Robert, at 09:37 on Mar 01 1996, you wrote to Bill Grimsley...

BG> How long is a piece of string... ?

RJ> Twice the distance from the middle to one end!

Give the man a cigar!

RJ> Which came first, the chicken or the egg? The rooster!

That conundrum is far too esoteric for me.  Pass.

RJ> I've been sitting back, watching the on-going saga of "The
Modem-Ones" and
RJ> can't help but laugh 'cos even tho I've only got a 14.4k ,EC-less modem,
RJ> I've never had any problems with TMTL that could be attributed to the 
RJ> modem.

I'm not surprised, as the session seems to fail during EC negotiation.  Do
you have any idea how cruel it is to see CONNECT 28800/V34/NONE ?

RJ> The general idea seems to have swung around to it being a 
RJ> Telescum/line level problem. I tend to agree with this. 

Yeah, that's almost a foregone conclusion now.

RJ> I was a Tech-In-training with the P.M.G. when I left school and one of the 
                                      ^^^^^
Jesus, you're as old as I am then (if not even older)!  :)

RJ> first things you learn is the characteristics of your average telephone 
RJ> line. They were designed and built for voice operation (modems weren't even 
RJ> thunk ov then) with a bandwidth of from 300cps to 3000cps and a nominal 
RJ> impedance of 600 ohms. The loop resistance can be as high as 6000 
RJ> ohms. These figures were quoted from subscriber to exchange. Telescum used 
RJ> to quote a maximum of -9dBm permitted line out-put to stop (read
"reduce") 
RJ> cross-talk between ajacent pairs of wire in cables. I suspect that most of 
RJ> the subscriber lead-in and street cables are still the original as-laid 
RJ> stuff which consists of single-strand, hard-drawn copper with on-average 11
RJ> wrapped-only joins between each subscriber and the exchange. If you want 
RJ> reliable data exchange, you have to ask Telescum for a "data
grade line". 

Upon which they STILL don't guarantee 28k8bps, although it's supposed to be
better than the 2400 baud they currently guarantee on residential lines.

RJ> If you complain about modeming on a normal phone line, Telescum will send a
RJ> token technician out who will put a very complicated looking meter across 
RJ> your lead-in block, try to look wise and then tell you that the problem 
RJ> MUST be in your modem 'cos your line is "balanced".
"Balanced" just means 
RJ> that he reads approx. the same impedance from both legs to earth (that is, 
RJ> provided you actually have TWO wires connecting you to the exchange. A lot 
RJ> of buildings that had another phone installed after the first one sometimes
RJ> had the original pair split and both phones used a single line with earth 
RJ> return.) It's a bloody wonder that  modems work at all when you start using 
RJ> micro-wave and optical links. Each channel in a link is frequency shifted 
RJ> at least twice and is not the same signal at the end as what started. Makes 
RJ> it a bit hard to get a reliable link for you foreigners #:^}

Part of the problem could also be the excessive line loading caused by
having too many instruments hanging off the same line (common in this day
of multiple phones, faxes, answering machines etc.) , which lowers the
impedance.  This actually happened here when I first went V.32bis with my
Intel 400, a few years ago.  My 14400 connects were very sporadic, and it
was some months before I saw fit to remove certain peripherals from my line
as a test.  Turned out to be an old answering machine, which stated a REN
of 1.0, but was quite obviously considerably greater than that.  I suppose
that one should be grateful that the old infrastructure works as well as it
does, all things considered.

Regards, Bill

--- Msgedsq/2 3.20