[Asterisk-Users] zaptel and low ring voltage

[Jim Van Meggelen]

asterisk-users-bounces at lists.digium.com wrote:
>>>>> #1 - the ring voltage for some calls is below acceptable levels
>>>> 
>>>> Possible, but also possible that there is too much loss on the
>>>> circuit. 
>>>> 
>>>> You can test the ringing voltage with a meter, it needs to be
>>>> between 90V and 110V. 
>>>> 
>>>> Beyond that you may need to use a transmission test set (such as a
>>>> Wilcom T136B). I got mine for $20 bucks on eBay. Using a butt set
>>>> and the test set you'll need to call a 1004Hz source from TELUS
>>>> and then check that you're within the following specs:
>>>> Loop mA:		23 or better (too hot is no good either, but I
>>>> doubt that's your problem)
>>>> Circuit loss:	between 0 and -8dB. 0 is really too hot, -3 to
-6 is
>>>> nominal, -8.5 is pushing it, but still within spec.
>>> 
>>> Not likely the above is the root-cause as that implies the pstn line
>>> is 'changing' characteristics on almost a daily basis (highly
>>> unlikely). Also, the spec's for the 3050 chip (on the tdm fxo
>>> module) indicates it will detect ringing voltage down to 40v (or
>>> lower on some some specific configurations).
>> 
>> 
>> I wasn't aware of this spec.  If this is the case, then looking for
>> zaptel tweaks or even circuit loss may prove to be fruitless.
> 
> I'd have to agree on the circuit loss. If circuit loss were
> an issue, it would be impacting audio levels/quality waaaaaay before
> ringing. 

Hmmm. I've seen numerous cases where circuit loss did not cause
noticeable audio degredation but did cause signalling problems. Most CPE
can adjust the gain on received audio to nominal levels, but if the
ringing voltage drops below the level set in the system for ring
detection, the bells might not ring. I guess in the case of the Zaptel
cards we'd need to figure out whether ringing voltage has been set to a
certain threshold, or if any AC signal between 15 and 68Hz and 40 to 140
volts is acceptable (in North America, 20Hz 90Volts is nominal).

On a Norstar you could always tell if circuit loss was a problem because
the system would have trouble pulling dial tone. Converstions would
often still work fine for the users because the automatic gain would
boost the amplitude of the received signal to nominal. If the circuit
wasn't too noisy, no one would notice.

Anyhow, I am not trying to claim that the problem here is circuit loss.
That can't be known without testing. But to suggest that listening to a
conversation on the line can tell you whether ringing is being received
at acceptable levels doesn't make much sense to me.

>>> If you search through the archives you'll see others have also had
>>> issues with the tdm using fxo modules. Lots of references relative
>>> to stop/restart asterisk, unload/reload wctdm & zaptel, reboots,
>>> hear nothing but white noise, can only reload zaptel so many times
>>> before a machine reboot is necessary, etc. Seems there has been
>>> (maybe still are) issues that are driver related, kernel related,
>>> and/or possibly distro related.
>> 
>> 
>> I've deployed several of these cards, and have done my fair share of
>> griping about them (on the alias) in the past.  From experience in
>> these cases, when the card/module stops 'working', it will not
>> service any incoming/outgoing calls until module reload.  This is
>> not at all the case with this current implementation.
> 
> Do you have any other logs/info that would suggest what might be
> happening? 
> 
> Maybe turn on debug in /etc/asterisk/logger.conf for awhile
> and see if any messages are created at the time of failure?
> (as in hung channels, zombies, ring detected but no answer, no ring
> even detected, etc.)