[Asterisk-Users] Qs about FXO/FXS cards

[Jim Van Meggelen]

Dorn Hetzel wrote:
> On Sat, Jan 01, 2005 at 07:23:58PM -0500, Jim Van Meggelen wrote:
>> [...]
>> What if, for example, the TDM400 issues were a cumulative thing? If
>> you had over 6dB of attenuation on the PSTN loop, coupled with
>> greater than 5V potential on the neutral-ground of your elecrical
>> receptacle, compounded by a cheap power supply, exascerbated by a
>> Via-chipset, would you not be virtually guaranteed some strange
>> behaviour? But if your PSTN was -3dB, your electrical feed derived
>> from a power conditioner, your power supply manufactured by PC Power
>> & Cooling, and a ServerWorks chipset-based MoBo, would your system
>> always be faultless? 
>> 
> Can you recommend any favorite motherboards?

That is the million dollar question. Chipsets and MoBos seem to change
so fast that I've lost confidence in my ability to make sense of it all.
The Intel and ServerWorks chipsets are generally well regarded; Via
chipsets are almost universally avoided for audio work. The
linux-audio-dev folks seem willing to give nVidia's nForce chipsets a
chance. In general, I would avoid PC-class motherboards, and go with
server-class motherboards. That being said, the ultimate goal would be
to find a way to build a reliable Asterisk system on *any* half-decent
motherboard.

Personally, I'm of the mind that power (the power supply, the AC being
supplied to the system, and grounding) plays as much of a role as the
motherboard does, but that is a working theory only. I wonder if clean
power on a lousy MoBo might serve as well as dirty power on a quality
MoBo. If one reads about power quality issues, the symptoms of dirty
power sound suspiciously similar to the kinds of problems people are
having with their analog Asterisk cards.

I'm also wondering about the TDM400s ability to handle PSTN loops at the
extreme limits. Since those TDM cards were probably developed largely in
a lab environment, the telco lines would have been simulated with a
channel bank or C.O. simulator. What happens when the lines fall out of
certain limits? Annenuation, loop current, and longitudinal imbalance
are all factors that proprietary PBXs are able to correct within fairly
wide limits - but they do have limits (a Norstar, for example, tends to
have trouble pulling dial tone when attenuation exceeds 7dB). Has the
TDM400/FXO been similarly optimized? It must have limits; what are they?

I think what we are all looking for is some empirical evidence of what
conditions cause the biggest problems. Is it the TDM400 that is to blame
(either hardware or drivers)? Or is it Linux? PC Hardware? Telco Lines?
Electrical? Grounding? A combination of some or all of those factors? No
one seems to know for certain.

Where much frustration comes from is the fact that a typical PBX simply
does not suffer from these troubles. We've come to expect that our
telecom equipment handles these little noises for us (so much so that
we're suprised to find that these are genuine engineering issues). With
Asterisk, some of the responsibility for correcting those problems falls
to us, the system designers. Unfortunately, a comprehensive engineering
methodology for analog devices on Asterisk does not yet exist. It's all
kind of hit-and-miss. 

Cheers,

Jim.


-- 
No virus found in this outgoing message.
Checked by AVG Anti-Virus.
Version: 7.0.296 / Virus Database: 265.6.7 - Release Date: 30/12/2004