T1 trouble over microwave
telcotech
service from the LEC to a point-to-point microwave system. The T1s
are currently delivered over a DS3 WAN and muxed down to DS1
interfaces. They're used for private networking between multiple
PBXs. The data side of the point-to-point microwave system is in
service and has been operating error-free for some time.
When I move a given span from the DS3 mux equipment to the microwave
equipment, the span constantly goes up and down. PBX equipment alarm:
Loss of frame (Red) due to loss of line synchronization. No alarm
indications on the Kentrox T-ServII CSUs, and I'm able to loop the CSU
on each end.
If I use a spare Kentrox CSU on one end of the span rather than the
original CSU, the span comes up and stays up – error-free for days.
Identical settings on each CSU. I've tested and experienced the same
results on two spans.
*All equipment configured for ESF and B8ZS.
*All CSUs: Kentrox T-ServII. Equipment and Network-side interface:
DSX-1
*Microwave radio interface: DSX-1. Equalization set for appropriate
cable length.
*Cable length between CSU and radio: Near-end: 500' Far-end: 375'
*Cable type between CSU and radio: 22awg ABAM
*Cable termination: RJ48X (either on RJ48X patch panel or biscuit
jack.)
Can anyone comment on why the spare CSU might perform differently than
the others? Same manufacturer, model, and settings as the others.
Does the performance of these things vary that much?
Any comments or recommendations appreciated. Thank you.
Kenneth A. Becker
Umm.. I've got a thought. When you had the T1's running through the
telco, nickle bet says that the telco was also supplying your
timing. Hence, all of your equipment would have been "line timed"
from the telco clock. That clock is accurate to parts in 10**-12 or so.
Now you're running your T1's through the microwave gear. Bet that
your CSU's and such are still line-timed. Hence, you have no primary
reference in your network, and all the PLL's in all the CSU's are
locked to nothing in particular (well, each other, but that's the
point) and they're drifting around in frequency. Get them off
frequency enough and you get what you're seeing.
Idea: Sit down with a piece of paper and draw a diagram of your
network. Pick some centrally located node in that network and put it
into free-run. Make sure that timing for everybody else is derived
off that node one way or another and your problems should go away.
Ken Becker
f
o
d
d
e
r
f
o
r
t
h
e
i
c
l
u
d
e
d
l
i
n
e
s
b
o
t
.
Steve Pinkston
Kentrox T-SERV II CSUs from unit to unit. That unit is a very old,
very stable design.
Ken's comment about clocking is relevant, but the T-SERV II CSU is
completely transparent to framing, line code AND clocking.
There are four LEDs on the CSU that will tell you a little about line
problems:
LOS means NO SIGNAL coming in to the Network Interface.
AIS means ALL-ONES signal coming in to the Network Interface.
KA means NO SIGNAL coming in to the Terminal Interface.
BPV means BIPOLAR VIOLATIONS (line errors) coming in to the Network
Interface.
You've made a good choice by using ABAM cable. Near-end crosstalk
(NEXT) is unlikely to be your problem.
You probably do not need to set the Equalization on the microwave
radio DSX-1 interface (or the CPE) to anything other than 0-133 ft.
The CSU has a very wide input sensitivity range (0 to -27 dB) on both
the Network and Terminal interfaces. However, it is possible to
overload its input by setting the Equalization on the connected
equipment to too high a setting.
Equalization boosts the level of the signal and increases its
high-frequency content, in other words, "louder and more treble."
Good luck
- Steve Pinkston
On Fri, 13 Jun 2003 09:48:50 -0400, "Kenneth A. Becker"
<kab1n...@no.lucent.spam.com> wrote:
>telcotech wrote:
>> I'm having difficulties attempting to migrate T1s from leased-line
>> service from the LEC to a point-to-point microwave system.
(..snip..)
telcotech
Thanks for the reply.
The DS3 from the telco should not have been providing timing. Here's
some additional information:
*PBX for far end of T1 provides timing for span. That PBX has a PRI
span from the telco and is loop-timed. Timing for all other
private-networking spans is derived from the ISDN PRI timing.
*Given the above, the timing type for the equipment on each end of the
test span is system timing.
What I find interesting is that I have one CSU that allows me to move
the test span onto the microwave equipment and keep that span up for
days -- error free. However, if I simply move the span over to the
microwave (using the same CSUs on each end), the span goes up and
down.
Thanks.
telcotech
Thanks for your response. I was having a hard time believing that the
T-Serv II's were going vary in their performance.
The problem was what you alluded to in your post. The EQ on the
radios needed to be adjusted (down to 0-133'.) Having done this for
both microwave radios, and having set both CSU's LBO to -7.5, the
problem was corrected. I have one span up now 2 days error-free.
I must have just had a special spare CSU that was able to cope with
the overload situation.
Thanks again.
BTW: I knew your name looked familiar. Early on in trouble-shooting
this issue, I read through your extended demarc white paper. Good
stuff.
Kenneth A. Becker
> Ken -
>
> Thanks for the reply.
>
> The DS3 from the telco should not have been providing timing. Here's
> some additional information:
>
> *PBX for far end of T1 provides timing for span. That PBX has a PRI
> span from the telco and is loop-timed. Timing for all other
> private-networking spans is derived from the ISDN PRI timing.
>
> *Given the above, the timing type for the equipment on each end of the
> test span is system timing.
>
> What I find interesting is that I have one CSU that allows me to move
> the test span onto the microwave equipment and keep that span up for
> days -- error free. However, if I simply move the span over to the
> microwave (using the same CSUs on each end), the span goes up and
> down.
>
> Thanks.
>
>
Hmm.. This gets more interesting.
However, the >>most<< interesting thing was the original complaint
that your equipment mentioned:
> PBX equipment alarm:
> Loss of frame (Red) due to loss of line synchronization.
Loss of line synchronization could mean several things, I suppose.
Being a sync maven, this meant to me (when I first read it) that the
PBX was seeing an off-frequency input signal, so far off frequency
that it couldn't stay locked and gave you that Red alarm. This isn't
as hard to detect, in hardware, as one might think. Since T1 lines
keep track of slips, if you got a lot of slips, it means that
somebody's off frequency. Since every slip causes a burst of errors
on the T1 payload going into CGA for excessive slips would be expected.
What I would also expect is that the CSU at the far end would be
having the same trouble - that is, excessive slips. Of course if
that alarm was disabled somehow and you were only running voice
traffic, you might not notice. If you're running packets across the
link you might not notice in that direction, either, if enough
packets get through.
Now, if >>that's<< happening, really, that means that either the PBX
is 'way out in left field (unlikely) or that the remote CSU is. The
fact that using a spare CSU would indicate to me that the spare CSU
just happens to do a better job in staying roughly on frequency.
A piece of T1 test equipment will quickly show who's running at what
frequency.
The other possibility that crosses my mind is that the signal level
or some such from the CSU to the PBX is at some low level. Not
enough, say, to cause LOS - but it might be just enough to flip bits
at the PBX from time to time. Now, a T1 line receiver wants to have
a certain number of transitions or else it's going to have trouble
keeping its receive clock locked onto the right phase. There's a
requirement if you're not running B8ZS or some such that you have to
have no more than 15 zero's in a row (or something like that). If
you don't do that the receive clock wanders and you get BPV and bit
errors and the like. I suppose that an intelligent enough T1 line
receiver could keep track of the number of zero's in a row and give
you the alarm you saw if it saw too many. On the other hand: Does
the CSU you're using have a switch setting for B8ZS or not? If one
CSU is set for B8ZS and the other isn't, then the one that's not
might give you the alarms you saw at the PBX. That is, the possible
settings for a T1 line is AMI or B8ZS. A line that's set for B8ZS
can receive AMI data without a problem, so long as the 1's density
is up to snuff. If you put a T1 inside a DS3 that kind of problem
tends to go away, since the DS3 has its own coding scheme for too
many zero's in a row.
Ken Becker
telcotech
As I was bringing up my issue on this newsgroup, I was also speaking
with the manufacturer of the microwave equipment. My questions
finally made their way to the proper resource on their support team.
The response that I received was as follows:
*I shouldn't be using a CSU between the microwave T1 interface and my
PBX.
*Because I'm using the CSU, it's possible that I'm overloading the
microwave T1 interface.
I will be following up with this person. In the meantime, can anyone
elaborate on this issue?
*Microwave T1 digital interface: DSX-1 (meets AT&T Pub 62411)
*Kentrox T-Serv II network interface: DSX-1 input and output.
*Kentrox T-Serv II equipment interface: DSX-1 input and output.
*PBX T1 line card interface: DSX-1 (I'm confirming this with
manufacturer.)
This information was included in my last post:
*Cable type: ABAM
*Cable length: 320 - 480' (depending on location)
*CSU location: approx. 10' from PBX.
I will be following the microwave manufacturer recommendation and will
be connecting PBXs to the microwave equipment without the CSUs inline.
Will also get a BERT on-site to look at the signal -- want to see if
I'm overloading the anything (either the radios or the CSUs.)
Any insight you could provide would be appreciated. Thank you.
telcon...@yahoo.com (telcotech) wrote in message news:<754aa19f.03061...@posting.google.com>...
> I'm having difficulties attempting to migrate T1s from leased-line
> service from the LEC to a point-to-point microwave system. The T1s
> are currently delivered over a DS3 WAN and muxed down to DS1
> interfaces. They're used for private networking between multiple
> PBXs. The data side of the point-to-point microwave system is in
> service and has been operating error-free for some time.
>
> When I move a given span from the DS3 mux equipment to the microwave
> equipment, the span constantly goes up and down. PBX equipment alarm:
> Loss of frame (Red) due to loss of line synchronization. No alarm
> indications on the Kentrox T-ServII CSUs, and I'm able to loop the CSU
> on each end.
>
> If I use a spare Kentrox CSU on one end of the span rather than the
> original CSU, the span comes up and stays up ? error-free for days.
Macy Hallock
+---------------
> I will be following the microwave manufacturer recommendation and will
> be connecting PBXs to the microwave equipment without the CSUs inline.
> Will also get a BERT on-site to look at the signal -- want to see if
> I'm overloading the anything (either the radios or the CSUs.)
Check the signal level, timing and quality. A Sunrise FT1 or similar
TTC test set should work nicely.
I don't see a CSU as helping or hindering the operation of this facility
as long as it is setup properly. Note that some CSU models will provide
very useful diagnostic info via a console (craft) serial port.
--
Macy M. Hallock, Jr. N8OBG 216.241.7166 fax:216.241.7522
APK Net, Inc. 1621 Euclid Ave. Suite 1230 Cleveland, OH 44115 USA
###
If the DSX signal is set to 100 feet, the most common setting, then the CSU
is needed to drive the signal to the transmission equipment which is over
400 feet away. The CSU also provides other protections for use in in the
PSTN. One of the most important functions of the CSU is as a part 68
protection device, which is not required on a private facility. So, maybe
checking the settings on the DSX ports of the PBX and Microwave to insure
that they are set to drive the signal the correct distance. Oversetting
this could manifest as a hot signal. The CSU is going to put out a signal
with higher power as well, and would require another CSU in front of the
Microwave to convert it back to a DSX-1 from DS1. I ramble. Sorry. Just
try it without the CSU's, and set the DSX-1 interface for 450 feet.
--
----------
Andy Kauffman
Telecom Consultant
www.ahk.com
"Macy Hallock" <ma...@apk.net> wrote in message
news:bdpjkd$9...@newstaff.apk.net...
telcotech
Thanks for the reply. One question: You mentioned that the CSU is
going to put out a signal with higher power. Do you expect that
output signal on the network side of the CSU will be DS1? My
understanding from the manufacturer is that the CSU network interface
output is DSX-1 (without line build-out.)
Regardless, I have had a chance to test another span and was able to
bring it up both with and without the CSU inline. EQ at the radio:
0-133'. EQ at the PBX: 0-133'. LBO at the CSU: -7.5dB.
Thanks.
"###" <ne...@ahk-no-spam.com> wrote in message news:<%7KMa.89462$R73.10620@sccrnsc04>...
###
user side produces a DSX-1 signal. The LBO on the network interface
"attenuates" the incoming signal to accommodate a signal that is to hot.
The settings are generally -7 db, -15 db, -22 db, etc.. The signal is
designed to travel 6,000 feet. The DSX-1 signal LBO increases the power.
Settings are generally 100 feet, 250 feet, 400 feet, and 650 feet (all
approximate). This signal can travel up to 650 feet without repeaters.
Unless the device you were using was a DSX-1 repeater, this is how it should
have worked.
--
----------
Andy Kauffman
Telecom Consultant
www.ahk.com
"telcotech" <telcon...@yahoo.com> wrote in message
news:754aa19f.0307...@posting.google.com...
Steve Pinkston
>Yes, the network interface, or WAN port of a CSU produces a DS1 signal. The
>user side produces a DSX-1 signal. The LBO on the network interface
>"attenuates" the incoming signal to accommodate a signal that is to hot.
>The settings are generally -7 db, -15 db, -22 db, etc.. The signal is
>designed to travel 6,000 feet. The DSX-1 signal LBO increases the power.
>Settings are generally 100 feet, 250 feet, 400 feet, and 650 feet (all
>approximate). This signal can travel up to 650 feet without repeaters.
>Unless the device you were using was a DSX-1 repeater, this is how it should
>have worked.
This is ALMOST correct.
LBO on a CSU attenuates the OUTBOUND (telco-facing) signal, not the
inbound signal. In fact, LBO does not just attenuate the signal, but
also rolls off the high frequency content proportionally to how they
would be rolled off by copper cable.
LBO on a CSU is often used to mitigate the effects of near-end
crosstalk in UTP extended demarc wiring. By reducing the outbound
signal amplitude, less of it will be induced into the inbound pair.
With 0.0 dB of LBO, the output of a CSU should be right at 0.0 dB
(DS-1) plus or minus a few tenths of a dB.
Equalization -- as noted by "###" -- is applied to the CPE-side port
of a CSU, and often on other DS-1 type ports. The default setting for
most ports of this type is 0-133 feet. Setting the port to higher
settings not only increases the signal level, but increases the high
frequency content to compensate for the attenuation of the copper
cable.
- Steve Pinkston
steve_pinkston_at_kentrox.com
Replace _at_ with @ to send me e-mail.