Monthly Archives: September 2014

Be ware of Unnecessary ActiveStandby SCC Switching on OSN 3500&7500

Abstract
Because 3.3 voltage detecting circuits of SSN1GSCC02 and SSN3GSCC02 boards are of low precision, unnecessary active/standby SCC switching occurs, resulting in a possibility that NEs are unreachable to the NMS. If the ASON feature is enabled on NEs, data configuration may be lost during the active/standby SCC switching.

GSCC Board

[Problem Description]
If the following conditions are met, there is a possibility that the problem is triggered.

  • The NE runs active and standby SSN1GSCC02 boards that were manufactured before 2010-04-06 or active and standby SSN3GSCC02 boards that were manufactured before 2010-12-28.
  •  The FPGA of the SSN1GSCC02 boards or SSN3GSCC02 boards is of V120 or an earlier version.

Fault symptom:
Unnecessary active/standby SCC switching occurs on the NE that runs active and standby SCC boards; the standby SCC board is reset and reports the COMMUN_FAIL alarm after the active/standby SCC switching.

Identification method:
When the following four conditions are met, it can be determined that the problem is triggered:

  •  Active/Standby SCC switching occurs many times on the NE that runs active and standby SCC boards.
  •  The NE runs SSN1GSCC02 boards manufactured before 2010-04-06 or SSN3GSCC02 boards manufactured before 2010-12-28.

To query the type of the SCC board and when it was manufactured, run the :cfg-get-bdinfo: SCC board ID command in the Navigator (for Huawei OptiX OSN 3500 NEs, the SCC board IDs are 17 and 18; for OptiX OSN 7500 NEs, the SCC board IDs are 24 and 25). Then you can find information about the SCC board type and manufacturing date by checking the “BoardType” and “Manufactured” items in the returned information.
For example, query the type and manufacturing date of the SCC board on the Huawei transmission equipment OptiX OSN 3500 as follows:
:cfg-get-bdinfo:17
VERSION-INFO
/$[ArchivesInfo Version]
/$ArchivesInfoVersion=2.0
[Board Properties]
BoardType=SSN1GSCC02
BarCode=02664510A1000253
/$Item=03026645
/eDescription=System Control and Communication Board
/$PCBVersion=C
Manufactured=2010-01-06
ManufactureCode=1
/$CLEICode=

  • The FPGA of the GSCC board is of V120 or an earlier version.

To query the version of the FPGA, run the :ver command in the Navigator and you can find the FPGA version by checking “Logic” item in the returned information.
BIOS 08.09.03T02 20091113 10:57:13 inactive
ExtBios 09.09.04T01 20091118 11:48:14 active
NeSoft(D) 5.21.16.13 20091215 18:35:29
Platform(T) 5.00.07.C03T13 20090720 11:27:06
AosSoft(T) 5.99.02.10T26 20070228 10:49:11
Logic (U1004)120
Dsp

  •  To query records in the black box file bb1.log, run the :log-query: SCC board ID,”bb1.log” command in the Navigator.

For example, to query the black box file bb1.log of the SCC board in slot 17, run the following command:
:log-query:17,”bb1.log”
The returned information contains one of the following records:
…….
### Hard Register is not registered ###
…….
[HF:5][HF:8][HF:9]
…….

[Root Cause]
The 3.3 V voltage detecting circuits of the SSN1GSCC02 board manufactured before 2010-04-06 and SSN3GSCC02 board manufactured before 2010-12-28 are of low precision. Even if the 3.3 V power supply fluctuates within the allowed range, the board detects that there is an exception, triggering active/standby SCC switching. The SCC board is reset and reports the COMMUN_FAIL alarm.

[Impact and Risks]
For an NE that runs active and standby SCC boards, if unnecessary active/standby SCC switching occurs when the active SCC board is backing up data in batches, there is a possibility that data is lost; in addition, the NE is unreachable to the NMS transiently. For an NE that runs only one SCC board, there is no impact.

[Measures and Solutions]
Recovery measures:
【TN-R-201307】Rectification of Unnecessary ActiveStandby SCC Switching due
to Low Precision of 3.3 V Voltage Detecting Circuits on OSN 3500&7500 INTERNAL
If unnecessary active/standby SCC switching occurs when the NE is backing up data in batches, data is lost and services are interrupted. For a non-ASON NE, clear the NE databases, reset the SCC board, and download data from the NMS to the NE; for an ASON NE, contact R&D engineers for help.

Workarounds:
None
Solutions:
1. For NEs that can be upgraded:
First upgrade the FPGA. Upgrade the SSN1GSCC02 board manufactured before 2010-04-06 and SSN3GSCC02 board manufactured before 2010-12-28 to V100R008C02SPC500+SPH502 or a later V100R008C02 version, or V100R010C03SPC203+SPH207 or a later V100R010C03 version.
If the version of the SCC board is V100R008C02SPC200 and cannot be upgraded to V100R008C02SPC500+SPH502 or a later V100R008C02 version, or V100R010C03SPC203+SPH207 or a later V100R010C03 version, upgrade the FPGA of the SCC board to V130.

2. For NEs that cannot be upgrade:
Replace the SCC boards. Replace the SSN1GSCC02 board with an SSN1GSCC02 board manufactured after 2010-04-06 and the SSN3GSCC02 board with an SSN3GSCC02 manufactured after 2010-12-28.
For an ASON NE, replace the faulty SCC board with the N4GSCC board based on actual situations.
Solution of material handling after replacement:
Return the boards to Huawei headquarters.

[Rectification Scope and Time Requirements]
For the rectification scope, refer to the SSN1GSCC02 List and SSN3GSCC02 List.
It is required that the rectification is complete before 2014-04.

[Rectification Instructions]
Refer to the Replacement Guide for the GSCC Boards on the NG-SDH Equipment.

TwitterLinkedInGoogle+FacebookPinterestTumblrStumbleUponRedditShare

Be care for SS49SCB Boards DIP Switch on Certain Sets in the Installation State Problem

Abstract: After certain sets of huawei transmission equipment OptiX Metro 1000 V3 (SS49SCB VER.B) restart upon power failure, the SW1,SW2 DIP switch of the NE is in the installation state. As a result, the NE configuration is missing, and the services are interrupted.

[Problem Description]

Trigger conditions:
The OptiX Metro 1000 V3 (SS49SCB board) restarts upon power failure or is manually reset.
Fault symptom:
After the device restarts upon power failure, all the NE configurations are missing, and the services are interrupted.

Identification method:
1. The NE software version of the huawei metro transmission OptiX Metro 1000 V3 is 5.37.05.12. Hardware version of SS49SCB is VER.B:
:sys-get-nesoftver;
VERSION-INFORMATION
NESOFTNAME    VER                     DATE           STATE
NSF1                   5.37.05.12        20080831       active
NSF2                   5.37.05.12        20080831     inactive
BIOS                    8.12.48               20100612      inactive
ExtBIOS              9.12.45               20100612       active
Pcb                      SS49SCB VER.B
Logic                   (U301)110
Total records :6

2. Run the following command:
:dm:0×70000000
If the result is not 0 (shown as follows, marked by red circle), it means that the OptiX Metro 1000 V3 NE (SS49SCB VER.B) encounters this problem after reset.

Run the following command

[Root Cause]
After the equipment test is complete, the DIP switch is not restored to the normal state. The DIP switches of certain SS49SCB boards are still in the installation state after the manufacture equipment test. The cause is described as follows: Normally, the actual unit test is complete within three minutes. The DIP switch fault can be detected. If the unit test is performed after the unit is powered on for 15 minutes, however, the NE with abnormal DIP switch status works properly after the equipment test. In this case, the unit test cannot detect the DIP switch fault.

[Impact and Risks]
All the NE configurations are missing after the OptiX Metro 1000 V3 device (SS49SCB board) restarts upon power failure or a cold reset is performed on the SS49SCB. The services are interrupted.
Products involved: OptiX Metro 1000 V3 (SS49SCB VER.B)

[Measures and Solutions]
Recovery measures:
NA
Workarounds:
NA
Solutions:
Power off the SS49SCB board, and remove the board. Then, set the SW1、SW2 DIP switch to the normal status. Insert the board. After the board runs properly, download the NE configuration through the NMS.
Note: The following figure shows the position of the SW1、SW2 DIP switch on the SS49SCB board. To set the DIP switch to the normal status means to switch bits to the ON side of DIP switch.

SS49SCB Boards)

Solution of material handling after replacement:
NA
[Rectification Scope and Time Requirements]
NA

Notice of Flash Memory Failure in WDM Products

Summary:
Because of defective incoming materials, the Spansion Flash memory units provided in a specific lot cannot erase Flash memory data. Huawei  transmission boards equipped with these Flash memory units occasionally encounter a software uploading failure, patch upgrade failure, or log recording failure; or the boards sometimes reset unexpectedly or fail to boot from a reset. So please have a caution on the fail flash memory in WDM(OSN 6800& OSN 8800)products.

Huawei osn 8800

[Problem Description]
Trigger condition:
The problem corresponding to this rectification notice is likely to occur when a board listed in the attachment is undergoing an upgrade/reset or backing up logs because of a Flash memory erasing failure. The probability increases with the frequency of erasing the Flash memory.
Symptom:
Upgrade failure: When an involved NE is undergoing a software or patch upgrade, the U2000 displays that loading the software or installing the patch has failed.
Backup failure: An SCC board backs up the database every 30 minutes. When an involved SCC board attempts to back up the database after a configuration change occurs in the database, the attempt fails and the board reports a DBMS_ERR alarm. The new configuration is lost after the NE is reset.
Unexpected reset or failure to boot from a reset: An involved board is reset upon a component failure. During the restart from the reset, the board performs a Flash read/write check and fails to boot.
Identification method:

The boards in the attachment are at the risk of having this problem.
[Root Cause]
The production process of Spansion is not well controlled. Due to production errors, the grating oxide layer on the Flash chip wafer is uneven. Horizontally, the oxide layer has become thinner because of corrosion. In a high voltage application (programming or Flash erasing), the oxide layer will be punctured and consequently erasing the Flash memory will fail. For a board equipped with a Spansion Flash memory unit, the Flash memory is erased when the board is undergoing an upgrade/reset or is backing up logs. The probability that the board fails to boot increases as the Flash memory is erased more often.

[Impact and Risk]
The boards listed in the attachment will occasionally encounter a software uploading failure, patch upgrade failure, or log recording failure; or the boards sometimes reset unexpectedly or fail to boot from a reset.

[Measures and Solutions]

Recovery measures:

Set aside all the boards and repurchase all you need ones in professional Huawei optical transmission product distributor.

 

Troubleshoot Link STM-64 using BPA and DCU in OSN 7500II

Background
How to troubleshoot link STM-64 in Huawei transmission equipment OSN 7500II using BPA and DCU because although have followed the theory guidance about power input range below , but there always find alarms B2_SD, B1_SD_B2_EXC, MS_RDI, R_LOF appears in its link (by NMS U2000)

Link information and BPA specification
1. Link : STM-64

2. Board : SLD64

3. Distance : 96 km (with one site as OTB passthr), as like picture by OTDR below:
picture by OTDR
4. Patrol is OSN 7500II    Jatiluhur is OSN 7500II
5. Topology
Cause Analysis:
This is caused of non liniear factor.
BA output is so high (+14db), the distance is so far (about 96 Kilometres), so when the fiber quality is not good, it have a big nonlinear affection on service
A. Pre-troubleshoot:
1. Please prepare:

More XFP that support > 80 km – colored XFP(1550.12nm,193.4THHZ),
More Attenuator (5db, 7 db, 10 db, 15 db) ,
More Patch cord (LC/FC, LC/LC) ,
Core cleaner (alcohol, dust free, etc) ,
2 pcs OPM (in each of site is mandatory to bring) ,
Laser / OTDR >> to get actual fiber loss and distance (mandatory to bring, enough in near site or far site only) ,
Local connection to monitor using U2000
2. Please let familiar about BPA link connection guide
3. BA and PA must be used as one pair. That mean cannot use BA or PA seperately
4. Install fiber connection each of site only for TX direction first without attenuator
5. using OTDR to check OSP/OTB firstly (note: no patch cord in opposite site connect to its OTB during OTDR)
6. Fiber check by OTDR
7. Integrate one direction from TX in each of site
B. Integrate one direction (TX from Jatiluhur to RX Patrol )
Each of site stay focus to one direction only :
First measuring the power TX from Jatiluhur to RX Patrol :
install fiber connection in TX Jatiluhur without attenuator first in Jatiluhur with step by step measure by OPM until Patrol
Write down the power in each of point, each of patch cord, the purpose for avoiding if there is bad fiber also or not
(in advance, we found out that 2 pcs fiber in Patrol was bad/cannot transmit signal, after changed it can transmit)
Put attenuator and measure power to match the range in RX Patrol
C. What we meet when integrated BPA link 
D. Step by step troubleshooting 
1. The first one we considered, the board and XFP is ok or not
So we Try to loopback itself in Patrol (SLD64-out to SLD64-in)
2. BA output is 14~15dbm, when the fiber quality is not good, it have a big nonlinear affection on service,
Then we try put 15 db at BA-out to decrease the BA output power, but the received power is -33dbm
3. Then we try put 12 db at BA-out to increase the input power of PA, then alarm dispeared
E. The reverse direction
measuring the power TX from Patrol to RX Jatiluhur :
Follow the same steps with opposite direction (TX from Jatiluhur to RX Patrol ) ,
install fiber connection in TX patrol without attenuator first in Patrol with step by step measure by OPM until Jatiluhur ,
Write down the power in each of point, each of patch cord, the purpose for avoiding if there is bad fiber also or not
(we found out that 2 pcs fiber in Patrol was bad/cannot transmit signal, after changed it can transmit) ,put attenuator with same range attenuator with first measuring.
attenuator
F. Result
then in that link used BPA and DCU in source (sample: Jatiluhur) and sink (sample: patrol) site already clear from alarm, and also can be created in NMS U2000
Suggestions and Summary:
From case study above, have conclusion that although now the link don’t have the alarm. But we notice that the received power(-31dbm) is lower than specification(-28dbm). That is mean in the later , the fiber degradation, the link don’t have any margin to work normal and must go to site to decrease the attenuator on.
It is suggested that If the received power of PA is higher than suggested value(-26), please add attenuator on BA out port firstly to decrease the nonlinear affection.

Pay attention to the Software Version Compatibility Issue in OptiX OSN 1800

Summary:
For Huawei transmission equipment OptiX OSN 1800 V100R003C01SPC200 (5.67.03.23) or an earlier version, services on the TNF2LQM or TNF2LDGF2 board that uses the service chip of version V110 will be unavailable after the board is initialized because of a bug. The problem can be resolved by upgrading the OptiX OSN 1800 to V100R003C01SPC300 (5.67.03.25) or a later version.

Huawei OptiX OSN 1800

[Problem Description]
Trigger conditions:
The problem occurs when the following conditions are present:

  • The TNF2LQM or TNF2LDGF2 board uses the service chip of version V110.
  •  The NE software version is V100R003C01SPC200 or earlier.

Symptom:
Services are unavailable after the board is powered on and configured with services.
Identification method:

Run the following command to query the service chip version on the board:
:woptp:$bid,1,83,1,ff,23,00,04,10,00,2 # “bid” indicates the slot ID of the TNF2LQM or TNF2LDGF2 board in the hexadecimal format.
The following information may be returned:
Optp cmd : ff23
00 04 10 00 02 87 51 00 11 # “00 11″ indicates that the service chip version is 110.
If “00 11″ is displayed in the returned information, the board is involved in this warning.

[Root Cause]
For Huawei OptiX OSN 1800 V100R003C01SPC200 and earlier versions, the TNF2LQM and TNF2LDGF2 boards that use the service chip of version V110 have a bug in board initialization.

[Impact and Risk]
Affected scope: in and out of China
Risk: The services are unavailable.

[Measures and Solutions]
Recovery measures:
Upgrade the OptiX OSN 1800 to V100R003C01SPC300 or later and perform a cold reset on the involved board.
Workarounds:
None
Preventive measures:
See “Recovery measures” as described above.

Huawei Ascend Mate7 starts sale in China

Huawei formally launched its latest large-screen flagship smartphone Ascend Mate7, and Ascend P7 of sapphire version and Ascend G7 to the domestic market, the prices of Mate7 standard version and high-end version are 2999 yuan and 3699 yuan respectively.

Huawei Ascend Mate7

Huawei’s consumer BG CEO Richard Yu once declared in his blog, “Huawei has put extremely high R&D capital into Mate7, which is a top flagship smartphone with superior performance, user experience and sense of quality, it is also the best large screen smartphone by now, the value of which is definitely more than 830USD.”

The standard version of Huawei Mate7 is of two colors: silver and black, which has a 2G RAM and a 16G ROM, it has three standards of China Mobile, China Unicom and China Telecom, the pre-sale will start on September 22nd and formally on sale in September 28th.

The high-end version of Mate7 is of amber golden and adopts a combination of 3G RAM and 32G ROM, it is also equipped with the exclusive Huawei gold membership privilege services card, the value-added services include exclusive service hotline, a boot gift bag of Huawei mobile phone application market, free of charge to use a 160G online storage in 2 years, this phone was formally on sale in Huawei mall on September 12th.

Mate 7 is first launched in Berlin on September 4th, and drew much attention in the following IFA, it is regarded that Mate 7 will challenge Samsung in large-screen smartphone market. Richard Yu also revealed that, the launch of Mate 7 in between of Galaxy Note4 and iPhone 6 is a “contest” with Apple and Samsung, by providing the consumers with a comparison and choice.

Cations for the Cable of the OptiX OSN 8800 Fan Assembly Is Incorrectly Connected

Summary: Under certain circumstances, a batch of fan assemblies of the Huawei Optical transmission OptiX OSN 8800 T32 or T64 intelligent optical transport platform have a speed regulating problem in high temperature.

Huawei OSN 8800

[Problem Description]
When the version of the fan board is changed according to the requirement, the layout and cabling engineer modifies the slot of the fan board in the fan assembly from fan 1, fan 2, fan 3 to fan 2, fan 3, fan 1 (left-to-right). When regulating the speed of the Huawei OptiX OSN 8800 partition, if the high temperature alarm occurs in the third partition, the full speed command will be dispatched to the J7 terminal. However, the J7 terminal is connected to the fan of the second partition, which results in the board of the third partition is always in high temperature. Modify the speed of the fan in the second partition to a higher rate based on the policy of the adjacent speed modification. That is, the full speed command is dispatched to the J8 terminal. The J8 terminal is connected to the fan of the first partition. Consequently, the board of the third partition is always in the high temperature.

Figure 1 Cable connection

Cable connection

 

Symptom
The NE reports a TEMP_OVER alarm. In the fan assembly, all fans are in full speed except the fan in the third partition.

Figure 2 Partition of a NE

Partition of a NE

Diagnosis Method
Step 1 Check whether a TEMP_OVER alarm is reported. If it is reported, go to the next step. If it is not reported, the fan assembly is not involved in this precaution notice and no further action is required.
Step 2 Query the manufacturer information of the fan assembly, and check whether the fan assembly is included in the delivery list (find the attachment). If it is included, the fan assembly is involved in this precaution notice. Otherwise, it is not involved.

[Root Cause]
In 2009, the version of the fan assembly is changed according to the requirement. The new version has no inheritance to the earlier version because the slot ID of the fan is changed. The manufacturer assembles the device according to the old figure for the assembly. As a result, the speed regulating of the fan assembly fails to be performed.

[Impact and Risk]
The NE reports a TEMP_OVER alarm, which has no impact on the device operation.

[Measures and Solutions]
Temporarily restoration
Manually run the full speed command to the fan assembly for maximum heat dissipation.
Solution
Change the cable order of the fan assembly according to the R&D operation guide. After that, the fan assembly can be used as a spare part.
Install the R4C04SPH870 hot patch that will be published in January 2014).

 

Be Ware of Services Interruption on OptiX OSN 3500

Summary:
On a port of an SSN1PEG16 or SSN1PEX1 board of a version earlier than V200R011C03SPC100, both LAG and CAR are enabled. After the board is upgraded to a version between V200R011C03SPC100 and V200R013C10SPC100 (itself excluded), the board is repeatedly reset due to a software bug. As a result, services are interrupted.

Huawei OptiX OSN 3500

[Problem Description]
Trigger conditions:
The problem occurs when the following conditions are met: On a port of  Huawei Service board SSN1PEG16 or SSN1PEX1 of a version earlier than V200R011C03SPC100, both LAG and CAR are enabled. The board is upgraded to a version between V200R011C03SPC100 and V200R013C10SPC100 (itself excluded).
Symptom:
An SSN1PEG16 or SSN1PEX1 board is repeatedly reset and services on the board are interrupted.
Identification method:
Check whether the current NE version is V100R009C03 or a V200R011 version earlier than V200R011C03SPC100.
Check whether the target version of the upgrade is between V200R011C03SPC100 and V200R013C10SPC100 (itself excluded).
Check the board configuration. Check whether a port on the faulty board is configured with LAG and CAR.

[Root Cause]
This issue is caused by a software bug. When CAR is enabled for a slave port in a LAG on an SSN1PEG16 or SSN1PEX1 board of V200R011C03SPC100, the scenario that the board version was smoothly upgraded from another historical version is not considered. When the SSN1PEG16 or SSN1PEX1 board gets online, the database does not apply for the CARID resource for the slave port. Therefore, after the slave port gets online, it fails to obtain a CARID resource and the default CARID of 0 is used. As a result, verification of board parameters fails and the board is reset.

[Impact and Risk]
Services are interrupted.

[Measures and Solutions]
Recovery measures:
If the NE has two SCC boards, switch over the active and standby SCC boards.
If the NE has only one SCC board, warm reset the SCC board.

Workarounds:
During an upgrade, before activating an SSN1PEG16 or SSN1PEX1 board, switch over the active and standby SCC boards. For details, see the Guide to Prevent Interruption to Services on SSN1PEG16&SSN1PEX1 Boards of Huawei OptiX OSN 3500 After an Upgrade. Here for more huawei transmission equipment
Solution:
Upgrade the faulty NE to V200R013C10 or a later version.

What will happen if you falsely match Software Package with OSN 8800 Platform Subrack

Summary: When an Huawei OptiX OSN 8800 platform subrack is falsely connected to the master subrack of V100R006C01SPC100, SPC200, SPC300, or SPC500 version as a slave subrack, the master subrack will automatically match the software package for the slave subrack. As a result, the TN15AUX board in the OptiX OSN 8800 platform subrack resets repeatedly and cannot be managed.

huawei osn8800

Problem Description
Triggering conditions:
The OptiX OSN 8800 platform subrack is connected to a master subrack of the V100R006C01SPC100, SPC200, SPC300, or SPC500 version as the slave subrack.
Symptom:
If the Optical transmission equipment OSN 8800 subrack is falsely connected to a master subrack of the V100R006C01SPC100, V100R006C01SPC200, V100R006C01SPC300, or V100R006C01SPC500 version as the slave subrack, the master subrack will automatically match the software package for the slave subrack. That is, the TN15AUX board software of the R6 version is matched for the TN15AUX board in the OptiX OSN 8800 platform subrack of the R7 version. The TN15AUX board resets repeatedly and the subrack cannot be managed because the TN15AUX board software of R6 version has defects.
Under this circumstance, this problem persists even if the faulty OptiX OSN 8800 platform subrack is connected to a normally operating master subrack of the R7 version again. Therefore, replacing the faulty TN15AUX board is the only method to solve the problem.
Identification method:
1. The version of the master subrack is V100R006C01SPC100, V100R006C01SPC200, V100R006C01SPC300, or V100R006C01SPC500.
2. The slave subrack is the OptiX OSN 8800 platform subrack.
Root Cause
The TN16AUX board software in the software package of the V100R006C01SPC100, V100R006C01SPC200, V100R006C01SPC300, or V100R006C01SPC500 version applies to both the TN15AUX and TN16AUX boards.
If the TN16AUX board software of the V100R006C01SPC100, V100R006C01SPC200, V100R006C01SPC300, or V100R006C01SPC500 version is installed for a TN15AUX board, the board will reset repeatedly.

Impact and Risk
The TN15AUX board resets repeatedly, and the OptiX OSN 8800 platform subrack cannot be managed.

Measures and Solutions
Recovery measures
Give up the  faulty TN15AUX board and go to a  Huawei transmission equipment vendor to buy a new one.
Preventive measure:
Prevent the OptiX OSN 8800 subrack from connecting to the master subrack of V100R006C01SPC100, V100R006C01SPC200, V100R006C01SPC300, or V100R006C01SPC500 version as a slave subrack.
Solutions:
No solution is available.

Have You Ever Meet Of The OCTL Port Application in the OptiX OSN 1800

Summary
An OCTL port on a TNF1SCC board can only be connected to an OCTL port on a CTL board of an OADM frame through a straight-through network cable. Due to the fact that the OCTL port appears similar to a common network port, the OCTL port is likely to be misconnected and short-circuited, causing a blown fuse on the TNF1SCC board and a power failure on the OADM frame.

[Problem Description]
Trigger conditions:
The normal application scenario is: An OCTL port on a TNF1SCC board is connected to an OCTL port on a CTL board of an OADM frame through a straight-through network cable.
Any other application scenario is possible to trigger this problem. For example:
A. An OCTL port on a TNF1SCC board is connected to an OCTL port on a CTL board through a non-straight-through network cable.
B. An OCTL port on a TNF1SCC board is connected to a non-OCTL port (for example, an SW port).
Symptom:
1. When a TNF1SCC board is connected to an OADM frame, the OADM frame is not powered on and its RUN indicator is off.
2. The OADM frame cannot be managed.
The following is a schematic connection diagram of  Huawei OSN product like OptiX OSN 1800 OADM frame and an OptiX OSN 1800 I chassis.

Huawei Optix OSN1800 OADM

Identification method:
1. The TNF1SCC board is running normally but the OADM frame is not powered on, its RUN indicator is off, and it cannot be managed.

2. When the OADM frame is connected to an OCTL port of another TNF1SCC board, the OADM frame is powered on normally with an RUN indicator turned on and can be managed. It can be determined that the OCTL port on the previous TNF1SCC board is faulty with a blown fuse.

[Root Cause]
An OCTL port on a TNF1SCC board, which supplies power to the OADM frame, appears similar to a common network port. If the OCTL port is connected to a network port, the OCTL port will be short-circuited, leading to a blown fuse in the protection circuit of the OCTL port on the TNF1SCC board. This fault, however, does not affect the operation of the TNF1SCC board.
The following is a schematic power supply diagram of the OCTL port on the TNF1SCC board.

TNF1SCC board

Functions of the fuse:
The fuse on the TNF1SCC board is used for protecting power outputs to the OADM frame and implements hot swap protection together with the transient voltage suppressor (TVS), ferrite bead and capacitor included in the protection circuit shown in the blue box. The operating principle of the fuse is as follows:
A. When the connected OADM frame is working normally, the fuse current is about 0.2 A, far lower than its rated current 2 A. The fuse is working normally and serves as a power line.
B. When the connected OADM frame is faulty, for example, is short-circuited, the fuse has a transient current far higher than the rated current 2 A. As a result, the fuse blows and the protection circuit is opened to isolate the power supplies and components inside the TNF1SCC board. The opened protection circuit can be restored only by replacing the TNF1SCC board.

Table 1 Pin arrangement of the OCTL port

table1

Table 2 Pin arrangement of a standard network cable (straight-through cable)

table2

[Impact and Risk]

In site deployment or expansion, the OADM frame connected to a TNF1SCC board cannot be powered on or managed.
[Measures and Solutions]
Recovery measures:

Replace the TNF1SCC board.

Workarounds:
None.
Solutions:
Replace the TNF1SCC board.
Material handling after replacement:
The TNF1SCC board is returned for repair.For more transmission board you can visit Huawei transmission network.