dlw@ucbtopaz.CC.Berkeley.ARPA (David Wasley) (06/19/85)
In article 'pttesac.152', dated 20 Jan 85, Marnix van Ammers described
one possible explanation for the commonly observed interference
associated with 212 modem usage. "Sync slip" in the T1 trunk interface
was assummed to cause a few "bits" to be generated by the receiving
modem. This results in the characters DEL (0xFF) and '{' (0xFB) [assuming
a mark parity bit - I've never actually checked that] being seen by the
receiving DTE. It seems always to be that pair, in that order, but
with varying frequency depending on the actual routing of the call.
I never saw any follow up on Marnix's article. Can anyone add further
technical detail or explanation for this phenomenon? Explanation of the
dibit encoding and data scrambling might be useful - I know some of it.
Why does it always happen in the originate-to-answer direction?
We've finally convinced the local telco folks to do some serious testing
in our Northern Telecom DMS100 centrex but they're not sure what to
look for. Their test equipment is geared towards signal-to-noise and
signal level problems. They say they can also look at "phase hits."
One additional fact: at least in our case, we see that symptom in calls
originating and terminating on the same switch. That is, there are no
T1 trunks involved: it is a call local to our centrex. And, the DMS100
switch is all digital from line card to line card.
Any (serious) technical advice will be most welcome. Thanks,
David "It can't be happening" Wasley
...!ucbvax!dlw
dlw@Berkeley.ARPA