View source for Line monitoring

{{stub}}
In security systems where intercom stations are used in emergency situations only, maybe just once per year, it’s extremely important that the equipment works when it’s finally used.  To make a security system useful in practice, it’s necessary that the equipment has built-in fault monitoring

== ASLT subscriber lines ==

=== Wire test ===

==== Error detection ====
The AlphaCom’s line boards ASLT (4-wire TouchLine station) and ATLB (2-wire telephone) monitors station wires continuously based on the fact that certain currents flows during the stations when everything is OK, and that either wire break or short-circuit creates abnormal currents that are detected and reported.  
The line boards has an on-board microcontroller which performs the line monitoring along with other tasks.  This means fast response (errors are reported after 2-4 seconds), and there is no load on the functional software that runs on the AMC board.

NOTE:  The ASLT board determines automatically whether a station has display or not depending on the ab-wire current (display stations draw a constant 10 mA  to power the display).  Some line error conditions may be detected to be wrong station type rather than a line problem.   When the exchange is in doubt, a so-called “extended check” is performed - see detailed description below.

==== Error state:  Station down ====
When a station wire is detected to be faulty, the station is assumed useless and is blocked by the exchange software.  It’s no longer possible to dial to the station, i.e. attempts are rejected with a failure tone.  (Group Calls will be connected, as there is no check of each station’s error state due to group set-up speed).
Other station tests does not cause station down, they are just reported.

==== ASLT ====
ASLT performs continuous line monitoring, which can NOT be disabled.
Trick:  If you need a “portable” station that can be plugged in and out without causing line error reports, you can connect resistors close to the exchange to form a “dummy station”, and then modify the portable station so that it lacks the corresponding resistors (contact support for details).

=====TouchLine station without display=====
* Break of all wires (station un-plug) is detected, but needs 1 to 3 minutes before report.
* The ab-wires (speaker) encodes the functions Off-hook, On-hook, M-key, C-key where off-hook is close to wire break, and C-key is close to short-circuit.  <br\>
- Off-hook is not possible to separate from break, and is reported if the handset is off for more than one hour and no dialling takes place.
(AMC feature, from version 07.20. Timer can be modified using TST,  .ex_profile.timeouts.off_hook_too_long = 36000 (0x8ca0), new value e.g. 200 for 20 seconds.).<br\>
- M key in idle for more than 10 minutes (from AMC 07.12/07.60: 1 hour 50 min) causes “station down”, then change to display-station mode.<br\>
- Short-circuit close to the station may not be detected on long lines (1 km) as the wire loop resistance limits the current to the normal C-key area.
* The cd-wires (microphone, keyboard) encodes the functions OPEN, PRIVATE, digit-key where OPEN is close to wire break, and digit-key is close to short-circuit.  Single error on cd is not reported.
* Note that some cross-connects like c-wire shorted to b-wire is not possible to detect as the line currents change very little.
=====TouchLine station with display=====
* In this case the ab-wire is supplying power for the display, by adding a constant current of 10 mA to the ab-wires.  
<br\>
- Break is here detected immediately as there is a guaranteed 10 mA current in normal Off-hook.
<br\>
- Short circuit on long lines is even more difficult to detect here as the C-key current is  higher.
* From AMC 07.20:  You can select a special station type “Interguard”, which means that line monitoring shall report deviations from display station limits immediately.  Station type CRM 3 & 4 is included also.

===== ASLT Extended Line Check =====
In order to distinguish between the otherwise indistinguishable "station disconnected" and "non-display station with off-hook" states, a so-called "extended check" is initiated every 23 seconds if the station is presumed disconnected or in off-hook idle. 

PS: The station is still considered idle when passively listening to audio programs, simplex (program) conferences or group calls.

The extended check utilises 4 closely spaced polarity reversals on the CD wire pair (double flash in station LED) to do this. As a side effect there will be double flashes on a station every 23s when the handset is left (or has fallen) off-hook so long that the handset dial tone has timed out. This is not a bug -- it is a feature! Consequently, this is done for display stations also.

When the station is actively used or a change of station is being detected, no such checks are done. All activities will, however time out and return to idle, allowing the plugging in or out of a non-display station with handset off to be detected eventually, no matter what.  <br\>
Exception: Conversation does by default not time out -- time-out must be programmed explicitly.

===== ASLT error codes =====
The ASLT reports a code value to pinpoint a problem.  Note that only the first detected error is flagged, and that there may be additional errors on the same station.<br\>
80 (1000 0000) = AB Shorted<br\>
40 (0100 0000) = AB Open<br\>
20 (0010 0000) = CD Shorted<br\>
10 (0001 0000) = CD Open



=== Tone test ===

==== General ====
To monitor a greater part of the total system than LM1 does, it’s possible to run a “tone test”.  A tone is then sent from the exchange via the line card to the station speaker.   This  acoustic tone is picked up by the station microphone, and returned via the line card to the exchange.  Thus a complete test of the audio path is performed, i.e. both wires, electronics and electro-acoustical devices.

Tone test can only be run on typical loudspeaking intercom stations, as it must be able to generate a tone which the microphone can pick up clearly (well-defined difference between “silence” and “tone”).  This fact has some implications:

* The test should be run in low-traffic periods of the day both to reduce background noise, and to avoid disturbing staff.
* It is possible to skip stations that does not have sufficient acoustic crosstalk, like handset stations, goose-necks witch close-talk microphone, etc.  Stations that are used frequently, like guard stations, are “monitored” by humans, and may also be skipped.
* Some exchange functions like group call and high priority program/conference will be active while the speech channel is tested.  Stations using there features should be skipped, or the test must be run outside normal working hours.
* Each station is tested once.  If the test fails, the test is repeated a few times with a delay in between each test.  If a repeated test gives correct result, the station is assumed OK.  A station is reported faulty only if all repeated tests fails.


=== Day-to-day use ===
==== Automatic ====
The tone test is run automatically once per hour during night time (configurable), at low priority.  The test uses 3 seconds per tested station, i.e. 20 stations per minute.  A 550 liner uses some 30 minutes for a full test.

The exchange’s log port will print one line per each station that changed state (faulty/OK) during the test.<br\>
Example log:
 14/09-19:00 TT === ToneTest START   ===
 14/09-19:04 Line Audio  ERROR  Code: 1F  Dir:   101 Phys:   1 [N001 M65 B01-1]
 14/09-19:06 Line Audio  OK     Code: 45  Dir:   101 Phys:   1 [N001 M65 B01-1]
 14/09-19:07 TT === ToneTest END     ===  Errors: 0(0) Skip: 19 OK: 3(1)

==== Manual ====
A service man must be able to start the test during commissioning and before/after repair to get a full line error status of the exchange.  This is done by dialing a directory number that starts a full test, running full speed with 0,3 seconds per tested station, i.e. 180 stations per minute, or 3 minutes on a full 550 exchange.

The exchange’s log port will print one line per station that have “line audio” error, plus one line for each station that was skipped either due to busy (possibly handset off-hook error) or line wire error status.

Example log:
 14/09-19:03 TT === ToneTest START   ===
 14/09-19:04 Line Audio  ERROR  Code: 1F  Dir:   101 Phys:   1 [N001 M65 B01-1]
 14/09-19:04 TT Skip: Line Down (Faulty)  Dir:   104 Phys:   4 [N001 M65 B01-4]
 14/09-19:04 Handset Off ERROR  Code: 00  Dir:   105 Phys:   5 [N001 M65 B01-5]
 14/09-19:04 TT Skip: Line Down (Faulty)  Dir:   106 Phys:   6 [N001 M65 B01-6]
 14/09-19:04 TT Skip: Board Missing       Dir:   131 Phys:  31 [N001 M255 B64-1]
 14/09-19:04 TT Skip: Board Missing       Dir:   132 Phys:  32 [N001 M255 B64-2]
 14/09-19:04 TT Skip: Board Missing       Dir:   133 Phys:  33 [N001 M255 B64-3]
 14/09-19:04 TT Skip: Board Missing       Dir:   134 Phys:  34 [N001 M255 B64-4]
 14/09-19:04 TT Skip: Board Missing       Dir:   135 Phys:  35 [N001 M255 B64-5]
 14/09-19:04 TT Skip: Board Missing       Dir:   136 Phys:  36 [N001 M255 B64-6]
 14/09-19:04 TT Skip: Line Down (Faulty)  Dir:   152 Phys:  52 [N001 M65 B09-4]
 14/09-19:04 TT Skip: Line Down (Faulty)  Dir:   153 Phys:  53 [N001 M65 B09-5]
 14/09-19:04 TT Skip: Line Down (Faulty)  Dir:   154 Phys:  54 [N001 M65 B09-6]
 14/09-19:05 TT === ToneTest END     ===  Errors: 1(1) Skip: 21 OK: 0(0)
====The log info====
=====Error Code=====
The tone test uses two audio level measurements, one of silence and one of tone.  The tone level must be somewhat higher than silence to pass the test, and if it fails, the difference is shown here as Code (hexadecimal).  If the value is 00, it means identical or negative difference.

Autoload value for limit is 0x28, decimal 40, which corresponds to some 16 dB.   When an error is reported, the Code will be below 0x28.  When the station is reported OK, Code will be above 0x28, typically 30-50.

=====Test statistics=====
When the test finishes, one line with statistics is printed.  <br\>
Statistics includes both total counts, and also changes last test in parentheses (0).
* Errors: The number of stations actually tested with tone, and that failed.
* Skip:	Stations that are configured as testable, but coud not be tested right now either due to wire error 	“Line down (faulty)”, “handset off” timeout, or in use (“resource busy”). <br\> 
	This means that you should consider these as possibly faulty stations, and re-run the test later.
* OK:	The number of stations actually tested and passed in the tone test.<br\>
The sum of (errors + skip + OK) are constant, equal to the number of testable stations.

=== Test Algorithm ===
====Activation:====
[[Image:Bilde18.jpg|right|250px|thumb]]
* Automatic
:* Enabling on system level on/off.  Autoloaded off.
:* Start hour, end hour (runs once per hour in this period).  Autoload 00:00, 00:00.
* Manual, by dialing feature code.  
:* Feature dir.no is autoloaded.<br\>
COS autoloaded only for superuser.
:*Manual activation is allowed even if automatic is disabled.  <br\>
The function can be used as a commissioning test for all systems, where the service man hooks up his laptop PC to the log port just to test the installation before leaving.
* Manual activation while a automatic/manual test is running will just restart the test with new test parameters.

 
==== Algorithm: ====
[[Image:Bilde19.jpg|right|250px|thumb]]
* The test works like this is:<br\>
Connect speaker to silence for 200 ms, then measure level.  Connect speaker to 500 Hz for 200 ms, then measure level.  If the difference is larger than 16 dB, OK, else failure.
* Tone test runs only if station is idle and free.<br\>
The station is allocated while it’s used, with normal (low) priority.  Normal allocation has the following effects:<br\>
- Tone test will not disturb normal operations in the exchange<br\>
- Stations with wire error and non-connected stations can’t be allocated, so they are skipped automatically<br\>
- Audio program is switced off.  Notifier blink/bleep is switched off. <br\>
- Allocation with result “busy” causes periodic re-try, like unsucessful test (gives a camp-on effect)
* Tone tests handles only loudspeaking stations<br\>
- ASLT positions with “station type - PNCI” are skipped<br\>
- AGA ports with “port use - station” are tested<br\>
- all other board types are skipped (ATLB, ASVP, AGA program,. AGA multimodule, AGA AlphaNet)<br\>
- it’s possible to configure “disable tone test” on a per-station basis 
* Tone test may be disturbed by noise
&rarr; Reports unsuccessful test, but does not shutdown the station.  It’s important to keep the functionality of the station which may be left even after a unsuccessful tone test, e.g. the push-button may still be able to signal emergency. 
* Testing of noisy stations<br\>
- A station is accepted as OK after a single silence/tone sequence, i.e. minimum test time.<br\>  
- If a test sequence fails, the station is re-tested up to a configurable number of times (default 3), each with a delay of 5 seconds in between, thus reducing the chance that temporary noise disturbed a test.<br\>
- The station is allocated and released for each attempt.  This strategy also handles possible faulty level (duplex) scanners, which have no self-test.

=== TST NVRAM configurable options for Tone Test ===
The following algoritm parameters can be modified using the TST Console NVRAM command:<br\>
'''.ex_profile .line_monitoring .tt_retries'''<br\>
:Autoload value 3.

:This is the number of re-tries on each station that fails the tone test.  Between each re-try the long 3.6 seconds tone_test_timeout is used (see below).  If a re-try gives an acceptable measurement the station is declared OK, else it’s declared a failure when all re-tries have been done.

'''.ex_profile .line_monitoring .tt_limit'''
:Autoload value 40 (hexadecimal 0x28).

:This is the limit for the difference in audio level between 500 Hz and silence, where 40 corresponds to some 16 dB.

:If there are problems with background noise etc., it’s possible to reduce this limit somewhat.  Inspect the error log, and check the code field to see what values you get in practice (observe that the code field is hexadecimal).   Then set the limit below this.  Reducing the value too much reduces the value of the test, as differences in background noice may be enough to pass the test over a few re-tries! 
 
'''.ex_profile .timeouts .sli_measure_duration'''
:Autoload value 1, i.e. 100 ms (plus 100 ms minimum).

:This is the duration of the silence and the 500 Hz tone periods.
If there is a problem e.g. with echoes, try setting this value to 5 or 10 (0.5 to 1 second).  Note that the actual measurement is just 20 ms, this setting just delays when it’s taken.

'''.ex_profile .timeouts .tone_test_timeout'''

:Autoload value 35, i.e. 3.5 seconds (plus 100 ms minimum).

:This is the delay between each tested station in automatic (night-time) mode, which you can observe as time between each station blink/beep.  The automatic test can be run once per hour, so setting this value higher means that the test may not finish in one hour if there are a few errors.  Setting it much lower means that the exchange becomes more busy when the test runs.

'''.ex_profile .timeouts .skip_test_timeout'''

:Autoload value 2, i.e. 200 ms (plus 100 ms minimum).

:This is the delay between each skipped station.
The test progresses quckly to the next station when no test can be performed (board not in position, no test configured, or station has som other error).  This value can be reduced to increase the speed of the manual test somewhat, on the expence of a more busy exchange.

=== Troubleshooting ===
The test depends on a good ratio between silence and tone.  To improve the measurement quality, you can try different things:
* Increase measurement duration time, so that audio stabilizes e.g. in rooms with echo
* Increase number of re-tries, which increases the likelihood that one test passes 
* Increase speaker volume, so that the tone becomes louder than background noise
* Reduce microphone sensitivity.  The level measurements which are done by the duplex scanners on the ASLT boards are non-linear at high levels and thus may not distinguish noise from noise+tone very well.
* Be aware of stuck call buttons, M-keys etc.  The exchange is allowed to call to (and tone test) such stations, even if they may function not very well audio-wise.

=== Error reporting ===

Error reporting was designed to inform personnel when an error occurred, on standard exchange devices.   Each error is reported only once, and only when the station becomes faulty, NOT when it’s OK again.   Staff should inspect the faulty station immediately, and must keep written records of the total exchange error status.

==== System “Faulty line” event ====
This event is on the system level, i.e. all stations will be processed identically by the actions set up here.  It’s possible to assign one or more actions during commissioning.  Typical:
* MST, i.e. mail to station.   One or more stations can be notified via the mail queue that a station is faulty.
In addition to notifying a local station, an additional mail can be sent to a station in another AlphaNet node.
* MPP, i.e. mail to pocket pager.  One pager can receive display calls when line faults occur.  This will normally be in addition to the MST action.  Pagers often have short displays, and can present reduced error identification only.

Configuration:  AlphaPro RCI screen, below the 60 RCI’s in the scroll box.  
==== Station “Line error” event ====
This event is per station, i.e. unique error actions for selected stations.  It’s possible to assign one or more actions during commissioning.  Typical:
* RCO, i.e. control an digital (relay) output which switches over e.g. a power amplifier to a standby unit.
(this event is reported both for error ON and OFF, but is not restored after reset - se below).

Configuration:  AlphaPro Users/Event screen.  (from AMC version 07.00)



==== Exchange log port (serial printer)====






The log port prints a lines when a fault occurs, identifying the station and the physical position 
(Node, Module, Board-port) it’s connected to. 
14/09-18:44 Line Wire   ERROR  Code: 10  Dir:   106 Phys:   6 [N001 M65 B01-6]
Configuration:  AlphaPro System Info/More screen.
3.4	Data protocol support
3.4.1	AMC 07.20 and earlier versions
There are no direct error reports in the exchange’s AlphaCom data protocol.
It’s possible to use the mail to station information though.  Every time a station receives a mail a broadcast is made.  It’s then possible for an external computer to read out the mail entry using additional data commands, and thus get the mail’s display text which contains information about the kind of error, and the technical address of the faulty station (or device).
This information is available when the error occurs only, and there is no information when the error is corrected (the external computer can poll the faulty station by attempting conversation set-up, which is rejected until the station becomes OK and free).
3.4.2	AMC 07.30 and newer versions
See separate chapter describing data protocol messages and routing possibilities.
3.4.3	TouchLine data protocol 
The TouchLine data protocol never reports errors.
(The upcoming AMC 08.00 will be able to send text to the TouchLine port from the Event Handler.  This can be used to generate TL error messages based on error events.)

== ATLB subscriber lines ==
You can activate line monitoring of (modified) telephones in AlphaPro per physical number.   It’s possible to select the limit for the current, which is necessary to handle various line voltages and line lengths.

===== Telephone with parallel resistor =====
* An on-hook telephone draws very little current.  To be able to detect a wire break, it’s necessary to connect a resistor in parallel close to (or inside) the telephone.
* An off-hook telephone draws a lot of current which is not possible to separate from short-circuit, and is reported if the handset is off for more than a one hour and no dialling takes place.

===== ATLB Error Codes =====
Currently, ATLB reports identically to ASLT.  (As there is little information in that, this may be changed in the future to the 8-bit line current value which is compared to the value entered in AlphaPro as monitoring limit).

:40 (0100 0000) = AB Open

== Supervision of IP stations ==
The IP substation is supervised in two different ways.

# Each IP-sub-stations have its own unique TCP socket for data communication. This socket send TCP-alive frames every 5 second, thus a broken socket is discovered within 10 seconds and reported as line error from AlphaCom.
# From the Application layer AlphaCom is sending a "PING" approximately every 90 second. The stations reply with a "PONG". If one PONG is missing the station is reported as line fault.

IP station down is now reported to the event handler "faulty station line" ON/OFF
and reported on the log port in the two following ways:
 
 930: Feb 13 16:17:58 src@Node1 AlphaLog: N001.D065:.Exch:@D65.B25-1  Line Down: IP ST Data Com St 20  "1001 -"

(Data communication is lost, no PING reply or socket closed) 

 933: Feb 13 16:21:34 src@Node1 AlphaLog: N001.D065:.Exch:@D65.B25-1  Line Down: IP ST License St 20  "1001 -"

(This happens if you enter a new license key with no IP stations licenses or not enough licenses, then some stations will be out of business)

 
The system station will report like ordinary line faults with reference to master board and first port since the IP station is not related to any physical AMC board position: