed
to the AC resistance.  But generally I think that will not
result in lower levels where one party can barely be heard, and
is more likely to simply make one phone dead when the other is
off hook.

Another result of multiple telephone sets being off hook is a
little more difficult to explain.  The volume of the speakers
signal that is fed from the telephone set's transmitter to its
receiver is significantly less than the volume sent down the
line to the distant end.  That is, the talkback volume is
reduced.  Part of the reason is just because we can hear
ourselves talk anyway, but the major reason is to equalize the
signal loss that the speaker hears to make it similar to the
signal loss (over the loop, through the network, to the distant
telephone set) that the distant party hears.  That loss is
assumed to be about 9 dB, or about 1/8th the power level.

But of course if there are two phones off hook, they hear each
other *before* that 9 dB of loss over the network.  If the two
phones are identical, they split that signal as well as the
incoming signal (a 3 dB loss).  So while the distant party is 3
dB down from normal, the other phone set is 6 dB above normal!
Very annoying!

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

In article <87oeezlifa. @barrow.com>, Floyd L. Davidson says...

I take it that you're talking again about variability across phone products. If
so, that would be in accord with my recollection that one of our phones hits on
the line heavier than the rest. Can't remember which one it is, though...maybe a
systematic experiment is in order...

Good, except subscriber loops are nominally 900 ohms, not 600, like analog
trunks. I think there is an additional problem with electronic phones in
that most of them are line powered, needing to draw current from the talk
battery to run their electronics. A second off hook phone may effectively
short out the first's power supply.

I think the problem in this case is most likely that one of the phones is
taking all the power from the others, either because your line is long and
there's simply not enough power left for more than one phone, or because of
the design of one of the phones. In other words, it's essentially a problem
of DC loading not AC impedance.

It's true that different impedances at audio frequencies can cause some
variation as others have pointed out, but most phones are designed with an
impedance quite close to 600 ohms (or to a more complex one approximating
two resistors and a capacitor), because it's easy to do and bad echos can
happen otherwise. In the old days, this impedance was regulated because the
whole performance of an analogue network depends on it.

Many phones are simply not designed to be used "in parallel" like this, and
methods of coping with this problem and of limiting the current drawn vary
widely from country to country. So the chances of successfully connecting
phones in parallel depend a lot on when and where your phones were
manufactured, it's no longer guaranteed.

--
Phil McKerracher
www.mckerracher.org