Just wondering
Bookmark this page:
 Yahoo!
 Google
 Windows Live
 del.icio.us
 digg
 Netscape
|
|
|||||||||||||
|
Posted by Jimbo on September 23, 2007, 4:15 am
Please log in for more thread options this. I was in charge of a Nortel network of over 100 PBX's connected through OC48 / DS3 and T1's. There were ATM switches connecting the optical portions of the networks. The site I am wondering about was main site>OC48>T1.
I had one campus that always popped up with a T1 error about every 2 to 6
days(sorry, over the years I don't remember the specific error code). It was never enough to interrupt service but still it was a problem. I logged into the NIU (don't tell Qwest) and didn't see any errors. I did a head-to-head test with T-bird 2209's over the weekend many times and found the pipe to be clean. I couldn't figure it out. I tried swapping out TMDI cards and even swapped out CSU's with other CSU's that worked fine at other sites. Still, I kept getting errors. I even had Qwest swap out cabnets and NIU's and my self swapped out cables to the CSU and from the CSU to the OPT 11c. Still errors. I tried different bonding grounds and different brands of CSU's, even tried AC power to the CSU vs. DC. Still errors. I finally ran a cable directly from the NIU to the OPT 11C (no CSU) and all the errors were gone for good. I'm confused. What the hell is going on here? The CSU was configured the same as all other sites and there were other OPT 11C's in the network by the same install tech with the same configuration. We even set the thresholds the same but still had errors only at this site. I even made sure that the pairs were with in a couple of db (22 for one and 23 for the other) and feed out of the same xboxes. TDR tests showed them to be the same. Qwest replaced both doublers (only one per T1). The software RLSE was 23 at the main site and 25 at the problem site. However, we had 3 other sites with RLSE 25 and no problelms from them. Like I said, I am no longer in charge of this site but as far as I understand, they are still running directly from NIU to OPT 11c to keep from getting the errors. Just a note, the customer was influentual enough to get Nortel to listen to every word he had to say and he had a LOT to say if it ment he could get some free service from us/Nortel. He got a computer log of EVERY error that the switches spit out and always wanted us to explain every one of them every day. On another install job of over 20 Opt11c's (all through Passpost, back in 2004) we had a problem that some T1's would not come up or would come up then go down. With T-birds we found the T1's to be OK most of the time but the only way that we could keep a connection through the passports was if we could get the Qwest tech to install a regen-NIU that keeps the db at 0. That was towards the start of my career in telecom but, still I think of it today and wonder why? T1's seem simple in the text books but, seem to have strange errors that can not be explained by logic when you get into the field. I would love to here any comments about any part of this post. | |||||||||||||
|
Posted by Bill on October 1, 2007, 8:32 pm
Please log in for more thread options >I am no longer in the telecom biz but wonder if anyone knows how to explain
>this. > I was in charge of a Nortel network of over 100 PBX's connected through > OC48 / DS3 and T1's. There were ATM switches connecting the optical > portions of the networks. The site I am wondering about was main > site>OC48>T1. > I had one campus that always popped up with a T1 error about every 2 to 6 > days(sorry, over the years I don't remember the specific error code). It > was never enough to interrupt service but still it was a problem. I logged > into the NIU (don't tell Qwest) and didn't see any errors. I did a > head-to-head test with T-bird 2209's over the weekend many times and found > the pipe to be clean. I couldn't figure it out. I tried swapping out TMDI > cards and even swapped out CSU's with other CSU's that worked fine at > other sites. Still, I kept getting errors. I even had Qwest swap out > cabnets and NIU's and my self swapped out cables to the CSU and from the > CSU to the OPT 11c. Still errors. I tried different bonding grounds and > different brands of CSU's, even tried AC power to the CSU vs. DC. Still > errors. I finally ran a cable directly from the NIU to the OPT 11C (no > CSU) and all the errors were gone for good. I'm confused. What the hell is > going on here? The CSU was configured the same as all other sites and > there were other OPT 11C's in the network by the same install tech with > the same configuration. We even set the thresholds the same but still had > errors only at this site. I even made sure that the pairs were with in a > couple of db (22 for one and 23 for the other) and feed out of the same > xboxes. TDR tests showed them to be the same. Qwest replaced both doublers > (only one per T1). The software RLSE was 23 at the main site and 25 at the > problem site. However, we had 3 other sites with RLSE 25 and no problelms > from them. > Like I said, I am no longer in charge of this site but as far as I > understand, they are still running directly from NIU to OPT 11c to keep > from getting the errors. Just a note, the customer was influentual enough > to get Nortel to listen to every word he had to say and he had a LOT to > say if it ment he could get some free service from us/Nortel. He got a > computer log of EVERY error that the switches spit out and always wanted > us to explain every one of them every day. This is actually a common problem. Some CSU's don't like to have DSX levels on their receive port. The regenerators are overdriven, and it can result in bit errors that can be interpreted by the TE (terminal equipment) as BPVs/Framing errors/CRC errors depending on a few other variables. The fact that the switch is working with no CSU means that NIU receive towards the switch is pretty close to DSX levels. If you either put a pad on the receive towards the CSU, or configure the NIU for no regen, the CSU will likely be happy - the switch won't work without the CSU, but that's OK. >
> On another install job of over 20 Opt11c's (all through Passpost, back in > 2004) we had a problem that some T1's would not come up or would come up > then go down. With T-birds we found the T1's to be OK most of the time but > the only way that we could keep a connection through the passports was if > we could get the Qwest tech to install a regen-NIU that keeps the db at 0. > That was towards the start of my career in telecom but, still I think of > it today and wonder why? The inverse problem of above.... The Passports likely did not have regeneration (generally refered to as a CSU - they really do more than just provide a loopback) on the NI port. DSX levels are within ~5 db of 0 dbdsx (~ 6 volts peak-to-peak into ~100 ohms). Lower than that, and things may not work without a regenerator. > T1's seem simple in the text books but, seem to have strange errors that
> can not be explained by logic when you get into the field. > > I would love to here any comments about any part of this post. DS-1 is actually pretty simple, but there are a few issues that can trip you up. | |||||||||||||
|
Posted by Jimbo on October 2, 2007, 1:10 am
Please log in for more thread options Thanks Bill for the reply. It seems like such a simple fix and you would
think just about anyone who has worked for a few years with DS-1's would know this. However, between me, some other techs (some of whom have been senior Nortel certified techs for decades) and the few Qwest T-1 techs never came up with the ides of putting a pad in line. | |||||||||||||
| Similar Threads | Posted |
| Just wondering | September 23, 2007, 4:15 am |