[asterisk-users] Asterisk not reading fast DTMFs, was: PBX -> PRI -> * -> Telco not working

Mikel Lindsaar raasdnil at gmail.com
Wed Nov 19 03:46:34 CST 2008


>
> I plug the NEC back straight to the Telco and all works well again.
>>
>
> I just got on the phone to Digium and we've raised a ticket with some pri
> intense debugging going on. I'll update the list on findings.
>
> On Wed, Nov 19, 2008 at 10:32 AM, Brent Davidson <
brent at texascountrytitle.com> wrote:

> I have a weird thought...  Is the PBX possibly passing the digits both
> inband and via PRI signaling so Asterisk is getting two digit streams at the
> same time and totally freaking out?
>
>
You know.. that is probably it....

What the NEC system is doing I think is when you pick up the POTS phone to
dial, you go to the NEC's LCR program (least cost routing).  It then reads
the first digits of your call.

When it determines how to route your call (in our case, we have made it
route everything out to the PRI) it then must send the digits out via
PRI signaling.

Maybe it captures three digits before deciding what to do, so it sends them
out via PRI signaling.

It would also capture the remaining digits and send them too via
PRI signaling, but then the analog phone is ALSO sending the remaining
digits via inband audio and then asterisk gets the first three via
pri signaling, and the last 5 via inband, and instead of putting the
pri signaling first and the inband second, is interleaving it.

This must be how the Telco actually managed to router the call.  Because it
must go 'pri signaled digits first, inband second'.  Because if you take the
pri signal digits (which we assume are the first three) and put them at the
start, you can see the number, all in the correct sequence.

Thanks for this idea, I'm going to send it off to Digium and get it added to
the ticket.

Mikel

-- 
http://lindsaar.net/
Rails, RSpec and Life blog....
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.digium.com/pipermail/asterisk-users/attachments/20081119/39c433ed/attachment.htm 


More information about the asterisk-users mailing list