Here's our Policy:
Four to five questions and answers will be published.  They will be archived on ACES Int'l BROADBand.

Q & A:

Q1.  What are the "Acceptance Test Procedures" for conducting an "Acceptance Test" in fiber optics?   Why can there be conflicting results?

A1.  I have researched many textbooks and standards for Acceptance Test Procedures. The following steps reflect industry-standard practices. Notice that the standards define two "Methods" of testing a fiber optic link or segment and which would provide conflicting results.

Step One: Test and document per industry standards.

The ANSI/TIA/EIA 568B.3 standard addresses installation parameters and testing for both copper and fiber. Specific to fiber optic testing are the ANSI/TIA/EIA 526-7A for single-mode and ANSI/TIA/EIA 526-14A for multi-mode testing. The purpose of an Acceptance Test is to ensure that testing a fiber optic link using an analog signal measurement method will support the digital transmission application of the network.

For the purpose of this article, I will concentrate on multi-mode applications for LAN applications. ANSI/TIA/EIA 455-XX has a variety of publications particular to manufacturers of fiber optic components and testing for cable assemblies. OSP (Outside Plant) has it’s own set of Bellcore standards as does DWDM (Dense Wave Divisional Multiplexing) and WDM applications.

Step Two: Use a good quality optical loss test set (OLTS) or quality power meter/light source to test for light loss.

The best test and simplest instrument to use for fiber optic Acceptance Testing is either a calibrated optical light loss test set (OLTS) or a quality power meter and light source (see question #2). The power meter is analogous to the volt-ohm-milliammeter (VOM) used in electronics. Most are compact and portable with digital readouts, but measures optical power levels either in (1) absolute levels of dBm, (2) in relative units of dB.

Optical power (loss) is the quantity most often measured in fiber optic systems and is the most important property of passive optical LAN components. It determines what part of an optical signal is lost within the component and how much passes through it. (Measure dB loss = Pin in dBm – Pout in dBm/length).

Think of the light source being the saltshaker and the power meter as the peppershaker. You need both to perform an Acceptance Test. Basically, the light source "launches" an absolute level of light (Pin) referenced in dBm (decibels per milliwatt) and the power meter measures the "received" light (Pout) in dBm. The law of physics tells us there will be a loss in light as the distance increases. Less light loss for shorter distances and more light lost for longer distances (like a flashlight).

Loss is measured in a relative unit of dB (decibels) to length derived from the logarithmic ratio of the two quantities of Pin and Pout. Note: A power meter by itself can also be used to check the output power from a optic transmitter.

The OLTS is a unit comprised of both the power meter and light source or the salt and peppershakers all in one. It allows the user to do bi-directional and possible dual wavelength testing without having to swap meters. The new Level III UTP compliance testers mimic an OLTS with their add-on fiber probes.

Step Three: Use and reference the correct test reference cable method to the application and document the procedure and test results.

An Acceptance Test is testing the installed fiber link or segment to a standard so that it will support the IEEE LAN application – the most common is Ethernet.

Turn on the power meter and light source and allow warm-up time per manufacturer instructions.

Ensure that both the light source and power meter are set to the same wavelength (850nm) and the power meter is testing at dBm (not dB).

When testing for 1000BaseSX or Gigabit Ethernet, it is required to use a mandrel-wrapped launch reference patch cable. Take a good quality 1-3 meter patch cord (less than 21dBm or 0.2 – 0.4db) and wrap it around a 5/8ths mandrel five-six times (Molded plastic mandrels are available for under $20.00, but you can use a white board market pen), slowly remove it from the coiled fiber, and secure the wrap loosely with ultra-small tie-wraps or electrical tape. This will help stabilize the high order modes and will also your stabilize your meter fluxuations. I use these modal wrapped reference patch cables for all my testing at the light source only. Some refer to it as a modal conditioning cord. (And, I have found a second benefit that people will quit stealing your test cables for the network)

Test both launch TX "A" and receiving RX "B" patch cords individually with a loss no greater than –21dBm or 0.2dB to 0.4dB each bi-directionally. KEEP ALL THE CONNECTORS CLEAN! Cap them when your through testing and clean them again before use.

Chart 1.1 shows the ANSI/TIA/EIA 526-14A Method A with two reference cables zero’d out that measures only one mated pair connector. Use of LED source.

Chart 1.2 shows ANSI/TIA/EIA 526-14A Method B and 568B.3 using one reference cable zeroed out that measures two mated pair connectors. Use of LED source with mandrel-wrapped launch TX reference cable for testing Gigabit Ethernet.

 

Your first reading in the example of Chart 1.1 below registers -20.5 dBm. The second reading is –23.5dBm. The difference is the loss or 3.0dB and not 3.0 dBm!. Some testers have the feature of "zeroing" out the reference readings and perform the mathematics for you. The terms "zeroed out" or "zeroing out" means the reference TX and/or RX patch cord mated-pair connector(s) measurement is taken out of the tested optical link segment readings.

___________________________CHART 1.1______________________________

Per ANSI/TIA/EIA 526-14A Method "A" (or ISO 61280-4-1 Method 1):

Set reference level using TWO reference cables zeroed out – measures one mated-pair connector

Mandrel wrap:  optional, but strongly recommended for all testing on launch cable only

(Measures –20.5dBm)

LED Light Source <---TX Launch Patch Cable "A"  ---> (<SM Adapter>) <---RX Receiving Patch Cable "B"---> Power Meter

(Measures – 23.5 dBm)

Measurement link loss plus (ONE) mated-pair connection – using SM Adapter

LED Light Source <---TX "A" ----> (<SM Adapter>) <----FO Link Under Test-----> <> <---RX "B"--> Power Meter

(Result is 3.0dB)

______________________________________________________________

______________________________CHART 1.2______________________________

Per ANSI/TIA/EIA 526-14A Method "B" (or ISO 61280-4-1 Method 2) and 568B.3:

Set reference level using ONE reference cable zeroed out– measures two connectors

Mandrel wrap on launch cable optional but required for testing 1000BaseSX

(Measures –20.0dBm)

LED Light Source < ------------TX Launch Patch Cable "A" --------------> Power Meter

(Measures –23.5 dBm)

Measurement link loss plus (TWO) connections – using (TWO) SM Adapters

LED Light Source <--TX "A" --> (<SM Adapter>) <-- FO Link Under Test--> (<SM Adapter>) <---RX "B"--> Power Meter  

(Result is 3.5dB)

______________________________________________________________

Why the two Methods A and B for testing 10/100BaseFX and 1000BaseSX? Lets refer to the beginning of this question. We are testing with an analog or sinusoidal (Sine) wave in order to test for the required digital application. We also must consider the launch of the light and how it propagates through the fiber link – cable and all components. The traditional methods of testing fiber does not apply to the new Gigabit Ethernet applications! Therefore, the required use of a mandrel-wrapped launch cable will help stabilize the high order modes for more accurate readings.

Both the A and B methods are both correct as well as controversial! Method A references or zeros out both launch TX "A" and receiving RX "B" reference cables, but fails to recognize the second mated-pair connection. Method B references or zero’s out one launch TX "A" cable only, thereby recognizing both mated pair connections. However, (and this becomes argumentative) if a bad receiving RX "B" patch cable is used, it would be incorporated into the fiber link segment test measurements data. Only use low loss patch cords!

Now you can understand why there can be conflicting test measurements taken from two different individuals using different test equipment and different test methods – if referenced at all!

CYA! Always document your test method procedure for future reference!

Given the much lower budget loss of Gigabit Ethernet compared to 100BaseFX, research is under study to more accurately measure it. LED light sources using modal wrapping is the most practical way today as it acts like a filtering process by mixing high-order modes to achieve EMD or equilibrium mode distribution stability. However, it does not control cladding modes.

For repeatable and even more accurate measurements, mode distributions in the fiber must be more controlled in the launch process. The intent is to reduce the overfill conditions that exists when using an LED source. In other words, you may be exhibiting higher losses with an LED source.

Alternative techniques being considered include VCSELs and laser sources. It seems logical as the active equipment in Gigabit Ethernet uses these same light sources. These higher-end sources provide a narrower spectrum, which also keeps the chromatic dispersion in check thus lower loss readings. Another technique is actually immersing the fiber’s cladding into liquidized cladding mode stripper material that has a higher refractive. This causes the cladding modes to be passed into the stripper.

Here are some more precautions when using a power meter and light source:

Step Four: When testing multi-mode fiber use the worse case measurement. For multi-mode fiber cable, set your optical loss test set or power meter/light source both at 850nm. Ever heard of the life’s sure sayings of death and taxes? Well, the third is if it passes the specifications @ 850nm it will pass @ 1300nm.

One of the major advantages of fiber is it has substantially less attenuation than copper. The new ANSI/TIA/EIA 568B.3 standard defines the "Dual Window Capability" for 62.5/125um and 50/125um multi-mode fiber @ 850nm and @ 1300nm and single-mode @1310nm and @ 1550nm.

It is also important to have a basic understanding of metrics. Fiber is; (1) international, which is metric-based, and (2) it is easier to express minuet measurements. ALWAYS verify that length is given in feet or meters. 1 meter (m) = 3.281 ft or 39.4", 10 m = 32.81 ft, 100 m = 328.1 ft, 1000 m (1 km) = 3,281 ft

The attenuation dB loss measurements are LINEAR or directly proportional to length. For example, 3.5db/3,281ft = X/1000 ft X = 1.07 dB/Mft

Per ANSI/TIA/EIA 568B.3 DUAL WINDOW CAPABILITY

62.5um and 50.0um/125um multi-mode fiber

Wavelength       62.5um Bandwidth       Attenuation (dB)          Per 1000 ft (Mft)        50.0um Bandwidth

850nm                   160 MHz*km                  3.50 dB                          1.07 dB                    500 MHz*km 

1300nm                 500 MHz*km                  1.50 dB .                        .45 dB                      500 MHz*km

8.3um/125um single-mode fiber:

1310nm                 1.0 dB/km                         .50 dB/km                                            Unlimited

1550nm                 1.0 dB/km                         .50 dB/km                                            Unlimited

                                                                        TABLE 1.1

What can you derive from these specifications? Testing multi-mode @ 850 nm will provide worst-case measurements. Single-mode loose tube OSP fiber experiences less attenuation than tight-buffer designs. When calculating Budget Loss (See Question #5) this becomes very important.

This leads into the question as to why test @1300nm if the operating wavelength has a higher performance or less attenuation/higher bandwidth? Traditionally, it was the operating wavelength for FDDI (Fiber Distributed Data Interface) -- the backbone interconnection between closets. As we get into Gigabit and upwards, it may play a more influential future role. Most of the time, required testing @1300nm has been perpetuated by uninformed specification writers.

You will also notice that the attenuation test specifications get much more tighter with shorter distances from 10/100BaseFX to Gigabit Ethernet. This will be used to determine budget loss. Some networks will work fine for 100BaseFX equipment, but will not support 1000BaseSX! (See worksheet exercise in question #5)

Step Five: Perform bi-directional testing.

The ANSI/TIA/EIA 526-7A and –14A standards require just one-way or uni-directional testing in short cable runs using Method A or B. No matter how short or long the optical link segment is, it is strongly recommended to perform bi-directional testing.

It is reasonable to understand that attenuation is evil… but tolerable. As an installer, you cannot control bandwidth or it’s evils of dispersions (modal and chromatic), but you sure can manage attenuation. Fiber is not affected by EMI and temperature fluxuations (to a degree) like copper and can go longer distances (more advantages!). It is the "events" in the fiber plus the fiber cable length that will effect attenuation; splices, connectors, adapters, bend radius, switches, all contribute to total system attenuation.

A logical (but false) assumption is that dB loss of a connector or splice is the same in either directions. Because of the possible slight differences in the core sizes (49um to 51um which is normal manufacturing tolerances) in matching two fibers together combined with the eccentricity or the centering (polishing) of the cores when terminating a connector, the light’s high-order modes may behave differently from one direction to the other. Going from a smaller (49um) core to a larger (51um) will not exhibit geometrical loss. However, be aware that going in the opposite direction from the larger (51um) to the smaller (49um) would cause up to a .3dB loss.

Security cameras typically use a simplex (one-way on one fiber) signal. Most LANs applications use multi-mode fiber transmitting a half-duplex (one-way both ways on two fibers) signal and long-distance broadband uses single-mode fiber transmitting a full-duplex (both ways $imultaneously on one fiber) signal. For most LAN applications, the installer does not know which direction the user will be transmitting the signal -- therefore the rationale of bi-directional testing. Both readings are recorded and divided by two thus providing an average measurement. This type of documentation will also indicate that the worst case measurement of the two readings in both directions and does not (1) exceed minimum and/or maximum standards and, (2) the average is within the desired range.

This brings up the question of what role does an OTDR play in Acceptance Testing? Remember that the power meter and light source measures optical power loss of the transmitted signal. There are many benefits to using an OTDR, but it should be used only for (1) taking a "finger print" of the complete installation to ensure compliance to the design specifications (complete with event measurements and distances, total length, assessment of the fiber uniformity) and, (2) troubleshooting. In fact an OTDR is NOT needed if any optical link is less than 150-200 m (500-600 ft) and the whole link segments should exceed 500 meters (1,640 ft).

An OTDR is an indirect method of attenuation measurement. It measure backscatter levels (the opposite direction of the light) not the level of the transmitted signal. It needs room – lots of distance-- to work properly. Plus they’re expensive! More on this topic in our next newsletter.

So what defines the Acceptance Test? Remember the goal of any fiber datalink physical layer is to match the capabilities of all components to ensure reliable transmission. The Acceptance Test verifies that the installed optical transmission link segment meets the established performance criteria for successful deployment of the network application – per industry standards.

What industry-standard Acceptance Test Procedures are used? Reference the industry 568B.3 and 526-7A and -14A standards. Use quality optical loss sets and/or power meters and document the procedure using the industry-standard testing Methods A or B. Use good, CLEAN mandrel-wrapped launch and receiving reference patch cables with CLEAN single-mode adapters (couplings). Finally, test multi-mode for worst-case measurements @ 850nm bi-directionally to ensure more accurate and consistent measurements.

A2.  First, attenuation by definition means "signal loss". We measure the loss of amplitude of a sine wave as it propagates down the fiber. Think of a calm swimming pool. Suddenly, someone does a barrel roll into the center and causes this big splash!. At the point of contact (the TX "transmit" side of the signal) the waves (or "sine waves") are big and tall or peaked like a witches hat. As the waves move out ("propagate") towards the sides of the pool (the RX "receive" side of the signal) they become smaller (loss of amplitude) with distance – the waves are "attenuated". They also become flatter – "dispersion"! The objective is to try and keep that wave tall and peaked throughout the distance of the fiber. When the wave becomes too flat and spread out, the receiver may not be able to interpret the signal

This loss of signal is measure in dB or decibels. We measure the TX transmit side first in dBm (decibels per milliwatt) since we are measuring the power of the signal. Power is measured in watts like a light bulb! In fiber applications, the power is very minute (milliwatts, microwatts, nanowatts, etc.) We then measure the RX receive side. The loss of amplitude or the signal loss is determined be comparing the TX dBm to the RX dBm and the net result is measured in dB.

Be aware that attenuation is a direct function of the wavelength and is usually fixed wavelengths of 850nm, 1300nm or 1550 nm. Measured in "db", this "decibel" term means a "ratio of two quantities" of the Power transmitted to the power received. It is also logarithmic (not linear). If you graph a logarithmic measurement along the X and Y axis, it would look like a bottle rocket dying out in flight versus a linear measurement that would be a straight continuous shot.

ATTENUATION (dB) = -10 log Pout/Pin

(Note: The negative sign is added to give attenuation a positive value because of the law of physics! The output power is always lower than the input power for passive components – a flashlight beam gets weaker with length)

Why is all this important? Because you must understand the relevance of decibles. In optical and telephone transmissions equations, 0 dB equals one milliwatt (0dB = 1mW). For every added 3dB of attenuation you loose 50% of your power! (Inversely removing 3dB will double the power). Your objective as an installer is to reduce attenuation through proper installation and termination techniques.

For example, if you induced 3dB into a system or onto a 100-watt light bulb by using a dimmer switch, it would only be as bright as a 50-watt light bulb. Add another 3dB (totaling 6dB) and it is only as bright as 25 watts, 9dB represents 12.5 watts, and so forth. The more dB the dimmer the light. (This also is the basic principle of variable attenuators.)

Now compare these fiber attenuation decibel specifications to copper Category 6. UTP 24awg is 67dB per only 1000 ft – or an extrapolated loss of 99.9999 % of the signal! Now you can understand the reasoning behind the 568B.3 standard for the Basic or Permanent Link of 90 meters or 298 ft and Channel of 100 meter or 328 ft. Since loss is linear and directly proportional to length, the copper loss is now ONLY 21.4dB and 24.1dB respectively - or approximately only 99.1-99.9% of the signal is only lost. HEY! Put fiber in your diet and FTTD (Fiber To The Desk)!

Q3.  What defines a good, quality power meter and light loss set?

A3.  I cannot over emphasize that being a certified ACES Int’l fiber professional means quality tools and test equipment. Be careful in selecting a quality field tester! A field power meter and light source test set ranges in price from $400.00-$4,000.00. A good quality field power meter/light loss set will be in the $1,000.00-$2,000.00 range. This is determined by many factors, but the most critical are: (1) the type of sensing detector photodiode used, (2) operating wavelengths, (3) dynamic and power range, (4) resolution, (5) responsivity and, (6) repeatability. If you are limited in your budget, (7) bells and whistles features should be second priority.

Silicon photodiodes generally have a spectral range of 400nm to 1100nm limiting them to the 850nm for multi-mode testing only. I have found that they are also generally less consistent in their sensitivity measurements and therefore less reliable than other options. Their low price makes them attractive. Germanium is the most practical choice as it has a range from 800nm to 1600nm. Indium Gallium Arsenide or InGaAs is the higher-end photodiode with a range from 800nm to 1800nm and is preferred for single-mode testing. Thus Germanium (and InGaAs) can be used for both multi and single-mode testing at the following wavelengths;

The most common wavelengths at which power meters are calibrated are 665, 820, 850, 1300 for multi-mode, 1300 and 1550nm for single-mode (665nm is used for plastic fibers).

Dynamic range is determined by subtracting the receiver sensitivity level (e.g. –70dBm) from the light source output power range (e.g. –20dBm) resulting in a (50dB) difference. Divide this by the 3.5dB/km @850nm standard for multi-mode glass; 50dB/3.5dB = 14.28 km or 8.86 miles (without any events) to determine the dynamic range. The typical field test set’s fiber optic dynamic range is from –50dBm to -70dBm.

Resolution determines whether the reading will be one or two decimal places or selectable to 0.1 or 0.01dB. Two is preferred.

Responsitivity is the response time directly related to the sensitivity versus wavelength or the switch rate of the source.

Repeatability is related to re-mating cycles and getting consistent same power level readings. Go ahead and check your current tester or one you are thinking of buying on the same cable at least five times. Disconnect completely each time and observe the measurements. How much do they vary? A 0.2dB variance is typical.

Bells and whistles or the added features will add to the price, too. The purpose of these features is to reduce testing and documentation time by offering such features as; bi-directional testing, multi-wavelength capabilities, test data storage with software download, packaging, and accessories. The advantage of the Level III fiber probes is that it produces the same software report as it does for copper. The downside is that you have only one tester for both copper and fiber.

Q4.  The ANSI/TIA/EIA 568-B Cabling Standard is now in effect.  I am a newly Certified ACES Int'l Professional Fiber Optics and Professional Data Cabling Installer.  Who publishes the Standards, where can I get a copy of it and do I have to pay anything for it? 

A.4  To obtain a copy of the complete TIA/EIA Building Wiring Standards contact Global Engineering at 1-800-854-7179 or www.global.ihs.com.   Sorry, it’s not free. The cost  is $595.

A5.  Standards committees (dubbed TR4X.X, and comprised of volunteers) are typically made up of representatives from cable and connector manufacturers, other standards organizations (e.g. IEEE), some large end users, and telecom distributors. To be more competitive, it is the best way for any these organizations to stay abreast of the new technologies and future standards.

In 1989, I was the Director of International Sales and Marketing for Thomas and Betts’ telecom division – formerly Nevada Western. Nevada Western was one of the originators of the modular patch panel concept. And, at the time, we were one of the largest players in this market. One of my job responsibilities was also new product development. What better ways to keep pace than to join the standards organizations.

As more companies entered the market, it became necessary to develop standards in order to promote the technology. Give credit to Anixter who actually began the "Levels" program which evolved into the "Category" program. Northern Telecom – now Nordx CDT – also sponsored the committees’ development.

"Neaner Neener… I make a better patch panel than yoooooou! My company makes better and prettier cable than yours!" These were the cries of the market in the mid-late 80’s. So, let’s put the "proof in the pudding". The TR41.8.1 committee was formed to develop standards for commercial building cabling (and connectors) – now the 568B standard -- and I was one of the original-founding members. I stayed on the committee for the next five years as well as the IEEE 802.3 Token Ring and X39.5 FDDI standards committees.

In each of these, there’s allot of work to do. Subcommittees do a variety of tasks from research and development, reporting on other standards organizations findings (have to stay especially close to IEEE), setting up test labs for the new technologies, etc. Additionally, once a proposed standard is agreed upon, it has to be set in motion with third party testing organizations like UL (Underwriters Laboratory) and ETL (Edison Testing Labs) – P.S. Did you know the Thomas Edison establish this lab? Give him credit. He opened up his discoveries to the market with standards committees, too, in order to promote his new "technologies"!

I will say it was very rewarding and enjoyable, too. I gleaned so much from being in the heart of telecommunication’s technology, meeting people from around the world, and developing business contacts – which some developed into good friendships. To participate as an active voting committee participant, your company has to be a supporting member. The frequent traveling and related travel expenses – plus being away from your job responsibilities -- can make it quite expensive. So, it is not something practical for an individual or small VDV installer..

A6.  My advice would be to be to first get in contact with the manufacturer and the local sales representative in your area. They might have or can acquire the operation and maintenance manual – even for older outdated models. They can also determine of it needs calibration. Splicers are in the high-end of the fiber market. Any manufacturer who is no longer in business was probably bought or absorbed into another company. A few phone calls or web searching should determine the new owner.

In the interim, "cleanliness is next to godliness". Use 99% isopropyl alcohol, fiber-rated Q-tips, Kim wipes, and compressed air can to clean thoroughly. Lube any moving parts (knobs and hinges) with DROPS from a needle with WD40. Or most hardware or techie stores have needle lubes for electronics. Make sure you do not lube any of the fusion splice or alignment areas as it may damage the electronics.

Precision is defined by the accuracy and the durability of the splice. The newer models are almost all automatic once the glass has been cleaved. They do the "X", "Y", and "Z" axis alignment, seal it, and test it. Older models, which you might have, are more labor intensive. The glass alignment is the most challenging step. A good splice will have a 0.1 – 0.05dB and below loss and you can yank on it without it coming apart. Does your meet those criteria? 

A7.  This column is too short to instruct someone on the proper design for a fiber-cabling infrastructure. However, these guidelines should help you in putting together a Loss Budget.

Think of a Loss Budget as a checkbook. The balance or deposit is provided by the equipment manufacturer or the IT network administrator in the form of maximum attenuation. Or, use the Application Specifications as a guide.

The ANSI/TIA/EIA 568B.3 standards use the following guidelines:

Table 7.1

Lets use an example of a backbone situation design where there the router is patched in Building A’s ER (Equipment Room) and the switch is in Building B’s TR1 (Telecommunications Room or Closet). There’s 200 ft of 62.5um/125um OFNR (Optical Fiber Non-conductive Riser-rate) fiber cable from the A-ER to building A’s EF (Entrance Facility). Here, it is patched or cross-connected onto 1000 ft of loose tube 62.5um/125um OSP (outside plant) cables. This OSP run has a splice in the pull box. The cable enters building B-EF where it is spliced onto OFNP 62.5um/125um (Optical Fiber Non-conductive Plenum-rated) cable. This cable is routed 100 ft into B-TR1 and patched to the switch. Total specification designed optical link segment is 426 m (1,300 ft).

FIBER PLANT DESIGN

BLDG A 

Router                       A-ER                                                  A-TC1                                 A-TC1

Ü ----3’ Patch----Þ Û Ü --------200’ OFNR-----------Þ Û Ü ------3’ Patch-----Þ Û Ü ---------400’ OSP-------

 1                            2          3                                               4         5                                6          7

BLDG B

                  Splice                                                                   B-EF                                       B-TC1                        Switch

--------------- Ä ----------------------600’ OSP------------------- Ä -------100’ OFNP--------Þ Û Ü ----3’ Patch----Þ

           8                                                       9                            10      11                   12

Notice that the Budget Loss Worksheet in Chart 7.2 below has both the "Attenuation per 568B.3" and "Recommended Target". Follow the standard first. "Recommended Target" represents (1) a quality installation by professional ACES certified installers given today’s technology with terminating connectors and splices, or (2) specified manufacturers specifications.

For ease of identification when putting together a graphical displayed budget, number the events. Patch cords are too short to be relevant to any loss.

BUDGET LOSS WORKSHEET CHART  7.2

Link/Segment                               Attenuation per 568B.3                             Recommended Target

1-2/3 patch cord (2 connectors)       .75 X 2 = 1.50dB                                        .5 X 2 = 1.00dB

3-4 200’ OFNR                               1.1 X .2 = .22dB                                        1.1 X .2 = .22dB

4/5-6/7 patch cord (2 connectors)      .75 X 2 = 1.50dB                                       .5 X 2 = 1.00dB

7-9 1000’ OSP                                 1.1 X 1 = 1.10dB                                      1.1 X 1= 1.10dB

8 Splice                                                          .30dB                                                    .10dB

9 Splice                                                           .30dB                                                   .10dB

9-10 100’ OFNR                                 1.1 X .1 = .11dB                                      1.1 X 1 = .11dB

Budget Loss Total:                                            6.53dB                                                  3.63dB

FIBER PLANT DESIGN

BLDG A

Router                       A-ER                                                  A-TC1                                   A-TC1

Ü ----3’ Patch----Þ Û Ü --------200’ OFNR-----------Þ Û Ü ------3’ Patch-----Þ Û Ü ---------400’ OSP-------

1=.75dB                2 =.75dB  3              .22dB                 4 = .75dB  5                           6 = .75dB  7               .44dB

.50dB                       .50dB                    .22dB                        .50dB                                     .50dB                   .44dB

BLDG B

                  +Splice                                                                B-EF                                       B-TC1                         Switch

--------------- Ä ----------------------600’ OSP------------------- Ä -------100’ OFNP--------Þ Û Ü ----3’ Patch----Þ

            8 = .3dB                  .66dB                 9 = .3dB      1.1dB      10 =.75dB= 11        12 = .75dB

                  .1dB                  .66dB                       .1dB      1.1dB             .50dB                       .50dB

Total 568B.3 Loss = 6.53          Total Recommended Loss = 3.63

CHART 7.3    

Compare the Budget Loss design to the Application Specifications Table 6.2

APPLICATION SPECIFICATIONS

                                                              Max. Supportable Distances   MaxChannelAttenuation(dB)

Application      Wavelength      /Bandwidth        62.5um      50um       62.5um        50um       Source

10BaseFL            850nm         160 MHz*km       2000 m     2000 m      12.5              7.8            LED

100BaseFX         1300nm         500 MHz*km       2000 m     2000 m      11.0             6.3            LED

1000BaseSX        850nm                                                                         3.2              3.9

1000BaseSX                               160 MHz*km      220 m                                                            Laser

1000BaseSX                               200 MHz*km     275 m                                                             Laser

1000BaseSX                               400 MHz*km      500 m                                                             Laser

1000BaseSX                                500 MHz*km      550 m                                                             Laser

1000BaseLX         1300nm                                                                        4.0                3.5

1000BaseLX                                400 MHz*km       550 m         550 m                                           Laser

1000BaseLX                                500 MHz*km        550 m                                                             Laser

Application               Glass                  Max. Dist.      Optical Link        Budget       568B Target      Recommended

10BaseFL         62.5/160Mhz* km         2000 m             426 m            12.5dB          6.53db                    3.63 

100BaseFX       62.5/160Mhz* km         2000 m             426 m             11.0dB          6.53db                   3.63

1000BaseSX     62.5/160Mhz* km           220 m              426 m              3.20dB          6.53db                   3.63

1000BaseSX     50/400Mhz* km              500 m              426 m              3.9dB           6.53db                   3.63

1000BaseSX     50/500Mhz* km               550 m              426 m             3.9dB            6.53db                   3.63

1000BaseLX     62.5/400Mhz* km             550 m              426 m            4.0dB             6.53db                   3.63

TABLE 7.2

The Budget Loss design of 568B.3 loss of 6.53dB and Recommended Target loss of 3.63dB will both support 10BaseFL and 100BaseFX applications on 62.5um/125um 160Mhz* km optical glass -- as it is within the budgets of 12.5dB and 11.0dB and 2000 m (6,562 ft) maximum backbone supportable distance.

However, it would not support 1000BaseSX with 160Mhz* km glass that has a very low budget of 3.2dB and maximum distance of 220m (722 ft). The options would be to first specify a manufacturer connectors and splices with the Recommended Target low loss and either;

Specify 50.0um/125um 400 or 500Mhz* km optical glass to increase the distance to 500-550 m and allow a budget of 3.9dB

Use 1300nm equipment with 62.5um/125um 400 Mhz* km optical glass to increase the distance to 550 m with a 4.0dB budget. Note:

Use a fiber repeater to boost the signal.

Which option would you choose? What are the consequences?

1) If most of the active equipment and the majority of the installed fiber plant is 62.5/125um, it doesn’t make it practical since you will exhibit up to 1.0-2.0dB loss when going from 62.5 to 50/125um fiber. There’s goes your budget!

2) 1300nm equipment adds costs. It is twice to three-times more expensive than 850nm. However, the optical fiber cable loss calculations for the design would be reduced by a third using 1300nm wavelength reducing the total loss another .8dB!

3) Fiber repeaters add costs, too. You would have to determine how many you’d need and evaluate all the options.

To verify compliance to the installation, apply the Acceptance Test Procedures as outlined in Question #1. Your documentation or Method you used should be provided to the customer -- whether hard copy or soft copy. This will be your Acceptance Test – ensuring that your fiber installation is tested and documented properly, and that the attenuation measurements satisfy the budget loss requirements for the application and the devices on the network.

In the next newsletter, I’ll explain how to determine power watt to dBm conversion and vice versa.

How To Submit A Question:

e-mail: conradb@crossbowcom.com

Non-published questions will be answered off line within 10 days.

Bo

D.A. Bo Conrad, RCDD

President/Director

CrossBow Communications

ACES Int'l Home Page

Copyright© 05/15/2001 ACES L.L.C.

All rights reserved