Thanks everybody for the help so far on Asterisk issues.
Here's what has helped:

1) The wiring from T.P.C. was going through a dogs breakfast
of cables and coupling blocks.  I replaced all that with
a 66 punchdown block and a few appropriate length cables.
That got rid of some noise, which helps with echo training.

Why noise?  Every time you have a non-gas-tight connection,
here in moist Oregon, you get corrosion, which acts like a
series of tiny short-duration batteries. That is, voltage
pulses added to the circuit, and noise.  

2) I had a couple of analog phones acting as "emergency
backups" plugged into the lines in parallel with the
Asterisk FXO cards.  Unplugging those also reduced noise,
and I will add some switches to turn them on in case of
power failures.

3) fxotune helps.  Thanks for the clue.

4) I appreciate the suggestion of Digium's HPEC, and may end
up using it, but it seems that the Kool Kids are using the
open source "oslec" echo canceller instead of the default mg2.
I haven't managed to get this to compile and load properly yet, 
but eventually it will fall to the mighty hammer of "googling
for error messages".  Hindered but not stopped by the name 
migration from zaptel/zapata to dahdi.

----

BTW, one of the things I hoped to connect was an LED kludge
I built that lights up when a line rings - that was our way
of letting the assistant know to put the current caller on
hold and pick up the other line, since the call-waiting 
indication on the Cisco phones is too subtle.

I designed the kludge for the standard voltages from T.P.C.
+50 volts on hook,  25 Hz and 300 volts (+200V/-100V), 
peak-to-peak for ringing.  Sadly, the wimpy little FXS
cards I have produce 50 volts on hook as before, but the
ring signal is only 130V p-p, (+65V/-65V).  I can rewire
the LED stack for the lower voltage, but it will draw too
many ringer equivalents if I plug it in to T.P.C.again.
If anyone else has an Asterisk system with FXS cards
connected to analog phone desksets, perhaps I can bring
my oscilloscope over and measure those, and see if it
is my cheap cards or (more likely) this is non-standard
behavior that started with the first zapata designs.

Keith

-- 
Keith Lofstrom          [email protected]         Voice (503)-520-1993
KLIC --- Keith Lofstrom Integrated Circuits --- "Your Ideas in Silicon"
Design Contracting in Bipolar and CMOS - Analog, Digital, and Scan ICs
_______________________________________________
PLUG mailing list
[email protected]
http://lists.pdxlinux.org/mailman/listinfo/plug

Reply via email to