T1 Background Information - HDSL, SDSL

Comparing hicap t1 to HDSL and SDSL

T1 is an old but extremely well supported and reliable technology that was standardized in the late 60s. T1 as changed over time to not specify the layer 1 implementation at the electrical level so much as it specifies the service and interface that must be delivered. Thus, many creative companies will deliver a "t1" that is emulated using DSL technology. A standard t1 uses 2 pair of twisted copper wire (4 wires) and B8ZS line coding (Binary 8 zero substitution). B8ZS allows t1 technology to be resistant to other pairs running other line codings (crosstalk) in the same copper plant. T1s require specially conditioned lines (no taps or analog signal boosters) and they require digital repeaters every 6000 feet of cable. A pure t1 is also called Hicap service (High Capacity Digital service by telco engineers). In the mid 90s many telcos began using HDSL to deliver t1 loops. HDSL can go up to 12,000 feet w/o a repeater. HDSL uses ISDN's 2B1Q line coding--2 bits encoded into a four-state (quaternary) symbol. This is the exact same line coding that a Basic Rate ISDN line (2x64Kbps B channels + 1x16Kbs signalling channel - D channel) uses. HDSL became so popular that in many areas over 70 to 80 percent of all t1s are delivered using HDSL technology. The smart jacks (line power and termination cards) convert the 2B1Q coding to B8ZS such that our CSUs that we plug the t1 into still understand the service as a standard t1 facility. HDSL also uses 2 pairs of copper wire, though it is not quite as tolerant as standard t1 to crosstalk from other pairs, as I understand it. The literature conflicts on this point somewhat. The HDSL 1 spec is what is widely deployed. The HDSL 2 spec states that the same capabilities of HDSL 1 can be delivered over 1 pair of wire. A lot of the literature talks about SDSL as being basically HDSL 2.

SDSL Differences

Now, SDSL is where things diverge. SDSL is rate adaptive. That means that if line conditions are not favorable the bit rate will be lowered to try to maintain at least some data flow. HDSL is *not* rate adaptive. The other big difference is that SDSL again uses only one pair. This makes it far cheaper to provision, and it also makes it far more likely to crash or have downtime figures that exceed HDSL and certainly t1. SDSL is also much more vulnerable to crosstalk. This means that as new facilities are added a working SDSL line may suddenly not work so well b/c of proximity in the bundle to some other service. It is critical to understand the weaknesses of SDSL when choosing providers. A hicap or HDSL t1 to a tier 1 provider (Worldcom/UUnet, Sprint, ATT, C&W, Qwest are the teir 1s) will usually cost at least $1000/month. Some exceptions might occur if your business is in a park where fiber to the building is in place. If that is the case, often times there is a local provider willing to run MAN (metropolitan area network access) ethernet, which offers greater speed and inexpensive connections (VPN server with a firewall and an ethernet NIC). Many tier 2 providers (regional ISPs, and national ISPs who lease fiber from tier1) sell SDSL connections. Many of these companies have also filed for chapter 11 in the last year because these connections were priced at a level that made profits impossible. Any t1 that is sold for less than $800 a month is almost assuredly an SDSL t1. Caveat emptor. One final note: ADSL also uses 1 pair. This is what drives residential DSL. The bit rates are adaptive and the uplink is usually 20 to 30 percent capacity of the downlink.

AIS Alarm

Special notes on the infamous "All 1's condition"

When an upstream data source is not functioning, the DACs t-1 switch
that recieves the "empty" signal will generate all 1's to the
customer. The customer gear will show an alarm indication and AIS alarm will show up on the router. The t-1 will appear to be down to the end customer; no data can be transferred across the line.

Execute a show controller t1 or show service-module command depending on the type of t1 card in the cisco router. In the case where you have multiple t1s coming into a router, make sure to read the controller or interface number so you can match up the correct t1. Cisco routers number the first controller or interface at ZERO not ONE always.

Output will look similar to the below; in this case 4 t1 interfaces exist, and the 4th - service module serial 0/3/0 is having the AIS:

Interface Serial0/3/0
Module type is T1/fractional
    Hardware revision is 1.2, Software revision is 20080312,
    Image checksum is 0x4144A7, Protocol revision is 0.1
Transmitter is sending AIS.
Receiver has no alarms.

The Transmitter is sending AIS is the key. This tells us that the t1 local loop between the telco switch and the customer prem is working. The problem is NOT on the local loop, but is instead in the network - somewhere upstream.

Why AIS is important

Often telco techs will "test" only the local loop part of the t-1, from the customer prem to the first telco DACs. This is because the most common part of a t1 circuit to have a problem is the local loop. These local loop tests will pass of course, because in AIS condition the local loop itself is fine. Once the local loop has been tested, an inexperienced tech will push the problem back on the "customer equipment" to avoid further work/expense on the part of the telco. Pushing the support team to escalate the issue is often easier when you mention AIS as it forces the techs to start to look at their switching paths end to end instead of focusing on you (the customer) having a "dead router".

How does AIS get generated

The upstream DACS, Frame-Relay or MPLS Switch, SMDS Hub, or ATM switch port is down, misconfigured, or disconnected from the network (partitioned)
and not sending signal from the upstream network to the local
Central Office or customer. As soon as the port is swapped, or
re-enabled, all 1's will clear, and layer 2 will come up. Some CSUs
can read the bits coming in off the line. If you have such a CSU, be
on the lookout for all 1s, as this means the repair call should be
directed farther up the food chain than the local central office.

Always ook for the AIS alarm in the service module or show controller output. This is
fairly common, so be sure you can identify AIS quickly.

From the Cisco Web Site:

AIS: alarm indication signal. In a T1 transmission, an all-ones signal
transmitted in lieu of the normal signal to maintain transmission
continuity and to indicate to the receiving terminal that there is a
transmission fault that is located either at, or upstream from, the
transmitting terminal.

If you are receiving an AIS alarm on the router csu this means that the
last mile is o.k. (demarc, smart jack, router csu/dsu are all ok), but the
problem is somewhere upstream (Local Bell Central Office, MCI, etc.)

AIS Alarms are sometimes caused by telco grooming

Telco DACS troubleshooting

Situation Presents as:
1. layer 1 is clean
2. Telco can loop smart jack, and they can loop the router WIC or controller built-in T1 CSU
3. but they cannot do a payload loopback (individual channels loopback)

Problem: Telco has a DACs assignment issue.

Sometimes a t1 loop must pass through multiple DACS at multiple
Central Offices at multiple companies (Verizon, ATT, Qwest etc). If one of
the DACS is setup to map the wrong channels, then these symptoms will
result. Telco tech must be encouraged to check *all* DACS.

Notes below are specific to MCI/Worldcom now Verizon frame/t1 networks

Normally T1 and fractional T1 services run on an aggregate channelized ds3 when the
service is 384K and greater. If the service is less than 256K, it runs on
standard t1 service. 256K service can be configured either way. Techs
will sometimes have to move a circuit termination point from one switch to
another if spare port capacity on the DACS is not available when an upgrade
is being performed.

Grooming teams will come behind install and upgrade jobs to clean pvcs,
and relocate circuits.

T1 Troubleshooting on Cisco Routers with t1 cards

Pt to pt 1 troubleshooting:

Some point to point T1's still need both CPE equipments to have
clock set to line as the network provides the clock. The signal strength on the t1 line should be around 0 or a few db negative. A mismatch of clocking between our hardware and the central office crossover equipment could cause a signal to be high, such as -17db or so. The normal CO tester won't look at this sort of thing; they only do the loopback pattern testing. You will need to ask them about the signal strength of your line is you suspect this sort of problem.

DEBUGGING T1 issues

Before troubleshooting any aspect of a connectivity issue (e.g., ISDN, CAS, modem) you should always verify the physical integrity of the T1 line. You should always check the status of the T1 controllers and verify you are not receiving any errors. show controller T1 x will give the snapshot of the T1 physical layer status. There should not be any framing errors, Slips, or line code violations.

Following is the sample output of show controller T1 0 and what to look at:

show controllers t1 0

T1 0 is up.

Applique type is Channelized T1

Cablelength is long gain36 0db

No alarms detected.

Version info of slot 0: HW: 4, Firmware: 16, PLD Rev: 0

Manufacture Cookie Info:

EEPROM Type 0x0001, EEPROM Version 0x01, Board ID 0x42,

Board Hardware Version 1.32, Item Number 73-2217-5,

Board Revision B16, Serial Number 09356930,

PLD/ISP Version 0.0, Manufacture Date 18-Jun-1998.

Framing is ESF, Line Code is B8ZS, Clock Source is Line Primary.

Data in current interval (8 seconds elapsed):

0 Line Code Violations, 0 Path Code Violations

0 Slip Secs, 0 Fr Loss Secs, 0 Line Err Secs, 0 Degraded Mins

0 Errored Secs, 0 Bursty Err Secs, 0 Severely Err Secs, 0 Unavail Secs

The main items to look at in the above output is

* The status of the line
* Alarms
* Linecode and Pathcode violations
* Slip Secs

The line status will tell us if the T1 is either up, down, or administratively down. The Alarms section is very important and it will tell us what type of problem maybe present on the line. The presence of any alarms indicates a serious problem on the line.

It is recommended whenever you encounter a T1 that is in an alarm state that you verify the framing and linecoding parameters are configured correctly. Please refer to the show controller t1 commands in the Command Reference to find out all the possible values for the alarm state.

A common message you will see in the alarm field is "receiver has loss of frame." Some routers will also report a 'loss of frame' even when it should be a "loss of signal." So, make sure whenever you receive these errors that the T1 signal is present and the framing is correct.

Another message you might receive is "receiver is getting AIS." This means the receiver is getting an alarm indication signal (blue alarm). This is a framed or unframed all-ones signal in both SF and ESF formats transmitted to maintain transmission continuity. This is typically seen when the far-end CSU has lost its terminal side equipment. The "receiver has remote alarm" indicates the presence of a yellow alarm. This means the downstream CSU is in a loss-of-frame or loss-of-signal state. Therefore, the remote CSU has a red alarm.

The "transmitter is sending remote alarm" indicates that the local CSU has detected either a loss-of-frame or loss-of-signal condition. This indicates that the local controller has a red alarm. This message will be accompanied by a "receiver has loss of frame." Always verify framing and T1 signal when troubleshooting this problem.

If any of the above highlighted fields doesn't contain zeros, than here are some of the possibilities what might causing the physical problem. Following is the brief explanation of these fields.
Line Code Violations

Indicates the occurrence of either a Bipolar Violation (BPV) or Excessive Zeros (EXZ) error event.

A BPV error event for an AMI-coded signal is the occurrence of a pulse of the same polarity as the previous pulse.

A BPV error event for a B8ZS is the occurrence of a pulse of the same polarity as the previous pulse without being a part of the zero substitution code. An EXZ is the violation of the pulse density requirement.
Path Code Violations

Indicates a frame synchronization bit error in the D4 and E1-noCRC formats, or a CRC error in the ESF and E1-CRC formats.
Line Errored Seconds (LES)

A Line Errored Second, according to T1M1.3, is a second in which one or more Line Code Violation error events were detected.

In the T1M1.3 specification, near end Line Code Violations and far end Line Errored Seconds are counted. For consistency, we count Line Errored Seconds at both ends.
Slip Seconds

A Controlled Slip Second is a one-second interval containing one or more controlled slips.
Errored Seconds (ES)

For ESF and E1-CRC links an Errored Second is a second with one or more Path Code Violations OR one or more Out of Frame defects OR one or more Controlled Slip events OR a detected AIS defect.

For D4 and E1-noCRC links, the presence of Bipolar Violations also triggers an Errored Second.

This is not incremented during an Unavailable Second.
Bursty Errored Seconds (BES)

A Bursty Errored Second (also known as Errored Second type B) is a second with fewer than 320 and more than 1 Path Coding Violation error events, no Severely Errored Frame defects and no detected incoming AIS defects. Controlled slips are not included in this parameter.

This is not incremented during an Unavailable Second.
Severely Errored Seconds (SES)

A Severely Errored Second for ESF signals is a second with 320 or more Path Code Violation Error Events, one or more Out of Frame defects, or a detected AIS defect.

For E1-CRC signals, a Severely Errored Second is a second with 832 or more Path Code Violation error events or one or more Out of Frame defects.

For E1-noCRC signals, a Severely Errored Second is a 2048 LCVs or more.

For D4 signals, a Severely Errored Second is a count of one-second intervals with Framing Error events, or an OOF defect, or 1544 LCVs or more.

Controlled slips are not included in this parameter. This is not incremented during an Unavailable Second.
Severely Errored Framing Second (SEFS)

An Severely Errored Framing Second is a second with one or more Out of Frame defects or a detected AIS (Alarm Indication Signal) defect.
Degraded Minutes

A Degraded Minute is one in which the estimated error rate exceeds 1E-6 but does not exceed 1E-3 (see G.821 [15]).

Degraded Minutes are determined by collecting all of the Available Seconds, removing any Severely Errored Seconds grouping the result in 60-second long groups and counting a 60-second long group (a.k.a., minute) as degraded if the cumulative errors during the seconds present in the group exceed 1E-6. Available seconds are merely those seconds which are not Unavailable, as described below.
Unavailable Seconds (UAS)

Unavailable Seconds (UAS) are calculated by counting the number of seconds that the interface is unavailable. The DS1 interface is said to be unavailable from the onset of 10 contiguous SESs, or the onset of the condition leading to a failure (see Failure States). If the condition leading to the failure was immediately preceded by one or more contiguous SESs, then the DS1 interface unavailability starts from the onset of these SESs. Once unavailable, and if no failures present, the DS1 interface becomes available at the onset of 10 contiguous seconds with no SESs. Once unavailable, and if a failure is present, the DS1 interface becomes available at the onset of 10 contiguous seconds with no SESs, if the failure clearing time is less than or equal to 10 seconds. If the failure clearing time is more than 10 seconds, the DS1 interface becomes available at the onset of 10 contiguous seconds with no SESs, or the onset period leading to the successful clearing condition, whichever occurs later. With respect to the DS1 error counts, all counters are incremented while the DS1 interface is deemed available. While the interface is deemed unavailable, the only count that is incremented is UASs.

Cisco T1 PRI Troubleshooting

Contents

Introduction

Using the show isdn status Command

Using the debug q921 Command


Introduction

When troubleshooting a Primary Rate Interface (PRI), ensure that the T1
is running properly on both ends. If Layer 1 problems have been resolved,
look for problems on Layers 2 and 3. Use the show controller t1
command to verify that the configuration of the line matches that of the
remote end. Ensure that the framing, line coding, and clock source are
configured correctly.

Using the show isdn status Command

The show isdn status command displays a summary of all ISDN interfaces.
It also displays the status of Layers 1, 2, and 3. Complete the following
steps to check the status of the layers:

  1. Verify that Layer 1 is in the ACTIVE state. The status of Layer 1 should
    always be ACTIVE unless the T1 is down.

  2. If the show isdn status command output indicates that Layer
    1 is DEACTIVATED, then there is a problem with the physical connectivity
    of the T1 line. If the line is administratively down, use the no shutdowncommand to restart the interface.

  1. Ensure that Layer 2 is in the MULTIPLE_FRAME_ESTABLISHED state. This is
    the desired state for Layer 2, indicating that Layer 2 frames are being
    exchanged and Layer 2 initialization has finished.

  2. If Layer 2 is not in the MULTIPLE_FRAME_ESTABLISHED state, use the
    show controller t1
    EXEC command to diagnose the problem. For more information,
    see the T1
    Alarm Troubleshooting
    document.

    Since the show isdn status command displays a summary of the
    current status, it is possible that Layer 2 is bouncing up and down despite
    indicating a MULTIPLE_FRAME_ESTABLISHED state. Use the debug isdn q921
    command to verify that Layer 2 is stable.

    Following is an example of show isdn status output:

      maui-nas-03#show isdn status Global ISDN Switchtype = primary-5ess
      ISDN Serial0:23 interface dsl 0, interface ISDN Switchtype = primary-5essLayer 1 Status: ACTIVE Layer 2 Status: TEI = 0, Ces = 1, SAPI =
      0, State = MULTIPLE_FRAME_ESTABLISHED Layer 3 Status: 5 Active Layer
      3 Call(s) Activated dsl 0 CCBs = 5 CCB:callid=7D5, sapi=0, ces=0, B-chan=9,
      calltype=DATA CCB:callid=7D6, sapi=0, ces=0, B-chan=10, calltype=DATA CCB:callid=7DA,
      sapi=0, ces=0, B-chan=11, calltype=DATA CCB:callid=7DE, sapi=0, ces=0,
      B-chan=1, calltype=DATA CCB:callid=7DF, sapi=0, ces=0, B-chan=2, calltype=DATA
      The Free Channel Mask: 0x807FF8FC ISDN Serial1:23 interface dsl
      1, interface ISDN Switchtype = primary-5ess Layer 1 Status: ACTIVE
      Layer 2 Status: TEI = 0, Ces = 1, SAPI = 0, State = TEI_ASSIGNED
      Layer
      3 Status: 0 Active Layer 3 Call(s) Activated dsl 1 CCBs = 0 The Free Channel
      Mask: 0x807FFFFF Total Allocated ISDN CCBs = 5

    Notice that T1 0 (whose D channel is Serial 0:23) has Layer 1 as ACTIVE
    and Layer 2 as MULTIPLE_FRAME_ESTABLISHED indicating that the signaling
    channel is functioning correctly and is exchanging Layer 2 frames with
    the telco switch. The D channel (Serial1:23) for T1 1 has Layer 1 ACTIVE,
    but Layer 2 is TEI_ASSIGNED. This indicates that the PRI is not exchanging
    Layer 2 frames with the switch. Use the show controller t1 x
    command to troubleshoot. Refer to the T1
    Troubleshooting
    flowchart for more information

Using the debug q921 Command

The debug isdn q921 command displays data link layer (Layer 2) access
procedures that are occurring at the router on the D-channel.

Ensure you are configured to view debug messages by using the logging
console
or terminal monitor command.

Note: In a production environment, verify that console logging
is disabled by using the show logging command. If logging is enabled,
the access server might intermittently stop working when the console port
is overloaded with log messages. Enter the no logging console command
to disable logging.

Note: If debug isdn q921 is turned on and you do not receive
any debug outputs, place a call or reset the controller to get debug outputs.

Complete the following steps to ensure that the data link layer access
procedures are occurring at the router on the D-channel:

  1. Verify that Layer 2 is stable by looking for messages in the debug output.
    If the line is bouncing up and down, output similar to the following will
    appear:
  2. Mar 20 10:06:07.882: %ISDN-6-LAYER2DOWN: Layer 2 for Interface Se0:23, TEI 0 changed to down
    Mar 20 10:06:09.882: %LINK-3-UPDOWN: Interface Serial0:23, changed state to downMar 20 10:06:21.274: %DSX1-6-CLOCK_CHANGE: Controller 0 clock is now selected as clock source
    Mar 20 10:06:21.702: %ISDN-6-LAYER2UP: Layer 2 for Interface Se0:23, TEI 0 changed to up
    Mar 20 10:06:22.494: %CONTROLLER-5-UPDOWN: Controller T1 0, changed state to up
    Mar 20 10:06:24.494: %LINK-3-UPDOWN: Interface Serial0:23, changed state to up
  3. Verify that only service access point identifier (SAPI) messages appear
    on both the transmit (TX) and receive (RX) sides. For example:
  4. Mar 20 10:06:52.505: ISDN Se0:23: TX -> RRf sapi = 0 tei = 0 nr = 0
    Mar 20 10:06:52.505: ISDN Se0:23: RX <- RRf sapi = 0 tei = 0 NR = 0
    Mar 20 10:07:22.505: ISDN Se0:23: TX -> RRp sapi = 0 tei = 0 NR = 0
    Mar 20 10:07:22.509: ISDN Se0:23: RX <- RRp sapi = 0 tei = 0 NR = 0
    Mar 20 10:07:22.509: ISDN Se0:23: TX -> RRf sapi = 0 tei = 0 NR = 0
    Mar 20 10:07:22.509: ISDN Se0:23: RX <- RRf sapi = 0 tei = 0 NR = 0
  5. Verify that asynchronous balanced mode extended (SABME) messages do not
    appear. These messages indicates that Layer 2 is trying to reinitialize.
    The messages usually appear when poll requests (RRp) are transmitted and
    there is no response from the switch (RRf), or vice versa. Following are
    examples of SABME messages:
  6. Mar 20 10:06:21.702: ISDN Se0:23: RX <- SABMEp sapi = 0 tei = 0
    Mar 20 10:06:22.494: ISDN Se0:23: TX -> SABMEp sapi = 0 tei = 0

    If SABME messages appear, complete the following steps:

    1. Use the show running-config command to ensure that isdn switch-type
      and pri-group timeslots are configured correctly. Contact your Service
      Provider for the correct values.
    2. To change the isdn switch-type and pri-group settings, enter
      the following commands:
    3. maui-nas-03#configure terminal
      maui-nas-03(config)#isdn switch-type primary-5ess
      maui-nas-03(config)#controller t1 0
      maui-nas-03(config-controlle)#pri-group timeslots 1-24
  7. Ensure the D-channel is up using the show interfaces serial number:23
    command, where the number is the interface number.

  8. If the D-channel is not up, use the no shutdown command to
    bring it up. For example:

    maui-nas-03(config)#interface serial 0:23
    maui-nas-03(config-if)#no shutdown
  9. Ensure encapsulation is PPP. If not, use the encapsulation ppp command
    to set encapsulation. For example:
  10. maui-nas-03(config-if)#encapsulation ppp
  11. Ensure the interface is in loopback mode. Loopback should be set only for
    testing purposes. Use the no loopback command to remove loopbacks.
    For example:
  12. maui-nas-03(config-if)#no loopback
  13. Power cycle the router.

Verizon T1 Timing

Summary - covers confusion typical with pt to pt circuits

Timing in Verizon is a function of the terminating equipment. If the
circuit is point-to-point where the customer terminates both ends,
the customer provides timing with the "one side internal one side
line" setting. DS1 (T1) is synchronous. The best way to have
synchronous service is to have only one timing reference, or
*perfectly* identical multiple timing sources. Nothing is perfect, so we want
to go with one timing reference.

When Verizon terminates one end of the circuit, like in ISDN PRI,
Frame Relay, ATM and etc. or channelized T1 service then the CPE
should be set for line time.

On Verizon designed point-to-point circuits there are only two
terminations, one at each end. The "middle" network elements are
always designed to be transparent.

So if only one of the terminating CSU's should provide the timing
reference, what does the other one do?

The line setting tells the CSU to check the 1st, 8th, 16th, and 24th
frame bits (on ESF framed circuits) and calculate or derive the
reference. It then makes adjustment to the transmit data returning
to the far end. There is no "clock pulse" per se on a payload DS1.

It really doesn't matter how many CO's your circuit is transported
across, the settings are the same because this configuration deals
with the "payload" level. We always clock the "transport" level.

An example would be your point-to-point is DS1 rate, you clock the
DS1 we clock the transport HDSL and DS3 or above rates (like thru
SONET).

Full T1 Timing Brief

The FCC-specified framed DS1 rate is 1544000 bps (1.544 Mbs) +/- 50 bps.
The payload or usable data rate is 1536000 bps (1.536 Mbs).

DS1 service is full duplex and synchronous. This means that transmit and
receive speeds are the same and that all DS1 network elements operate at
the same relative speed.

A timing or clock source actually generates the DS1 rate and this source
is a function of Data Circuit Terminating Equipment (DCTE or DCE) in the
Verizon network.

A more common name for customer-owned DCE is CSU/DSU or WIC.

On DS1 circuits where the customer terminates both ends of the circuit,
called a point-to-point configuration, the customer CSU is responsible for
providing the timing source. This source is built-in by the equipment
manufacturer and required by the FCC.

Commonly the hub or otherwise hi-end CSU is set for INTERNAL source and the
remote or otherwise lo-end CSU is set for LINE or LOOP source.
If both CSU's are comparable, then the CSU most convenient physically or
one with remote access is usually set to INTERNAL source and the least
convenient CSU is set for LINE or LOOP source.

If both CSU's are set for INTERNAL source, then errors will be noted sooner
rather than later in most circumstances, sometimes immediately.

If both CSU's are set for LINE or LOOP source, then the circuit may work
properly for years until some internal event (heat, cold, static, power
surge, etc) damages the CSU timing chip or external event (Verizon network
failure) causes the circuit to become unstable. The circuit may not
synchronize and become stable until the CSU source configurations are
corrected.

NOTE: Sometimes the INTERNAL source setting will become corrupted and the
CSU will behave as if set for LINE or LOOP. If the circuit is showing slips
or acting erratically it is always a good idea to refresh or reset the
INTERNAL source by setting it temporarily to line, loop, or external and
then BACK to INTERNAL.

On DS1 circuits where Verizon terminates one end of the circuit like ISDN,
Fast Packet, or channelized service originating in a Verizon switch, D4
channel bank, or 0-1 digital cross-connect, the CSU terminating the end
user side of the circuit should be set for LINE or LOOP source.

Where multiple DS1's terminate in a facility, the timing source
configuation may be more complicated.

If multiple DS1s terminate in a single device it is common for one DS1 to
be set as a PRIMARY source and the others as ALTERNATE sources. When a
circuit is designated as a PRIMARY source it is said to be a FACILITY
source.

Rules for FACILITY sources:

Rules for FACILITY sources:

1. Dont use a payload carrying DS1 as a source.
2. Dont use a DS1 transported via SONET as a source.
3. Source Verizon terminated circuits to a Verizon source.
4. Everybody ignores rules 1, 2, and 3 so experiment.
5. If experimentation fails to provide desired results, seek professional
help.

(Verizon can provide a non-payload DS1 for use as a facility source, please
refer to a sales engineer for more information)

Sometimes customers will be certain that Verizon provides timing on
point-to-point circuits. Verizon does, at the TRANSPORT level. Verizon does
not provide timing at the PAYLOAD level on point-to-point circuits.

An example of this payload/transport level issue on a DS1 point-to-point
circuit is the DS1 is payload or user data level and the HDSL, DS3, or STS1
etc. through the Verizon network is the transport level.

More about LINE and LOOP sourcing..

The CSU device set for LINE or LOOP source examines multiple frame bits in
the received signal and calculates the rate. It then takes this calculated
rate and compares it to the current transmit rate. If the result is outside
a pre-determined limit, a correction is calculated and applied.

This process takes time and can introduce error in some instances. On a
discrete or single DS1 this small error is usually not noticeable.

At installations with multiple DS1s where the customer has specified
PRIMARY and ALTERNATE facility sources, remember that the PRIMARY source is
actually LINE or LOOP sourced and that the rest of the DS1s probably are
EXTERNALLY sourced to the PRIMARY source.

Errors introduced by the LINE or LOOP timing process WILL BE amplified and
distributed by the PRIMARY FACILITY source to any circuits or other devices
using it as a timing reference.

A small error in the PRIMARY source or timing distribution system can cause
significant problems!

For example in an installation with 8 DS1s, and DS1 0 as primary source and
DS1 1 thru 7 as alternates a single small slip at the Primary DS1 0 may be
cascaded through the rest of the circuits with increasingly severe effects.

DS1 1 may show 10 slips, DS1 2 may show 20 slips, DS1 3 may show 50 slips,
and so on.

NOTE: DCE may not use these specific sourcing terms. INTERNAL refers to a
source internal to the DCE device. LINE or LOOP refers to the received DS1
signal. DCE may define EXTERNAL source as LINE or LOOP, or FACILITY.

t1 Tariff Allowances (max number of allowed errors)

Verizon statement on ES


The tariff error allowance for DS1 service is 1080 ES per day or 10 UAS
per day. The modern networks are good enough now so that we try to make
them error free but that is a goal not a standard.

The modern networks are self-healing with automated diagnostics and
protection switching. The protection switching is required to execute
in less than 50ms. During these switching operations a small burst of
CVP could be experienced.


Richweb Opinion:

Richweb would suggest opening a ticket with Telco when the ES (errored seconds)
is greater than 40 or 50 ES a day. Most copper-based T1s will have anywhere from 3 to
10 ES a day, especially when the day in question involves a temperature fluctuation (hot to cold or cold to hot - copper cables expand and contract). A t1 that is delivered on fiber all the way
into the DEMARC will often never have any errors at all, except during grooming (usually in the middle of t he nite). Copper t1s that are delivered into a building via an exterior NID (pedastal) that is exposed to the weather can often exhibit unreliable performance in rainy weather as well. If you have a copper t1 that seems to very in quality with the weather then open a ticket and have the telco tech check the NID for stray wires, or connections that are too loose or too tight (heat related expansion can stretch a cable to become loose, or get to close to another cable causing cross talk or signal scrambling / shorting).

Applications running across a t1 WAN can run fine with occasional ES; users will usually never notice less than 30 or 40 ES a day unless they come in 10 to 15 sec bursts. Its when you have a t1 that is experiencing 60+ ES a day when your applications may to start to have some issues. At 300 to 500 ES, you probably will get user complaints, and notice apps stalling on
occasion.

UAS (Unavailable Seconds) are much more severe - essentially its a second where the entire framing is lost and the line is totally down. A 5 second burst of UAS can result in a minute or more for the line to resync and apps to recover. A t1 taking any UAS at all should be called in for repair.

t1 terms

Demarc

The demarc (short for demarcation point) is the last place (furthest point from the carrier operated Central Office) on a circuit that the carrier is responsible. With a T1 this is typically some sort of intelligent device (NIU or smartjack) that responds to remote loop commands and has a RJ48 jack that allows a patch core to be inserted to connect to a customer CSU or t1 wic (cisco routers). Some LECs hand off t1s on 66 blocks but this is generally a thing of the past now.

MPOE

The MPOE is the Minimum Point of Entry of a building. (Sometime called the MPOP- Minimum Point of Presence) This is typically the room in which the LECs Central Office cables are terminated. This is where any non-extended demarcs will be.

NIU

The NIU (Network Interface Unit) is a multifunction device that does surge protection and responds to loop commands, etc) i.e. a smartjack. The smartjack will almost always be located at the MPOE. Even if you get an extended demarc they won't move the it.

Try to make an extended demarc part of the order from the very beginning. Essentially you want the provider to be responsible for the whole thing, up to the 2 meter cable to your equipment. The carrier will try to test to the NIU, then the CSU. If they can run clean to the NIU but not the CSU they'll call you back to verify power- then dispatch someone to test from the extended demarc. This dispatch can be charged back to the customer if the customer is found to have incorrectly configured the router.

 

More info:

http://www.dcbnet.com/notes/9611t1.html

T1 Voltages and Crosstalk

Question:
The cisco router built-in CSU by default transmits at 0 dBm. It can be set for
-7.5 or for -15. If telco is sending at -7.5 and the cisco is sending at 0 dbm
could this cause a problem ?

Answer:
Yes, this can cause near end cross talk but only in longer DEMARC extensions.
The file below discusses this in detail:

Also, if interference is coming from a 3rd, foreign source that is leaking
energy, like a pots line, or even a high voltage electrical circuit that
is generating emf noise then many transmission problems can result.

Verizon Answer about differing signal strength and crosstalk:

The actual DS1 rate signal is preset to -7.5 dBm. It was common in
the past to set them for 0 dBm, but somewhere in FCC regulations in
tiny, tiny print the -7.5 setting is prescribed and Verizon tries to
adhere to it.

The 0 dBm from the Cisco should be fine, the HTU-R is pretty tough.
The only time I've noticed an issue with CPE power levels is when
they are connected thru a particular type of smartjack to a fiber
mux.

Some of the low-speed cards on the muxes are sensitive to input
levels. So it is a circuit-by-circuit kind of issue depending on the
design.

Crosstalk or other interference can be an issue. Generally you want
to run signal cables perpendicular not parallel to power cables, and
in the case of DS1 signal cables as far from 56-64K (like ISDN BRI)
signal cables.