Published on February 18, 2014
FIBER NETWORKS evolution, applications, standards, measurement methods of
3 LAWS OF NETWORKING #1 - Networks Never Go Slower Plan for higher speeds, increased throughput, reduced response time #2 - Networks Never Get Smaller Plan for more users, more traffic, more capacity #3 - Networks Never Stay the Same Plan for flexibility, reconfiguration, manageability
Applications Driving Bandwidth Interactive Video High Audio Desktop Video Quality of Service Voice Data Video Pipe Medical/Financial CAD / CAM Client/Server Apps Low Email File Transfer 10 100 1000 Information Capacity (Mbps) New real-time, information-rich applications require exponentially higher bandwidth
High Speed Networking As data rates increase, loss budgets get smaller. It is increasingly critical that the fiber plant is installed using best practices.
FACTS – SAFETY NOTE Under NO circumstances should you ever look into a fiber The wavelength being used cannot be seen by the naked eye Light from an active fiber if powerful enough, can do permanent damage to your eye sight The equipment you are using today cannot do permanent damage to your eye sight if exposed occasionally If you are told a fiber is not active, treat it as if it is NEVER stand in front of a fiber patch panel unless ALL of the fiber connectors have a protective cap on them Wash your hands BEFORE and AFTER going to the bathroom
FIBER TYPES Multimode (MM) Singlemode (SM) Core Cladding Buffer Example: 62.5/125 (62.5 m core) Example: 8.3/125 (8.3 m core)
FIBER DESIGNATION • The following now applies to ANSI/TIA-568-C and ISO/IEC 11801:2010 • • • • • OM1: 62.5 µm multimode fiber with a MBW of 200 MHz/km OM2: 50 µm multimode fiber with a MBW of 500 MHz/km OM3: 50 µm multimode fiber with a MBW of 2000 MHz/km OS1: 9 µm singlemode fiber New designations • • OM4: 50 µm multimode fiber with a MBW of 4700 MHz/km OS2: 9 µm Low Water Peak singlemode fiber
FIBER DESIGNATION OM1 OM2 Orange jacket Orange jacket OM3 Aqua jacket OS1 Yellow jacket
Fiber Splices Mechanical Quick Little specialized equipment required New techniques and splice connectors have improved the loss per splice (some < 0.1 dB) Good for emergency field repairs, low volumes Fusion Requires special, expensive equipment Hard to do under adverse conditions Low loss (can be < 0.05 dB) The only method for long links
FIBER ORIENTATION / MATING Typical fiber (core) orientation is either north, south, west, east or a combination of those Fibers for testing have almost centered orientation Ability to maintain best results and repeatability Expensive and hard to get
OPTICAL SOURCES Light Emitting Diodes (LED) Used for multimode: 850 nm or 1300 nm Wide beam width fills multimode fibers Wider spectrum (typically 50 nm) Inexpensive Lasers Used for singlemode: 1310 nm or 1550 nm Two Common Types: Fabry-Perot (FP) and Distributed Feedback (DFB) Narrow spectrum (can be less than 1 nm) Narrow beam width (does not fill multimode fibers) Highest power and fastest switching Most expensive (especially DFB) Wavelength Wavelength
COMMUNiCATION TX/RX Two fibers One fiber used for Rx, one for Tx One wavelength Used in structured infrastructure One fiber Two or more wavelenghts (using WDM) Telco/provider infrastructure
FIBER OPTIC EFFECTS ON A GOOD MEASUREMENT
Fiber Optic Measurements Optical Power an absolute measurement of power measured in dBm as a reference to one milliwatt of power Attenuation (Loss) the amount of light that is lost in a fiber path. Is measured in dB as a relative reading of power. Dispersion spreading over time of a ray of light as it travels through a fiber
Optical Power Optical Power is measured in dBm (0 dBm = 1 milliwatt) Some examples 0 dBm= 1. Milliwatt = 1000 microwatts -10 dBm = 0.1 milliwatts = 100 microwatts -20 dBm = 0.01 milliwatts = 10 microwatts -30 dBm = 0.001milliwatts = 1 microwatt Every 3 dB subtracted drops the power in half
Measuring Optical Loss Loss Power Power (in dB) Lost (%) Received (%) 3 50 50 10 90 10 20 99 1 30 99.9 0.1 40 99.99 0.01 50 99.999 0.001 Loss (dB) = 10* Log Measured in dB: Not a linear scale, but a logarithmic scale For every 3 dB down, received power drops by a factor of 2 For every 10 dB down, received power drops by a factor of 10 Power (received) Power (transmitted)
Measuring Optical Loss 1. Loss Is Measured As A Difference In Power Patch Cable Source 1. Meter Example: Measures - 20 dBm Then measure power after coming out of the fiber link Fiber Link Source Patch Cable Patch Cable Meter Adapter Adapter Example: Measures - 23 dBm 1. The loss is the difference in dB (3 dB in this example)
Sources of Loss in Fiber Optic Paths Dirty connections The fiber material Impurities Variation in material density Coupling losses between fibers Bends in the fiber
Losses From Dirty Connections Dirt moves from dirty to clean - it is a “VIRUS” Good Connector Fingerprint on Connector Dirty Connector
Sources of Loss within the Fiber
Coupling Losses Area mismatch Spacing loss Axis misalignment Angular misalignment
Bending Fiber Optic Cable Fiber is sensitive to bending If you bend the fiber too much, light escapes out of the fiber Singlemode wavelengths are more sensitive to bending loss than multimode wavelengths Datacenters, FTTx applications need special care when pulling fiber
Dispersion (limits link length) The main source of signal distortion in fiber optic transmissions Dispersion lengthens the transmitted light pulse as it travels down the fiber Dispersion is a critical offender for support of high data speeds When the fiber cable length is too long, pulses “run together” and the receiver can no longer decode the signal information
Applications limited by dispersion Modal Bandwidth (MBW) is the unit to classify dispersion in the cable 50 /125 µm has less dispersion – it’s core is smaller than 62.5/125 µm fiber Your loss may be lower than the allowed fixed loss, but if it exceeds the length found here (IEEE 802.3) there may be errors on the network IEEE 802.3ae contains no values for OM4, atm certification limit will be 300 m
How many are there? Application standards These include IEEE 802.3z 1000BASE-SX The optical loss allowed is a fixed value Cabling standards These include ANSI/TIA-568-C.0 and ISO/IEC 11801 (ISO/IEC 14763-3) The loss allowed depends on the number of adapters, splices and length of the cable
Application standards Fixed test limits are defined by ‘system’ specs Examples: 100BASE-FX, 1000BASE-SX, 1000BASE- LX, 10GBASE-S, ATM, Fibre Channel
Application standards Loss Limits Continue to Tighten 14 12 12.5 13 12.5 8 11 10 11 6 6 4 3.56 2 2.6 10BASEFOIL Token Ring 16 Mb FDDI/TP PMD 10BASEFL ATM Fibre Channel 100BASEFX 1000BAS E-SX 10GBASES 0 Token Ring 4 Mb dB 10 13 1986 1987 1993 1993 1994 1995 1998 2002 1989 1992
Cabling standards Test limits for installed fiber link are independent of any network application Limit is calculated, based on cable length, number of adapters, and number of splices Examples: TIA/EIA-568-C, ISO11801:2010, EN50173
ANSI/TIA-568-C.0 and ISO/IEC 11801:2002 Test Limit You are permitted: 0,75 dB per adapter (connector pair) *0,3 dB per adapter for MM and 0,5 dB for SM (ISO/IEC 11801:2010) 0,3 dB per splice You are permitted for multimode fiber 3.5 dB per km @ 850 nm 1,5 dB per km @ 1300 nm You are permitted for singlemode fiber (TIA) 1,0 dB per km @ 1310 nm and 1550 nm Inside Plant (ISP) 0,5 dB per km @ 1310 nm and 1550 nm Outside Plant (OSP) You are permitted for singlemode fiber (ISO/IEC) 1,0 dB per km @ 1310 nm and 1550 nm
ANSI/TIA-568-C.0 and ISO/IEC 11801:2002 Test Limit Let’s calculate the allowable loss for the link below: 850 nm: Adapters Splices Fiber = 2 * 0,75 dB = 0 * 0,3 dB = 0,1 km * 3,5 dB Allowable loss = 1,50 dB 0,00 dB 0,35 dB 1,85 dB
HOW TO MEASURE?
FIELD FACT FACT FROM FIELD: I DO NOT NEED TO TEST – WE CONNECT THE FIBER AND IF THE LED LIGHTS UP GREEN WE ARE GOOD TO GO!!! THE USER APPLICATION WILL BE JUST FINE...
CLEANING & INSPECTION No. 1 and the most crucial step before any testing on any kind of a network We need perfect Test Cords, Launch Fibers Dirt could move to core Will damage other connectors
Cleaning connectors Optimal Cleaning Method (dry or wet) 1. Use cleaning pads without cellulose 2. Apply a minimal amount of solvent (point 1) for very dirty connectors – DO NOT USE ISOPROPYL (IPA) ALCOHOL 3. Place end-face perpendicular to card in first corner 4. Swipe through “N” shape using gentle pressure and moving from wet to dry - NEVER use “8” method to clean 5. Check end-face with microscope or camera, place cap on test reference cord
Cleaning ports For dirty port use wet method (1st foam tip is wet to clean the dirty port, second foam tip is dry to pickup dirt from port)
Inspection Two methods Microscope Cheap – no built-in filter Higher cost – built-in protection filter Inspection camera Safest method for inspection of ports and connectors
Connectivity Verification The basic test before any testing is to use a VFL to determine: Where there is a break in the fiber panel Where there is an excessive bend in the fiber panel (typically an issue in FTTx networks and home installations) Verify correct polarity
Connectivity Verification Limitations of a VFL: The fiber can be broken in multiple places, yet you will still see light coming out at the end It is of no use in trunk cable, the light will never pass through the outer black jacket VFLs rely on the fiber jacket being translucent
5 Standard Ways To Test Fiber In 2010, the old OFSTP-14 was replaced by a new ISO standard. The TIA has adopted IEC 61280-4-1 as the replacement of OFSTP-14. Most of the two documents is the same, with some important exceptions. For insertion loss, three reference methods are are still approved, but the nomenclature is different - no more "Method A, B or C" designations- it's now 1, 2 or 3 reference cables. OTDR testing is now an approved test method as long as you use both launch and receive cables. Reference test cables with "reference grade connectors" are recommended. Methods are given for testing and verifying the loss of reference test cables. For multimode modal control, CPR with a mandrel wrap is gone, replaced by "Encircled Flux," a complex - and not completely debugged - method of measuring the source output.
TESTING (structured cabling) Network Designers may include two levels of testing in fiber test specifications: Tier 1: OLTS (Optical Loss Test Set) Testing the installed cable plant for link loss and verifying the cabling length and polarity Polarity for some backbone simplex applications may not need to be verified. Tier 2: Tier 1 plus an OTDR trace Testing for anomalies and assuring uniformity of cable attenuation and connector insertion loss. The higher level of testing providing quantitative measures of the installed condition and performance of the cabling system and its components. Evidence that cable is installed without degrading events (e.g., bends, bad connections, bad splices)
TESTING (structured cabling) Using the Right Method is Essential ISO/IEC TR 14763-3 specifies “Method 2 – one jumper reference” for testing the fiber link. A mismatch of the connector type between the OLTS and the link requires a modified method. Using the wrong method can result in a significant measurement error. ALL LIGHT SOURCES TAKE TIME TO STABILIZE
Multimode MANDRELS Required, not optional If I don’t use them? You may end up failing perfectly good links What do they do? They strip out the higher order modes in multimode TRCs Result is more repeatable, consistent optical power measurements when using LED sources. Do not use non bend insensitive fiber TRCs The mandrels will have no affect
Encircled Flux The New Standard in Multimode Testing Accuracy
OLTS – One Jumper Reference Tier 1 Setting Reference Added dB dB includes link + ALL connectors Supports a Permanent Link Measurement
OLTS - Modified One Jumper Reference Tier 1 The connector is a Setting different type then the Reference one on the OLTS Added dB includes link + all connectors dB
OLTS – Two Jumper Reference Tier 1 Unfortunately is the most intuitive method but NOT correct Setting Reference dB includes Link + ONLY ONE End Connector dB
OLTS – Three Jumper Reference Tier 1 The connector can be a Setting Reference different type then the one on the OLTS dB Includes Link/Channel BUT NO End Connectors dB Supports a Channel Measurement
OLTS – Method comparison Tier 1 The standards clearly define the referencing method, but the nomenclature is confusing Link End Connections Included in Loss Results Latest Versions Previous Versions TIA/EIA-526-14A (multimode) TIA/EIA-526-7 (singlemode) IEC 61280-4-1 (multimode) IEC 61280-4-2 (singlemode) 1 Connection 2 Jumper Method A Method A Method A.2 Method 1 Method A2 2 Connections 1 Jumper Method B Method B Method A.1 Method 2 Method A1 None 3 Jumper Method C Method C Method A.3 Method 3 Method A3
“Supplemental” OTDR Trace Tier 2 OTDR trace is a supplement to OLTS Determined cause of excess loss Troubleshoot fiber link Telco / Service providers still rely on OTDR testing This is good for veryfing the installation, still OLTS is required to meet the application requirements
Understand OTDR plots: Reflective events Almost always two mated fiber connectors (air gap) Could be a bad mechanical splice too On the OTDR trace, they are characterized as a spike Higher reflectance means better mating
Reflectance in connectors When the light travels down the fiber, if it sees a change in refractive index, their will be a reflection (reflectance) The most common causes are: Ideal world – no undercut Air gap between the connectors Dirt or residue left behind by the cleaning solution Good reflectance for UPC connector is ≈ -40 to -50 dB (recommended < -45 dB) Reality -best factory terminated connectors will have an undercut better than 50 nm (that’s 0.05 µm).
Understand OTDR plots: Non-Reflective events Almost always a splice, could also be a very good APC connection – can be an issue when using limit lines Good reflectance for APC is < -60 dB If you only see it at 1300nm (MM) or 1550 nm (SM), then it is a bend in the fiber On the OTDR trace, they are characterized as a dip in the trace
OTDR Bi-directional Averaging Do I need to do this? If the customer asks for it, YES ANSI/TIA-526-14-B; if the launch fiber has a different backscatter coefficient to the fiber you are testing, then you should ISO/IEC 14763-3; This is not necessary where the cabling under test comprises a single length of fixed cabling with terminating connectors and where the optical fibres of the launch and tails cords have the same scattering characteristics. Will always give the most accurate loss at each event Must use Launch + Receive Fiber Compensation in Setup
OTDR Launch fiber Each connector used in the Launch Fiber network MUST have the same connector type for mating Using short patch cords make results even worse Will give loss of the first connector
OTDR Receive fiber Receive Fiber Will give loss of the last connector
Common Measurement Issues Setting Reference When done do NOT remove fiber from the source, otherwise repeat reference method Negative Loss reading Sources not stabilized Bad reference Gainer (mismatch of MM cables used in link) Ghosting Long Attenuation dead zone Dirty or damaged connectors/mating Poor reflectance Bad/damaged launch fiber Dirty connetors Dirt on the fiber endface Good reflectance, but high loss only on one wavelength Excessive fiber bending
Passive Networks vs. Active Passive Optical Networks (PON) implement a: Point-to-Multipoint Topology Triple-play Fiber-to-the-Desktop (FTD) Has two (2) “active components” (OLT and ONT)
Proposed tests at each step of the way Proposed test each step of the way Network and Equipment Installation stage Verify total loss budget Link Characterization Coupler and Splices Characterization End to End Loss & Back Reflection Testing OLT and ONT turn up Maintenance stage How many ONTs are out? One? Some? All?
Network and Equipment Installation stage Verify total loss budget ≅16 dB for 1:32 ≅10 dB for 1:8 ≅7 db for 1:4 ≅ 2 – 3 dB ≅ 0,35 dB/km@1330 nm
Link characterization Done before splicing or connectorizing splitter From OLT towards splitter and from splitter to the ONT OTDR at wavelengths 1310, (1490) and 1550 nm
Coupler and Splices Characterization Done after splice and connectorizing process From splitter output back to OLT OTDR with 1310, (1490) and 1550 nm with launch fiber Backreflection or coupler ports should be -35 dB or better (ITU-T G.983.1) End to End Loss & Back Reflection Testing Done with a power meter and source at 1310, 1490 and 1550 nm Max Loss 25 dB for Class-B PON (20 dB Class-A and 30 dB Class-C) OTDR trace OLT->ONT and ONT->OLT to confirm back reflections to be -35 dB or more and splice losses less than 0,1 dB
OLT and ONT turn up Before any ONT turn-up Verify power at ONT coming from OLT at drop point Masure at 1310, 1490 and 1550 nm to get a go-no-go power reading (pass/fail criteria is user defined and depends on OLT)
Maintenance Stage ONT Outage when SOME are out Check power level at the closest faulty ONT No power – fiber defect between OLT and ONT Strong power (in the limits of OLT spec) – exchange the ONT Weak power – use fiber identifier to measure total power on fiber or visual fault locator for possible macrobending
ONT Outage when ALL are out Easy to overcome and fix Either OLT is out of order or a major fault has occurred between the OLT and the spliter point Check feeding (F1) fiber – OTDR trace for OLT side at 1550 nm OTDR facts when testing in PON environment 1625 nm must be used in live PON and fro ONT -> OLT If testing from OLT side a PON optimized OTDR must be used in order to see “through” splitters
OTDR OLT->ONT Example
WHAT IF? I rent fiber from a provider? Do I need to sign SLA? I want to test how fast my network is? WHAT DO I DO????
Ethernet testing To verify a link performance you do not need to understand standards Key paramaters for Ethernet testing Determining the layer of testing and the nature of the link under test Connectivity Physical connectivity Layer 2 and up connectivity Throughput Burstability Latency/Jitter
Connectivity Layer 1 Verify Link Up and MDI settings Verify Speed, duplex and flow control Layer 2 and 3 Layer 2, verify MAC connectivity ARPing the address Layer 3 connectivity, verify IP address is valid before doing any additional test PING
And what about T-PUT and others BER Testing Familiar concept Throughput Testing Better but not complete All in one RFC2544 Frame loss Burstability Jitter Latency RFC2544 Various Frame Sizes All tests in one set of measurements Provide A to B and B to A results to identify the direction of the network having problems (if any)
40G, 100G 40GBASE-SR4 and 100GBASE-SR10 MPO trunks (12 core fiber), utilizing several 10G fiber links Fiber standards for certification?
THANK YOU! Green Light / Red Light IS NOT ENOUGH YOU NEED TO TEST IN ORDER TO GUARANTEE THE PERFORMANCE OF THE USER APPLICATIONS
Optička merenja na spojevima ... optičkih pojava na optičkim vlaknima i spojevima ... ABSTRACT The method for measuring attenuation on fiber ...
Merenja CD i PMD (hromatske disperzije i disperzije polarizacionog moda) na optičkim vlaknima. ... na osnovu znanja, ...
Pri merenju hromatske disperzije na G652 optičkim vlaknima ... od talasne dužine na osnovu merenja u ... Troubleshooting Optical Fiber ...
Merenja na optičkim komunikacionim ... omogućuje optičkim vlaknima da provode svetlosni snop. ... (multimode fiber MM), ...
vlakna sa drugim optickim vlaknima (mehanickim ... Merenja na optičkim komunikacionim sistemima-1. ... OTDR_Merenja_na_optickim_kablovima.
brzinama po optičkim kablovima koji su instalirani po zemljovodnom užetu visokonaponskih vodova, i kod ... nastavljanja merenja na ostalim vlaknima.
... na optičkim vlaknima KBV Datacom d.o.o. spada u mali broj firmi u našoj zemlji koje su u mogućnosti da izvedu merenja ... Merenja CD i PMD ...
Scribd is the world's largest social reading and publishing site.