[Asterisk-Users] ISDN/CAPI outgoing calls - weirdness with ringing

[Armin Schindler]

On Sun, 25 Dec 2005, Michael J. Tubby G8TIC wrote:
> > On Sat, 24 Dec 2005, Michael J. Tubby G8TIC wrote:
> > > I changed the dial-string to include flags 'ob' as you mentioned
> > > (below)
> > > and now I get the following when I dial a BT phone number
> > > 
> > > - dial number, get:
> > > 
> > > Proceeding (in 100) briefly
> > > 
> > > - after a second or so:
> > > 
> > > Ringng Destination (in 180)
> > > 
> > > - double ringing tone:
> > > 
> > > BT style ringing generated by the exhange
> > > Cisco phone US-style ringing (generated by the phone)
> > > 
> > > these are overlaid on each other (mixed together)
> > > 
> > > 
> > > My hunch is that there's something not right with the call set up
> > > sequence
> > > and CAPI handling.
> > 
> > This is not a problem of CAPI. When you specify 'b' for early-b3, you
> > will
> > get the tones from the switch. If your phone adds its own tone, even when
> > it
> > receives progress tones, then it is incorrect (maybe wrong setup).
> > 
> > Armin
> > 
> 
> 
> However the difference that I see looking at the Cisco 7960 phone which
> shows a version of the SIP messages on its status line is:
> 
> 100 Proceeding
> 183 Session Progress
> 180 Ringng Destination
> 
> the order of which varies and depends on the dialled number.
> 
> Some dialled numbers go: 100->183->180 and these produce one set
> of alerting/ringing correctly.
> 
> Some dialled numbers go: 100->183 and stay in state 183 until the called
> party answers - these are the ones that produce no ringing.

Can you provide a verbose log level 5 with 'capi debug' ?
I would like to compare the capi messages. Maybe the switch just send an 
alerting message.
 
> If I add the 'o' to the existing 'b' flag then dial it appears to change the
> behaviour so that the phone goes 100-180 for all calls but some give
> me a single (phone generated US style ring) while others give the 'double
> ringing'.  The ones that produce double ringing are the ones that would
> have rung before, while the ones that now produce ringing (from the
> exchange) are the ones that used to be silent.

When using 'o', chan_capi is doing early-b3 from the beginning before 
sending any digits and you will get b3-data in each case.
Please send me a debug log of a connection with double ring-tone (no 183)
as well.
 

Armin