We have a telco colo PP with a couple of Widebank28 muxes aggregating
T1's for PA-MC-T3 cards. Something keeps upsetting the controller cards
in just one of the muxes. Sometimes cycling power will bring the mux back
online. Sometimes the controller card in the mux is actually fried.
We've replaced the entire mux twice. We've moved it off telco provided DC
power and onto CAC AC/DC power supplies, and the problem has continued, so
I'm fairly confident that one of the T1s is actually frying the mux.
Anyone seen this before?
Is there such a thing as a rack mount multi-T1 surge supressor, preferably
one that could be easily wired in series between the wire wrap DSX pannel
and the mux?
Jon Lewis | I route
Senior Network Engineer | therefore you are
Atlantic Net |
http://www.lewis.org/~jlewis/pgp for PGP public key
"Eat sushi frequently". - Avi
inet (AT) inet-access (DOT) net is the human contact address.
list (AT) inet-access (DOT) net is the list posting address.
See below URL for subscribe/unsubscribe and list options:

Wed, 23 Aug 2006, Jawaid Bazyar wrote:

You should be able to get lightning/surge suppressor fusible elements for
copper wire from a number of companies; Qwest uses Marconi a lot. I think
Lucent also makes them.

Basically inside a telco C all wire coming from the outside goes through
this type of power surge fusible element.

If that's the case [with Sprint^WEmbarq], then in theory, we shouldn't
need to do our own surge supression unless there's somehow voltage being
introduced inside the C, or their supressor isn't supressing.

I'm wondering if maybe PTS ring voltage is somehow making it into one of
our T1's. I was told today, when a circuit is canceled, they leave the
circuit intact, and just break it at the remote smartjack. It kind of
worries me that we have "disconnected" T1s with [the way we have the muxes
wired up] no easy way to break the connections on our end, which are still
connected to dormant copper loops.

Jon Lewis | I route
Senior Network Engineer | therefore you are
Atlantic Net |
http://www.lewis.org/~jlewis/pgp for PGP public key

"Eat sushi frequently". - Avi
inet (AT) inet-access (DOT) net is the human contact address.
list (AT) inet-access (DOT) net is the list posting address.
See below URL for subscribe/unsubscribe and list options:

Wed, Aug 23, 2006 at 04:18:20PM -0400, Jon Lewis wrote:
Wed, 23 Aug 2006, Jawaid Bazyar wrote:

>You should be able to get lightning/surge suppressor fusible elements for
>copper wire from a number of companies; Qwest uses Marconi a lot. I think
>Lucent also makes them.
>
>Basically inside a telco C all wire coming from the outside goes through
>this type of power surge fusible element.

If that's the case [with Sprint^WEmbarq], then in theory, we shouldn't
need to do our own surge supression unless there's somehow voltage being
introduced inside the C, or their supressor isn't supressing.

I'm wondering if maybe PTS ring voltage is somehow making it into one of
our T1's. I was told today, when a circuit is canceled, they leave the
circuit intact, and just break it at the remote smartjack. It kind of
worries me that we have "disconnected" T1s with [the way we have the muxes
wired up] no easy way to break the connections on our end, which are still
connected to dormant copper loops.

I haven't measured it, but I am told that T1s have quite a bit of
voltage and current themselves, as they must be capable of power inline
repeaters.

/*
Jason Philbrook | Midcoast Internet Solutions - Internet Access,
KB1IJ | Hosting, and TCP-IP Networks for Midcoast Maine
http://f64.nu/ | http://www.midcoast.com/
*/

"Eat sushi frequently". - Avi
inet (AT) inet-access (DOT) net is the human contact address.
list (AT) inet-access (DOT) net is the list posting address.
See below URL for subscribe/unsubscribe and list options:

Chris Adams wrote:
upon a time, jp <jp (AT) saucer (DOT) midcoast.comsaid:
>I haven't measured it, but I am told that T1s have quite a bit of
>voltage and current themselves, as they must be capable of power inline
>repeaters.

I haven't directly measured it, but a T1 definately _feels_ like more
voltage than PTS. :-)

<>

Avi: hi, our T1 is dead
Bell: can you give us the circuit number?
Avi: well, the bill says it is #<whatever>, but the box on the wall is
labelled #<othernumber>
Bell: please describe for me why you think it is down.`
Avi: because it is dead. it is down. it is not functioning

Bell: No, no, it's uh, it's resting

Avi: Look matey, I know a dead T1 when I see it, and I'm looking at one
right now.

Bell: No no it's not dead, it's, it's restin'! Remarkable line, the T1,
idn'it, ay? Beautiful bits!

Avi: The bits don't enter into it. It's stone dead.

Bell: Nonononono, no, no. It's resting.

Avi: All right then, if it's resting, I'll wake it up. <Yells into wall
socketELL, MISTER TELLY T1! I've got a lovely fresh news spool for
you if you show.

Bell: <Sends voltage down lineSee- it's working!

Avi: No it isn't- that was you.

Bell: I never!

Avi: Yes, you did!

Bell: I never did anything.

Avi: <Pulls wire out of socket and beats it against the side of the
server.ELL T1!!!! TESTING TESTING TESTING! This is your 9 o'clock
cron job download! <Watches bits physically falling on floor>

Now that's what I call a dead T1!

Bell: No, no, you stunned it.

Avi: STUNNED!?

Bell: Yeah- you stunned it just as it was starting to accept packets.
T1s stun easily, major.

Avi: Um, now look mate. I've had enough of this. This T1 is definitely
deceased, and when we hooked it up not more than a day ago, you assured
me that its total inability to transmit was due to it being tired and
shagged out after trying to pull across all of
last night.

Bell: Well, it's , ah, it's just pining for the ARPANET.

Avi: PINING FR THE ARPANET?!? What kind of talk is that? Look, why did
it drop every packet as soon as we plugged it into this Sparc?

Bell: The T1 prefers to drop packets from a Sparc. Remarkable line,
isn't it. Lovely bits.

Avi: Look, I took the liberty of dissassembling the smartjack last
night, and found that the only reason that it looked like it was working
it because you'd wired the lights to a battery.

Bell: Well, of course. If we hadn't, it might have burned the whole
smartjack out-VM VM!

Avi: VM? Mate, this line wouldn't voom if you put four million volts
through it. It's bleeding demised!

Bell: No, no, it's pining.

Avi: It's not pining, it's passed on! This line is not working! It has
ceased to transmit! Bereft of data, it lies in peace. If you hadn't
wired the lights it would have been recycled. It carries 0MB/sec! It's
no longer functional- it's shuffled off the backbone and gone to meet
its maker. THIS IS AN EX-T1!

Bell: Well, we better fix it then. <Pause'Fraid we're all out of
working T1s.

Avi. I see, I see. I get the picture

Bell: I've got an AL disk and a 2400 baud modem.

Avi: Pray, does it work?

Bell: Nnnnot really

"Eat sushi frequently". - Avi
inet (AT) inet-access (DOT) net is the human contact address.
list (AT) inet-access (DOT) net is the list posting address.
See below URL for subscribe/unsubscribe and list options:

Thu, 24 Aug 2006, Henry Yen wrote:

The T1's (well, actually HDSL 4-wire) that I've had installed that were
over repeatered spans were -137VDC, according to the telco installers.

I bet that's the voltage they were putting on the circuit (and it must
have been a long circuit) on its way out of the C, not what you could
measure at (and certainly not after) the smart jack. I'd also expect that
with colo inside the C (as in our case), the current loop starts and ends
on some distance from our rack, and we shouldn't see anywhere near that
sort of voltage on our end of the circuit.

Jon Lewis | I route
Senior Network Engineer | therefore you are
Atlantic Net |
http://www.lewis.org/~jlewis/pgp for PGP public key

"Eat sushi frequently". - Avi
inet (AT) inet-access (DOT) net is the human contact address.
list (AT) inet-access (DOT) net is the list posting address.
See below URL for subscribe/unsubscribe and list options: