I wouldn't be surprised if you were seeing it try to dial DTMF, then pulse, then DTMF again if the call didn't cut through very quickly.
I saw this with some fire panels that I was connected to once upon a time - they'd dial 7 digits, the adtran was configured to cut through on 10 digits or to give a post dial delay waiting for additional digit collection, and that took more than 3 seconds to fully connect so the panel would give up, hang up, and try pulse dial. Then it would hang up, try DTMF again, then hang up and try pulse again. Then it would fault and evacuate the school. Needless to say, that was an awkward call to show up on at an inner city charter school, pulling up to greet a bunch of evacuated students and an annoyed fireman who stuck around in a fly car just to give the all clear. Setting the adtran to 7 digit dialing cleared the issue, fwiw. That was on the TA900 series, but maybe something similar is happening here? -Paul > On Feb 13, 2023, at 3:47 PM, Jay Hennigan via VoiceOps > <voiceops@voiceops.org> wrote: > > On 2/13/23 11:48, Matthew Yaklin wrote: > >> --- The issue is that when the alarm panel dials the INVITE from the TA5k >> does not contain the correct digits dialed. A couple of digits are missing >> for example. Sometimes a call does manage to squeak through properly. If the >> digits do get to the Metaswitch properly the call completes fine. Same exact >> symptom with GR303. In SAS (meta's service assurance server that contains >> debug output) I can clearly review what digits reach the metaswitch via >> GR303. In this case I will see missing digits as I know the number the alarm >> panel is supposed to dial and a gap of time which should have contained >> digits. > > OK, so the FXS port of the local Adtran device isn't reliably decoding the > DTMF from the panel. Could be frequency, level, or twist. See if there are > knobs for receive level and/or impedance for the specific line port used by > the panel. I'd start with level, bump it up and down 3dB at a time and see if > it starts decoding reliably. If possible bracket it to see where it fails > again and set it in the middle. > > Changing impedance might help if it's twist. Sometimes the options include > series capacitance. Likely going to be trial-and-error. > > -- > Jay Hennigan - j...@west.net > Network Engineering - CCIE #7880 > 503 897-8550 - WB6RDV > > _______________________________________________ > VoiceOps mailing list > VoiceOps@voiceops.org > https://puck.nether.net/mailman/listinfo/voiceops > _______________________________________________ VoiceOps mailing list VoiceOps@voiceops.org https://puck.nether.net/mailman/listinfo/voiceops
