dya@unc-c.UUCP (03/24/84)
<Here's the beef....>
Well, it seems that the problem with the UDS modems in Orchard Park
was that there seems to be some extremely long time constant AGC/Compandor/
Noise Gate/something with a delta-G function... that was chopping off the
front of the training burst. Thus, the adaptive equaliser inside was getting
the wrong idea.
If we pulsed the lines with a short burst, then went back and repeated
it (before the AGC/whatever returned to a quiescent state) with sending real
data, the constellation from the modem is darn nearly PERFECT over these
extremely long loops.
NOW the question ois:
1) On local loops and ESS 2B, what kind of funny amplifiers are there
such as noise gates/AGC/analog companders with extremely slow attack
times ?
( This effect wasn't noticed on long distance loop through UDS's
own loop-through facilities )
2) How do we get RID of them ?
3) It doesn't seem to be in every ESS 2B, as they do work fine here
in Charlotte ESS office ( 535, 568 ), just Buffalo <Orchard Park> 662-
and Hamburg, NY..
Please !!. We aren't going to lose your job for you, we just want these
darn modems to work everywhere ( after all, 9600 bps Compuserve isn't bad ?)
for a research project. They * HAVE * to work in New York, and they seem to
work anywhere else. No one at New York Tel has a definitive answer, so we have
ESS 2B is either all digital or analog within the C.O. ).
David "they're companders, but it sure ain't DBX" Anthony
Reply to { akgua, decvax ! mcnc ! urp ! dya -- uncc is up and down these days }