OBi110 Disconnect detection in Spain (revisited)

[derketo-rick]

I have searched existing posts but not found any definitive statement of a working configuration to detect disconnection in Spain.

The tone sequence (Physical Interfaces->Line Port->PSTN Disconnect Detection->DisconnectTonePattern) is approximately:

425-30;10;(.2+.2,.2+.2,.2+.6)

as suggested [here][http://www.obitalk.com/forum/index.php?topic=3510.0] - three short 200ms tones then long 600ms gap - which is apparently rather unusual in a global telco context.

There are hints that support may have something definitive and I have sent in a request for info but if anyone knows some magic tone setting or other setting which is known to work please post. There could also be some other hardware setting I have not seen mentioned at all.

If I get anything to work I will post back here as well.

Thanks
Rick

[hwittenb]

In the ATA Administration Guide for their products, Cisco/Linksys says the disconnect tone for Spain is:
Spain—425@-10; 10(0.2/0.2/1,0.2/0.2/1,0.2/0.6/1)

which I imagine would translate to OBi as
425@-10;10;(0.2/0.2/1,0.2/0.2/1,0.2/0.6/1)

[derketo-rick]

Thanks for that hwittenb.

From my reading of pp148-9 of the OBi Device Administration Guide (Version 26.08.12 – 26 August 2012) I think the Cisco pattern definition should translate to:
   425-10;10;([1/]0.2+0.2,[1/]0.2+0.2,[1/]0.2+0.6)
where you can leave out the bits in square brackets as they indicate tone1 from the list of (one 425Hz) tone(s) at the beginning of the pattern definition.

To break that out, this is (if you remove line breaks added for clarity):
   freq1-dB1[,foptional_req2-dB2][,optional_freq3-dB3];
   secs_tot_duration;
   secs_pattern_duration(
       [optional_tone_number_defaulting_to_1/]t1_on_secs+t1_off_secs,
       [optional_tone_number_defaulting_to_1/]t1_on_secs+t1_off_secs,
       [optional_tone_number_defaulting_to_1/]t1_on_secs+t1_off_secs
   )

@hwitttenb: Do you have a pattern definition break-out like this from the Cisco/Linksys manual so I could compare to make sure I am not missing something here ?

By observation we seem to have a pattern of three 0.2 second 425Hz pulses (signal level unclear - this may be the problem but I have tried several dB levels), separated by equal 0.2 second gaps, with this three-beep pattern repeated after a 0.6 second gap, giving a total pattern length of 1.6 seconds. This repeats 25 times before switching to a screechy three-rising-tones service disconnect pattern. I suppose I could look for that pattern instead but that means a delay of over a minute before detecting the disconnect, and the other pattern should be matchable as it is quite distinctive.

I was hoping for a definitive response from anyone who has made it work with an OBi110 in Spain - anyone out there ?

Thanks again
Rick

[ianobi]

Rick,

I can't help you with the disconnect tone, but have you considered experimenting with "DetectCPC"? The OBi default "CPCTimeThreshold" is 450ms. This is way too high for some countries. For example for CPC to work here in the UK "CPCTimeThreshold" needs to be set at 90ms.

CPC is called different names in different countries. CPC stands for Called Party Control. When the called party hangs up, the local telco sends a short time of zero voltage back to the calling party. The OBi can detect this as a disconnection.

[drgeoff]

.... signal level unclear - this may be the problem but I have tried several dB levels ...

The level should not be critical.  It can only be set by the telco at their switch (aka local exchange or local office).  There will be an unknown amount of line loss (within limits) so the actual level at the customer cannot be accurately predicted.

[hwittenb]

derketo-rick

This is the Cisco documentation that I can find:

"This value is the tone script which describes to the SPA the
tone to detect as a disconnect tone. The syntax follows a
standard Tone Script with some restrictions. Default value
is standard US reorder (fast busy) tone, for 4 seconds.
Restrictions:
• 2 frequency components must be given. If single frequency
is desired, the same frequency is used for both
• The tone level value is not used. –30 (dBm) should be used
for now.
• Only 1 segment set is allowed
• Total duration of the segment set is interpreted as the
minimum duration of the tone to trigger detection
• 6 segments of on/off time (seconds) can be specified. A 10%
margin is used to validated cadence characteristics of the
tone."

Page 182:
http://www.cisco.com/c/dam/en/us/td/docs/voice_ip_comm/csbpvga/ata/administration/guide/ATA_AG_v3_NC-WEB.pdf

also
http://en.wikipedia.org/wiki/ToneScript

Edit:
I found the following OBi Device Administration Guide
Version 01092013 dated September 2013.
http://www.obihai.com/OBiDeviceAdminGuide#_Toc367543110

The Guide has a section on Tone & Ring Patterns plus a few limited examples.  I agree with your interpretation of the Cisco definition:
425-10;10;(.2+.2,.2+.2,.2+.6)

425 Hz at 10Dbm with the cadence 10-seconds of .2 sec on .2 sec off + .2 sec on .2 sec off + .2 sec on then 0.6 sec off for 1.6 sec total

You say by observation the sequence doesn't last just 10-seconds but lasts 40 seconds (1.6 x 25). That's a long time.  Assuming the 10-seconds is the "minimum" duration it sounds to me like the spec matches the observed.  Maybe there is something wrong with OBi's tone matching.

[derketo-rick]

Quick note to say thanks to all for help, especially hwittenb for his research. I wanted to follow up quickly but did not have chance to try anything before returning to UK. I will try a few other settings (eg CPC) at the weekend and let you know if get anything to work.

The previous topic about this ended with forum 'hero' member Stewart tyrying to help but saying he could not get the OBi to match this pattern when he tried a recording of it as his Google VMX message, and he therefore gave up, but there was a message from OBi support saying they were going to help the OP offline, so I hoped there was a known 'official' solution. Support did get back to me on my own enquiry but just gave the UK settings (which it shipped with) and suggested I try the forums !

The two main options, assuming we are correct in our collective suspicion that this pattern is for some reason 'unrecognisable' by the OBi, are the CPC, if Telefonica use that signal, and the shrieking 'you left the phone off the hook' three-rising-tone signal which cuts in around a minute and a half after the 'disconnected' pattern. CPC would be an ideal replacement for the mystic tones if I can get it to work as it will occur much more quickly.

Thanks again.
Rick

[hwittenb]

Whether you get a CPC (Calling Party Control) signal or a disconnect tone is a function of the PSTN Telco switch that your pstn line is connected to.  The CPC signal is a momentary loss of voltage.  It nominally lasts 500ms although, as Ianobi says, it can be much shorter than that.

The other option for disconnect detection is Silence Detection.  The Obi's default settings appear to set to detect that with
DetectFarEndLongSilence: YES
SilenceDetectSensitivity: Medium  (Low/Medium/High)
SilenceTimeThreshold: 60 seconds

Of course, this would occur after the disconnect tone and the subsequent shrieking.

[hwittenb]

Looking at the tone examples in the OBi Admin Guide, specifically the SIT Tone, perhaps the OBi could detect the Spain Disconnect Tone if it is configured like this:

425-10;10;(1/0.2+0.2,1/0.2+0.2,1/0.2+0.2,0/0+0.4)

Which I believe is
425-10 425hz @ -10 dBm
10 seconds minimum duration
1/0.2+0.2 --> Freq 1 0.2 sec on, 0.2 sec off
1/0.2+0.2 --> Freq 1 0.2 sec on, 0.2 sec off
1/0.2+0.2 --> Freq 1 0.2 sec on, 0.2 sec off
0/0+0.4 ----> Tone 0 (silence) 0 sec on 0.4 sec off

http://www.obihai.com/OBiDeviceAdminGuide

Looking at the .wav file for the Spanish disconnect tone that was posted as a reference in an another earlier posting, my analysis program says it is 422Hz.  You might try that if 425Hz doesn't do it.

Edit:  I would also try specifying shorter time than 10 seconds if 10 seconds doesn't do it.