[Asterisk-Users] W&M Wink timings for Nortel

[Rich Adamson]

> >>>I'm confused. Digium cards do not support E&M trunks, which are older
> >>>interfaces that include either a two-wire or four-wire audio "plus"
> >>>two additional wires called the "E" and the "M" leads. So, Wink
> >>>timing is irrelevant (unless I'm misunderstanding your question).
> >>
> >>Uh... my zapata and zaptel.conf exmaple files both show E&M signaling options.  
> >>I'm no telephony expert but I though that E&M signaling was just buried in the 
> >>RBS bits like any other CAS T1 signaling.
> > 
> > 
> > When I responded to Eric's original post, it sounded like he was trying
> > to interface to an older pbx trunk interface (physical interface).
> > Those physical E&M interfaces were as described above (each had either
> > four or six wires).
> > 
> > What the original post didn't say was the interface to the non-asterisk
> > pbx was a T1, and one (or more) of the channels were configured for E&M
> > signaling. Completely different then a physical E&M interface. (Someone
> > else had asked about a physical interface not more then a couple of
> > weeks ago, so the mind was oriented along those lines.)
> 
> The original setup was:
> 
> CLEC CT1 -> Channel Bank -> (analog) -> Nortel PBX.
> 
> Current setup is:
> 
> CLEC CT1 -> Astrisk -> CT1 -> Channel Bank (analog) -> Nortel PBX.
> 
> The Nortel does not have T-1 interfaces in it and the company is 
> unwilling to buy a T-1 card for the Nortel.
> 
> Only 4 channels of the T-1 are E&M Wink (they are for DIDs), the rest 
> of the channels are various other things.  For the voice channels that 
> we care about we terminate the channels into Asterisk.  For channels 
> we don't care about we use the Zappel DACS stuff to just patch the 
> channels thru.
> 
> Channels 9 - 12 are the E&M Wink channels.  This is what we see:
> 
> Failure #1:
> 
> -- Starting simple switch on 'Zap/9-1'
> Feb 18 16:13:25 WARNING[1965]: chan_zap.c:4723 ss_thread: getdtmf on 
> channel 9: Operation now in progress
> -- Hungup 'Zap/9-1'
> -- Starting simple switch on 'Zap/10-1'
> Feb 18 16:13:29 WARNING[1966]: chan_zap.c:4723 ss_thread: getdtmf on 
> channel 10: Operation now in progress
> -- Hungup 'Zap/10-1'
> 
> Failure #2
> -- Starting simple switch on 'Zap/9-1'
> Feb 18 16:13:25 WARNING[1965]: chan_zap.c:4723 ss_thread: getdtmf on 
> channel 9: Operation now in progress
> -- Hungup 'Zap/9-1'
> -- Starting simple switch on 'Zap/10-1'
> -- Executing NoOp("Zap/10-1", "EXTEN=189") in new stack
> -- Executing Dial("Zap/10-1", "Zap/G3/189||g") in new stack
> -- Called G3/189
> 
> In Failure #2 the calls still go thru, as you can see.
> 
> As you can see when a call suceeds I do not see the getdtmf WARNING 
> message.
> 
> We have no problems sending calls to the channel bank, only getting 
> them from the CLEC.  Since the CLEC used to be directly connected to 
> the channel bank and then to the Nortel I want to get the wink timings 
> for the nortel to see if making Asterisk match the Nortel settings 
> make any difference.

Since you're really dealing with signaling bits embedded in the T1
stream, my original comments relative to the physical interface obviously
don't apply.

In the old E&M physical interfaces, the E lead represented signaling
coming from the world (sort of E = everyone) to you, and the M lead
was you signaling to the world (sort of M = me). The two leads were
used to solve loop signaling issues, handle answer supervision, dial
pulse signaling, trunk startup, etc. The audio path was either two-wire
(eg, pbx interface) or four-wire (eg, LD circuits). I don't recall
any config options (from my telephony engineering days) relative to
wink duration, or any other timing adjustments.

My guess is the asterisk implementation for E&M signaling is probably
"one" end of that interface, watching the signaling bits and translating
those into something * can use to terminate the call. However, I'll 
be the first to admit I'm not a programmer and probably wouldn't 
recognize the code even if I seen it. Highly unlikely it is a 
fully-baked E&M emulation, and even higher probability the code can 
not source an E&M call.

Do you know if the incoming DID digits are actually dtmf, or is it
possible the digits are dial pulse on the E lead? (I'd have to guess
dial pulse from the looks of those messages.)

Since I have not personally heard of anyone actually using it, I'd have
to guess that resolving the problem will boil down to writing debug
code into asterisk, and then writing code that supports your needs.
Or, change the interface from asterisk to the Nortel to some other
trunking arrangement.

Rich