[Asterisk-Users] Anyone using * in a live production environment?

[Jorge Mendoza]

I'm in agree with all explanations regarding the echo and 2/4 wires 
conversion. However I'm wondering if there are other parameters like CPU 
and/or Asterisk configuration involved in the problem with more weight 
than hybrid. Otherwise how do you explain the difference in the 
following scenario:

1.- Crystal clear voice:

[phone1]----[pabx]-----[fxo gateway]------SIP-------[fxs 
gateway]------[phone2]

2.- A lot of echo:

[gnophone or xten]---[ * ]---------------SIP-------------[fxs 
gateway]-----[phone2]

The first scenario has four 2/4 W conversion. The second one has only 
two (or one?).
The * was running in differents CPU, PIII 500 Mhz, PIII 750 Mhz, 128 Mb 
to 512 Mb ram with not difference on echo.
We have installed a Mitel 3100 with IP phones at 40 kms within a 
wireless network with not echo at all.

Jorge

Stephen R. Besch wrote:

> Brian D Heaton wrote:
>
>> IIRC, proper functioning of the 2-wire to 4-wire hybrid depends on
>> proper balance between the the individual wires of the pair.  If you
>> upset the balance you're going to get all kind of problems.
>>
> I just finished modelling a standard 4-transformer hybrid coupled to a 
> balanced RC transmission line. Cross talk was zero when the hybrid was 
> balanced. Inserting a single resistor in  series with tip or ring 
> imbalanced the hybrid and cross talk appeared. This could be 
> completely compensated with the proper RC on the opposite side of the 
> hybrid, as predicted. It made absolutely no difference to the 
> cancellation if the resistor was split.  Since a balanced hybrid 
> appears as a pure resistance (complex terms are 0)  to the 
> transmission line, placing a simple resistor in series with the hybrid 
> (on either side) at the termination point will just look like 2 
> resistors in series and will properly terminate the line.  There 
> should be no effects at all from doing this other than the loss of 
> some energy in the termination resistor, which can be made up for with 
> a boost in Rx gain.
>
>>
>> Along the same lines, overdriving the hybrid is a big source of echo.
>
> That's because the cores saturate on transformer based hybrids.  This 
> is not as likely to occur with active hybrids built with op-amps 
> (which are found in almost all modern line cards), although it is 
> possible if the gains are high enough.  However the distortion from 
> the clipping would be far worse than the echo.
>
>> The correct way to set the RX/TX gain values would be to get the number
>> for the "milliwatt test signal" from a friendly telco tech.  You can
>> dial that number through * and then adjust the RX gain value so that the
>> signal is loud, but not overdriving the ADC on the analog interface
>> card.  A good starting point on the TX gain would be whatever the RX
>> gain ends up at.  The proper way to set TX gain would be to use a
>> digital version of the miliwatt tone (1004Hz at 0dBm IIRC) outgoing and
>> adjust it for a proper level through to another line on the same telco
>> switch or across a T1 span that you can observe with a test set.
>>
>> I don't recall the exact specs for what percentage of the full-scale
>> value milliwatt tones should be at on the digital systems.  Anyone have
>> the reference?
>>
>>             THX/BDH
>>  
>>
>
>
> _______________________________________________
> Asterisk-Users mailing list
> Asterisk-Users at lists.digium.com
> http://lists.digium.com/mailman/listinfo/asterisk-users