| Author |
Message |
Dan Lanciani
Guest
|
 Posted:
Mon Nov 22, 2004 2:23 am Post subject:
ISDN BRI line noise problem |
|
|
I've had a BRI ISDN line (off a 5ESS I think) working perfectly for about
four years. I use a Motorola Bitsurfr Pro II for voice and one synchronous
B channel of DOV. The voice path has always been noise-free and the data
use has never logged a single CRC error. Recently I started getting severe
static-like noise on the POTS port and lots of CRC errors on data. There
were also disconnects and complete loss of sync. Telco can find nothing
wrong (of course :).
Disconnecting and re-connecting the line sometimes (but not always) improves
the situation significantly, sometimes for many hours. By "improve" I mean
that things are completely usable, but there is still the occasional CRC
error. Swapping in a different Bitsurfr also sometimes (but not always)
improves the situation for a while. I believe this is simply the same effect
as disconnecting the line temporarily; however, it does seem to vindicate
the local hardware as I was able to go from severe problems with original
TA to severe problems with second TA on at least some occasions.
When I first had the line installed I added a 1000' spool of wire in line
just to test my loss headroom and the line operated perfectly that way
as well. So I don't think I had a marginal loss situation even at nearly
18,000' from the CO.
One thought I had is that the characteristics of the line are somehow
changing (water infiltration?) over time and a retrain operation makes
things right for a while--until something changes again. But that seems
a bit of a stretch. Is there any kind of common equipment in the switch
such that I might be getting a bad line termination device at times and
not at others?
Dan Lanciani
ddl@danlan.*com |
|
| Back to top |
|
 |
danny burstein
Guest
|
| Posted:
Mon Nov 22, 2004 2:23 am Post subject:
Re: ISDN BRI line noise problem |
|
|
In <1331230@news1.IPSWITCHS.CMM> ddl@danlan.*com (Dan Lanciani) writes:
[snip]
| Quote: | One thought I had is that the characteristics of the line are somehow
changing (water infiltration?) over time and a retrain operation makes
things right for a while--until something changes again. But that seems
a bit of a stretch. Is there any kind of common equipment in the switch
such that I might be getting a bad line termination device at times and
not at others?
|
I had a similar situation about five years ago, also using a BitSurfr. I
was down to a five percent or so ping throughput on the few times I could
connect (which, with error correction and retries, meant it was a lot
worse than even that.).
Voice was completely hopeless.
I put in a standard trouble report and things were ok a few days later.
One of the techs I saw in the building complex told me that the other tech
(so we're getting into a FOAF, fwiw...) was the one who had worked on it,
and they finally figured out that my ISDN circuit had been physically
interrupted. To wit, at the central office it had been switched from a
straigh-to-me copper pair over to one that passed through a univerasl
(voice channel) SL(i)C in the CO and back out of a SL(i)C in my complex
and hten back to copper wire. Oh, and that they were amazed anything had
gone through it...
I dunno if that was true, and it has absolutely nothing to do with your
specific problem, except to say that ISDN glitches do occur...
--
_____________________________________________________
Knowledge may be power, but communications is the key
dannyb@panix.com
[to foil spammers, my address has been double rot-13 encoded] |
|
| Back to top |
|
|
Loren Amelang
Guest
|
| Posted:
Tue Nov 23, 2004 1:45 am Post subject:
Re: ISDN BRI line noise problem |
|
|
On Sun, 21 Nov 2004 22:22:40 +0000 (UTC), danny burstein
<dannyb@panix.com> wrote:
| Quote: | In <1331230@news1.IPSWITCHS.CMM> ddl@danlan.*com (Dan Lanciani) writes:
[snip]
One thought I had is that the characteristics of the line are somehow
changing (water infiltration?) over time and a retrain operation makes
things right for a while--until something changes again. But that seems
a bit of a stretch. Is there any kind of common equipment in the switch
such that I might be getting a bad line termination device at times and
not at others?
I had a similar situation about five years ago, also using a BitSurfr. I
was down to a five percent or so ping throughput on the few times I could
connect (which, with error correction and retries, meant it was a lot
worse than even that.).
Voice was completely hopeless.
I put in a standard trouble report and things were ok a few days later.
One of the techs I saw in the building complex told me that the other tech
(so we're getting into a FOAF, fwiw...) was the one who had worked on it,
and they finally figured out that my ISDN circuit had been physically
interrupted. To wit, at the central office it had been switched from a
straigh-to-me copper pair over to one that passed through a univerasl
(voice channel) SL(i)C in the CO and back out of a SL(i)C in my complex
and hten back to copper wire. Oh, and that they were amazed anything had
gone through it...
|
When I read the original post I was thinking to ask if a SLC was
involved. There are ISDN line cards that slide in in place of voice
cards to let ISDN through the SLC. Whenever I see such problems it is
the line card in the SLC nearest me that is at fault. Sometimes the
error rate will be fine if either single channel is in use, but nasty
if both are up. Sometimes 56K will pass but 64K won't. Rarely does
power cycling or disconnecting and reconnecting my equipment help.
To answer the original question, my (SBC, NI-1) line leaves layer 2 up
24/7 and only cycles layer 3 with my usage. There is no significant
change in the underlying connection between "calls". But recently I
saw a log from another user with a different provider (Verizon?) where
dropping layer 2 at the end of each call appeared to be standard
procedure. If my system drops layer 2, there are several variables and
opportunities for problems introduced by having to re-connect layer 2,
and the behavior of the resulting connection may vary.
I have no idea how to see this activity on your BitSurfr - is there
any option to log "q921" events?
Loren |
|
| Back to top |
|
|
Dan Lanciani
Guest
|
| Posted:
Tue Nov 23, 2004 5:52 am Post subject:
Re: ISDN BRI line noise problem |
|
|
In article <vri4q0dddsvq559c9tl124r3br2n26r901@4ax.com>, loren@pacific.net (Loren Amelang) writes:
| On Sun, 21 Nov 2004 22:22:40 +0000 (UTC), danny burstein
| <dannyb@panix.com> wrote:
|
| >In <1331230@news1.IPSWITCHS.CMM> ddl@danlan.*com (Dan Lanciani) writes:
| >[snip]
| >>One thought I had is that the characteristics of the line are somehow
| >>changing (water infiltration?) over time and a retrain operation makes
| >>things right for a while--until something changes again. But that seems
| >>a bit of a stretch. Is there any kind of common equipment in the switch
| >>such that I might be getting a bad line termination device at times and
| >>not at others?
| >
| >I had a similar situation about five years ago, also using a BitSurfr. I
| >was down to a five percent or so ping throughput on the few times I could
| >connect (which, with error correction and retries, meant it was a lot
| >worse than even that.).
| >
| >Voice was completely hopeless.
| >
| >I put in a standard trouble report and things were ok a few days later.
| >
| >One of the techs I saw in the building complex told me that the other tech
| >(so we're getting into a FOAF, fwiw...) was the one who had worked on it,
| >and they finally figured out that my ISDN circuit had been physically
| >interrupted. To wit, at the central office it had been switched from a
| >straigh-to-me copper pair over to one that passed through a univerasl
| >(voice channel) SL(i)C in the CO and back out of a SL(i)C in my complex
| >and hten back to copper wire. Oh, and that they were amazed anything had
| >gone through it...
|
| When I read the original post I was thinking to ask if a SLC was
| involved.
There was none in the past. The repair guy who came out said nothing had
changed, but I have not walked the wire myself to be sure...
| There are ISDN line cards that slide in in place of voice
| cards to let ISDN through the SLC. Whenever I see such problems it is
| the line card in the SLC nearest me that is at fault. Sometimes the
| error rate will be fine if either single channel is in use, but nasty
| if both are up. Sometimes 56K will pass but 64K won't. Rarely does
| power cycling or disconnecting and reconnecting my equipment help.
Yeah, that's what is confusing me. Disconnecting and re-connecting the
line often noticeably changes the situation.
| To answer the original question, my (SBC, NI-1) line leaves layer 2 up
| 24/7 and only cycles layer 3 with my usage. There is no significant
| change in the underlying connection between "calls". But recently I
| saw a log from another user with a different provider (Verizon?) where
| dropping layer 2 at the end of each call appeared to be standard
| procedure. If my system drops layer 2, there are several variables and
| opportunities for problems introduced by having to re-connect layer 2,
| and the behavior of the resulting connection may vary.
Interesting. I don't think they are intentionally dropping layer 2, but
my doing so certainly changes things. It would be nice to know what the
echo cancellation parameters look like each time.
| I have no idea how to see this activity on your BitSurfr - is there
| any option to log "q921" events?
I've seen postings suggesting that the Bitsurfr has fairly extensive diagnostic
capabilities, but Motorola does not document them. I will look through the
postings that I've saved, but I remember being somewhat frustrated by the
lack of detail. :)
Dan Lanciani
ddl@danlan.*com |
|
| Back to top |
|
|
|
|
|
|
FAQ
Memberlist
Register
Profile
Log in to check your private messages
Log in