atcomsystems.ca/forum Forums Wiring - Telecom Carriers - Trunks & Services T1's, CSU's and DSU's Adtran TA 908 does not recover T1 after a service outage or power failure

I have, what I think anyway, a strange problem. We use TA 908 gateways and this is the only site where this problem occurs.

We have a T-1 providing ISDN-PRI service to a site. The gateway is used to connect the PRI to a PBX that can only be configured with SIP trunks. This part works perfectly.

The problem is when the gateway experiences a power failure or the T1 goes out of service, the T1 interface will go to Yellow Alarm. It can be recovered by using CLI to loop up then loop down the network. My saving grace is I have remote access to the gateway so I can quickly get the site up but I shouldn't have to do this.

I opened a case with Adtran and the engineer said that if we put a hard loop on the interface and it goes to up, this eliminates the hardware. We did this loopback including the cable that connects to the NID and it passed.

After much effort, we were able to arrange a vendor meet with the LEC. We showed the technician our trouble. The technician said there is nothing wrong with the circuit because when he connects his T-BERT it goes right to up. He was also kind enough to "back to back" with the gateway. When we did this, the gateway went to UP instead of sticking on Yellow Alarm. I pointed out that if the gateway was bad we should have expected the same result of the interface going to Yellow Alarm but this was brushed aside by, "I can connect my tester to the NID and it comes right up."

I can't disagree with the logic but I am also not 100% convinced there is not an issue upstream from us. I am going to try another gateway just to try and save face with the customer. I don't care if I look dumb, I just want this fixed.

Any ideas? My thinking is that a Yellow Alarm on my interface means something upstream from me is not working. I find it strange that looping then not looping the T1 clears the trouble immediately. I am further encouraged by the fact that connecting the the T1 test set goes right to UP. Since this is a physical layer issue it must (at least in my mind) be a service provider problem.

We have tried a new cable from the gateway to the NID, the NID is about 25 cable feet from the gateway. The entrance terminal for the site is in the same room as the NID and is approximately 25 cable feet away with no splices or obvious defects, and the LEC technician was kind enough to try changing the card in the NID. I was not able to convince them to try the cold spare unfortunately.

My goal is to get this customer to work like every other T1 I have ever worked with. The challenge will be proving to the LEC it is not CPE equipment. I am keeping an open mind though and, as explained earlier, will try a new gateway.

Thank you in advance for your assistance.

Mike Meyer

You're receiving a Yellow, meaning the far end equipment is seeing no signal. You loop it up and after releasing it clears, which to me means the far end equipment will sync to itself and releasing the loop causes your equipment to re-sync to the far end. I'm betting if you unplug from the NID wait a few seconds and plug it back in it will also re-sync. The best way to trouble shoot this would be to see if there is a signal from your equipment when you see the yellow by bridging into the circuit or putting a monitor on it. If you see a signal yet still receive the yellow the trouble is beyond you, a monitor could also be put at the NID.

I assume the LEC has remote testing capabilities on this, they could also see what signals they are getting when you see the yellow. Of course that takes the cooperation of the LEC.

My first thought is that the LEC is not sending the transition reset bit. The only way to see that is if the LEC comes back out, connects the T-Berd in the system in bridge mode, then have the CO yellow alarm the circuit and then have the CO restore the circuit. The tech should see the transition reset bit and your system should respond with an ack bit. Then do the same from the site and see, when your system sends the transition bit, the CO acks the sent bit with an ack bit.

Some systems use the same bit for the transition bit and the ack bit, so you may not see two different bits, but you should see the CO and your system sending bits to each other. If only ONE is missing, the circuit will not come up automatically.

Rcaman

Gentlemen - Thank you for your responses. I thought I knew T1 really well but, I have to admit, I am not too familiar with some of this.

We have a Sunrise T10 Tester so I should be able to insert the tester into the circuit in bridge mode and make some of these observations myself yes?

In what I am hoping is the simple fix, I am going to try replacing the unit just to see. If the new unit works (which I honestly don't expect), then I can bring the gateway back to our shop and tinker with it there.

Thank you again and, anything you can offer to further advance my knowledge, is appreciated!

Mike

Often a Central Office switch can send a Yellow Alarm when the carrier is removed from service. A switch can generate a Yellow Alarm if error thresholds are exceeded. It depends on how the carriers are provisioned in the CO switch.

Well it's either the equipment on either end or the circuit itself. Have you stress tested the T-1? Sometimes the circuit will go out and not reset by itself.

How is the circuit configured ? (AMI, B8ZS etc.). Test the hell out of the circuit, loop up the far end and then run the tests - All 1s, then All 0s (if it's B8ZS) then 3-24, then QRS.

If that passes, have the LEC go Head-to-Head with you (tester to tester, no equipment at each end. Prove the circuit and then work on the equipment.

Sam

If your tester can read the patterns, then, yes, you should be able to see the data.

I had several PRI circuits that would do weird things like you are describing. The circuits originated in Verizon territory and terminated in Windstream territory. It was a nightmare to deal with as Verizon looked good and, at the terminating Co for Windstream, the circuit looked good. After I brought my Fireberd to the meet, it was discovered that, somewhere in Windstream's circuits, the circuit went from B8ZS to AMI and then back again. How did it work at all? Go figure. It took several months and a LOT of vendor meets to finally nail the problems down. I have to congratulate the Verizon tech in the Bala Cynwyd special Circuits group that worked on the trouble ticket all the way through and did identify the Windstream problem.

Don't give up. In my case, Windstream kept sending techs with multimeters to test the circuit. It wasn't until I put a Fireberd on the circuit and proved to Windstream's foreman the problem was in the circuit not my equipment.

Rcaman

Once again - Thank you for all the advice. I will let you know what transpires. ~ Mike

As promised, a status update. Replacing the unit resulted in no difference as I expected. My colleague and I were racking our brains. We combed through the config and nothing.

I noticed in the configuration commands for the T1 interface there was a command to disable Remote Alarm Indication. I figured we have nothing to lose trying this and disabled RAI. The T1 came up and would recover on LOS and power failure.

I ran this past the Adtran engineer and he called it a "band-aid that might hold". He still believes there is something faulty upstream from us or in the CO but I don't think we are going to convince the LEC to check those things based on the last vendor meet. I will just have to be happy with this solution. Main thing is the customer will be up after some type of failure not related to the gateway.

Big thank you to Justbill, Rcaman, and Silversam. Your advice was in the back of my mind and, while I didn't break out the tester yet, it definitely helped fire off the brain cell to prompt trying a command related to yellow alarms. ~ Mike