Measurements in fiber optic systems - transmission method.

This article summarizes the knowledge for the installer who faces the task of verifying the correctness of a fiber optic system. The article describes in detail all aspects related to the idea and procedures of measurement by the transmission method, i.e. using an optical power meter (OPM) and a light source (LS) or an optical loss test set (OLTS).
Contents:
Measurements in fiber optic systems: general description of methods.
Among the measurement methods defined by PN-ISO/IEC 14763-3 and PN-EN 61280-4-2 standards, we can distinguish those related to the so-called basic and extended tests. Basic tests ("Tier 1") make it possible to determine the compliance of a completed channel or fixed link with the requirements of a given application (e.g., one of the Ethernet applications) and possible certification of the network. Extended ("Tier 2") tests are usually used in addition to the basic test or when the basic test gives a negative result and it is necessary to verify the exact cause of this situation.
The aforementioned standards defining parameters, methods and measurement procedures are directly related to the ISO/IEC 11801 and EN 50173 structured cabling standards. Installers implementing such networks and performing certification should strictly adhere to the rules defined in the standards and use equipment that allows such tests and the issuance of the applicable certificate. In practice, the permissible equipment to be used is defined in the warranty conditions of the structured cabling manufacturer.
A series of ULTIMODE devices for testing and measuring fiber optic installations.
A series of ULTIMODE devices for testing and measuring fiber optic installations.
However, there is an area of systems that are not implemented with reference to structured cabling standards and for which no costly certification is expected. However, in such situations, investors often expect a document confirming that the system has been made correctly and will work when active equipment is connected to it. The installer also often wants to verify that the parameters of the made system, such as its attenuation, will allow it to be used for specific applications. How to proceed then? Which measuring devices to use? What standards to refer to? How to properly carry out the measurement procedure?
The division of methods into the basic tests and extended tests mentioned in the first paragraph is certainly what every installer should use in their work.
A basic test is a test that primarily measures the attenuation of the channel. Such a measurement – known as the transmission measurement (or transmission method) – uses a stable light source and an optical power meter. In a nutshell, these devices, connected to the two ends of a fiber-optic link, allow to measure the attenuation contributed by it. Attenuation (insertion loss ) is absolutely the most important parameter to consider when verifying a fiber optic system. This is where the capabilities of cheap and popular light sources and optical power meters generally end. Very expensive sets designed for network certification additionally allow measurement of such parameters as link propagation delay, link length and link continuity. What is extremely important, they also allow the evaluation of the compliance of the results with selected assumptions or standards and the preparation of a measurement report. The lack of such a possibility in the case of low-cost meters and light sources, causes installers to turn their attention to devices that allow extended tests – most often the OTDRs.
Transmission method measurement - basic fiber optic link test.
Measurement using a light source and optical power meter in accordance with PN-EN 61280-4-2 or ISO/IEC 14763-3:2014 is the basic way to verify the correctness of a fiber optic link. It can also form the basis for network certification for specific applications.
The idea behind the transmission method of measurement is simple. To the completed fiber-optic connection, usually terminated on both sides in switches, boxes, a light source of known power is connected on one side, and an optical power meter on the other. Test patchcords are used when connecting the devices.
The idea of measuring by transmission method.
Knowing the power of the light source that injects the signal into the optical fiber and reading the power on the optical power meter, one is able to determine how much of the source power has been precipitated, or in other words, what is the attenuation of the connection made. Most available light sources generate power at -5 dBm. If a power meter connected on the other side reads -8 dBm, for example, this will mean that the attenuation of the measured line is 3 dB.
However, performing a measurement as above, without the so-called zeroing procedure of the measurement system, is subject to a very high uncertainty and cannot be treated as a reliable measurement. The uncertainty of the measurement is due to several issues. The most important include:
  • uncertainty related to source power: the manufacturer's declared power level value of -5 dBm may actually be different. Ignoring the issues of warming up the device before the measurement (it should take 15 - 20 minutes), these devices may generate power slightly different from the declared one;
  • uncertainty related to the attenuation of the light source connector: when connecting the measurement patchcord to the light source, we generate additional signal attenuation of unknown value. The light source connector is the one that generates loss. This is due to the design and construction of the device itself;
  • uncertainty related to the attenuation introduced by measurement patchcords: when measuring using measurement patchcords, their attenuation is taken into account in the final result. Since these patchcords are not part of the measured path and the value of the attenuation contributed by them is unknown (in the extreme case it could be a significant part of the total), they should not be taken into account in the measurement.
To reduce measurement uncertainty, measurement standards PN-EN 61280-4-2 and ISO/IEC 14763-3:2014 prescribe a procedure called system zeroing, also known as measurement system calibration or reference measurement (performed with reference to another value). There are 3 methods of system zeroing: the 1 patchcord method, the 2 patchcord method, and the 3 patchcord method. They all involve the same thing - connecting the light source and the power meter with a measuring patchcord or patchcords, and then saving the obtained power as a reference value for the next measurement, which will already be the actual measurement on the made line. The name "system zeroing" refers to the fact that, as a rule, after connecting the devices with the measurement patchcord/patchcords, the user presses the "REF" or similar button on the meter, which ends up storing the currently read power in the device's memory and displaying a value of 0 dB on the meter's screen. From now on, anything plugged in additionally between the devices (in particular, the line you want to measure) will generate additional attenuation, which will be directly displayed on the meter's screen. The idea of zeroing the circuit with each of the three methods is presented below.
Transmission method measurement: circuit zeroing - 1-patchcord method.
Transmission method measurement: circuit zeroing - 2-patchcord method.
Transmission method measurement: circuit zeroing - 3-patchcord method.
After zeroing the circuit, unplug the devices, and then connect them to the switches to measure the attenuation contributed by the made line. When doing so, do not unplug the patchcord from the light source, since connecting and disconnecting the plug at this point generates slightly different attenuation values each time.
Consider the example from the beginning of the note, in which the attenuation of the measured line without resetting the circuit was 3 dB. Suppose the same line is measured now, but preceding the measurement by zeroing the circuit using the 2 patchcord method. Connecting a source with a claimed power of -5 dBm to the meter using 2 patchcords and an adapter, a power indication of -6 dBm is obtained on the meter. It follows that the meter patchcords contribute 1 dB of attenuation. In fact it is not quite clear how much attenuation the patchcords themselves contribute, because we still cannot be sure about the declared power of the source (if the source would generate a signal of -5.2 dBm, the attenuation of the patchcords is 0.8 dB). Nevertheless it is not important at this point. What is important is the measurement we make in the second step - in reference to the power value stored in the meter (in this case -6 dBm). The circuit is zeroed by pressing the REF button. After resetting the circuit, you connect the equipment to the measured line and get a value of -2 dB on the meter screen. The measured value of line attenuation stripped of the measurement uncertainties described above.
Each of the three methods of the measuring system zeroing, due to the use of a different number of patchcords in determining the reference power, will ultimately generate a slightly different measurement result. So which one should be chosen? Intuition here usually suggests the method of 2 patchcords, since that is the number used in the final measurement. However, it turns out that this method is less accurate than the 1-patchcord method, and it is 1 patchcord that should be used when zeroing the circuit whenever possible.
More about methods of establishing references.
As two patchcords are required to connect the light source and the optical power meter to the line to be measured, the most intuitive method of establishing the reference power (calibration of the measurement system) is the one using 2 measurement patchcords (otherwise known as reference patchcords, test patchcords or TRC (Test Reference Cords). However, it turns out that the most accurate method is the single patchcord calibration method. It is the one recommended as the most appropriate according to the ISO/IEC 14763-3 and PN-EN 61280-4-2 measurement standards, factory standards used by large operators, as well as the instructions of structured cabling system manufacturers.
The figure below shows the measurement range for each of the three methods of setting the reference power. It appears that 1 patchcord method actually allows the measurement of the entire line to be measured: from beginning to end including the start and end connectors. The 2 patchcord method reduces the measurement range by the attenuation of one connector (this is because the attenuation of 1 connector is taken into account in the reference process, while the result of a measurement using the 3 patchcord method reference, leaves out the attenuation of 2 connectors. Note that the following figure, although generally accepted by the standards, provides some simplification, since the attenuation of the connector (or connectors) during the establishment of the reference (i.e. the connection of the two reference plugs) is not the same as the attenuation of the connector (or connectors) in the measured line (i.e. the connection of the reference plug to the standards).
Transmission method measurement when 1 patchcord is used in the reference establishment process (calibration of the measurement system). The green markers indicate the measured attenuation range – from the start connector to the end connector including these connectors. The calibration method using 1 patchcord is therefore best method to establish the reference power.
Transmission method measurement when 2 patchcords are used in the reference establishment process (calibration of the measurement system). The green markers indicate the measured attenuation range – from the start connector to the end connector without one of the connectors. This involves taking into account the attenuation of 2 connectors in the calibration process of the measurement system. It is therefore less accurate than the 1 patchcord method.
///Transmission method measurement when 2 patchcords are used in the reference establishment process (calibration of the measurement system). The green markers indicate the range of measured attenuation – from the start connector to the end connector – without both connectors. This involves taking into account the attenuation of 2 connectors in the calibration process of the measurement system.
At this point, it is worth mentioning the reference patchcords themselves. According to the recommendations of the above-mentioned standards, the "top quality" patchcords, for which the attenuation of the connector does not exceed 0.2 dB (a value of 0.15 dB can also be found), should be used. This is because during the measurement, the start and end connectors of the measured line contain the pins of the above-mentioned reference patchcords (reference connector – standard connector connection). Therefore, the patchcord connectors should introduce as little measurement uncertainty as possible. In fact, during the operation of the line, these connectors will be replaced by standard patchcord connectors – e.g. when connecting active equipment or crossing the distribution frames. Consequently, the lower and more predictable/repeatable the attenuation of the reference connectors, the more accurate the measurement.
Manufacturers of measuring devices for network certification offer such "special" patchcords at multiples of the price of generally available patchcords. These patchcords, in addition to good transmission performance, usually have physical properties (e.g. reinforced design) that allow them to be used for longer periods of time with less risk of deterioration of their characteristics. While for network certification, the use of this type of patchcord makes sense, and may even be a necessity (measurement kits may not accept patchcords other than those recommended by the manufacturer), line attenuation measurements without certification may include the use of standard patchcords, i.e. not identified as TRC. It is important that such patchcords are manufactured in min. class B, according to PN-EN 61300 -3 61300 (IEC -3-34). This means an average connector attenuation of no more than 0.12 dB and a maximum of no more than 0.25 dB. Ultimately, therefore, the measurement uncertainty involved in using such a connector will not be significantly greater than that of a true reference connector. Certainly, however, these patchcords should be periodically replaced with new ones and cleaned regularly. The length of the measurement patchcord should be no less than 2 m. The use of shorter patchcords involves the risk of an error when establishing the power reference – it may be slightly higher than it should be, and this will result in a distortion of the final measurement result to the detriment of the tester.
Single-mode patch cord PC-1372-2 SC/UPC - LC/UPC, simplex, G.657.A2, LSZH, 2m
Connector attenuation grades according to IEC 61300-3-34 (IEC 61753-1)
Grades Attenuation [dB]
A < 0.07 medium < 0.15 max.
B < 0.12 medium < 0.25 max.
C < 0.25 medium < 0.50 max.
D < 0.50 medium < 1.00 max.
ULTIMODE patchcords are manufactured in attenuation class B according to IEC 61300-3-34 (IEC 61753-1) and can be used as measurement patchcords during measurements with the transmission method.
Back to the 3 methods for establishing reference power, you already know that the 1 patchcord method is the best one, as the other methods increase the measurement uncertainty by reducing the reference power due to the inclusion of attenuation of one or two reference connections. The 2 patchcord method should be used when the connector of the optical power meter is not compatible with the connector in the switch (e.g. when the meter is equipped with SC connectors and the switch with LC adapters). Then the 1 patchcord method is not feasible and it is necessary to use two measurement patchcords (e.g. SC-LC) and a centering adapter (e.g. LC-LC). The 3 patchcord method can be used when the line to be measured is terminated with connectors. However, since this method excludes the attenuation of the start and end connectors from the measurement (see above figure), it only makes sense to use it when this attenuation is an insignificant part of the attenuation of the whole line.
Interpretation of measurement results.
Verification of the accuracy of a fiber optic system using a light source and optical power meter involves generating just one numerical value for the attenuation of the entire fiber optic path and comparing it with the expected value. When certifying a network, the expected value is specific to the application for which the link is being certified - for example, when certifying a network for a 10GBASE-LR application, i.e. a 10 Gbps Ethernet connection on single-mode fiber, the maximum attenuation of the fiber channel can be 6.2 dB for a wavelength of 1310 nm. Attenuation values for other applications can be found in documents describing the specific standard or in structured cabling standards ISO/IEC 11801, EN 50173. When the measured connection is not subject to certification, the maximum attenuation value is calculated by adding up the theoretical maximum attenuation values of all the elements included in the fiber optic path.
The problem with calculating the approximate typical attenuation of a given line arises from the lack of clearly defined attenuation standards for individual events such as splices and connectors. It may turn out that, according to one criterion, a line consisting of 2 connectors, 2 splices and 500 meters of fiber should attenuate no more than 2.3 dB, according to another 1.5 dB, and according to yet another 0.82 dB! Such divergent values have their source in various documents: structured cabling standards, company standards of large operators, international recommendations, manufacturers' standards defining attenuation classes of fiber optic connectors, product data sheets and information passed from mouth to mouth, which after some time become a kind of unscripted industry standard.
It is crucial, therefore, that the person performing the transmission measurement, in addition to the correct procedure for measuring and establishing the reference power, knows how to strictly define the criteria for evaluating its result - so that the determination of whether the system was done correctly or not is not subject to any interpretation.
The following lists the elements of the optical path to be analyzed when calculating the maximum attenuation of a single-mode fiber optic line. For each of them, acceptable attenuation values resulting from the application of various criteria are given, and the value that we feel, will be appropriate for analyzing the vast majority of fiber optic connections is indicated.
Fiber attenuation Connector attenuation Splice attenuation

ITU-T Recommendations G.652.D/G.657.A:
0.40 dB/km (1310 nm)
0.30 dB/km (1550 nm)

Fibre manufacturers’ declarations:
< 0.35 dB/km (1310 nm)
< 0.20 dB/km (1550 nm)

Large operator factory standard:
0.40 dB/km (1310 nm)
0.25 dB/km (1550 nm)

61280-4-2/ISO/IEC 14763-3:
< 0.75 dB


Norma 61300-3-34:
connector class B < 0,25 dB
connector class C < 0,50 dB


Large operator factory standard:
max. 0,50 dB, but not more than 0,30 dB on average

61280-4-2/ISO/IEC 14763-3:
< 0.30 dB


Large operator factory standard:
max. 0,15 dB, but 0,07 dB on average


Generally assumed:
< 0.10 dB

Thus, it can be seen that, depending on the criterion adopted, when estimating the maximum attenuation of a given connection, very different results can be obtained, which may result in the measurement result being correct for one and incorrect for the other. Taking into account the real attenuation, correctly executed elements of fiber optic systems, it will be more reasonable to make more restrictive assumptions. In reality, the measured values in the vast majority of cases will be significantly lower anyway. To sum up, the proposed values of attenuation of individual events to be taken for calculations are:
  • fiber attenuation: 0.4 dB/km (1310 nm), 0.3 dB/km (1550 nm),
  • connector attenuation: 0.3 dB,
  • splice attenuation: 0.1 dB
Therefore, the previously cited example of a 500 m line terminated with pigtails spliced on both sides should attenuate no more than: 0.5 × 0.4 dB + 2 × 0.3 dB + 2 × 0.1 dB = 1 dB for the 1310 nm wavelength and slightly less (0.05 dB) for the 1550 nm wavelength.
Why is it useful to take measurements at 1310 nm and 1550 nm?
Verification of the correctness of a fiber optic system built with single-mode fiber optics should include measurement at 1310 nm and 1550 nm. Even if only 1310 nm SFP inserts are to function in this network, you should be sure that in the event of changing them for e.g. 1310 nm/1550 nm WDM inserts, the network will function correctly.
Measurements for the two wavelengths may give slightly different results and highlight some problems in the system that would not be identified with only one measurement. The first factor affecting the difference in the result is the different unit attenuation of the fibre for the different wavelengths. However, this is irrelevant for short distances. Only for distances of more than 1000 m can the difference exceed 0.1 dB and it should increase linearly by about another 0.1 dB for a further 1000 m. For shorter links, the measurement results should be similar with slightly less attenuation for the 1550 nm wavelength.
If the measurement for the 1550 nm wavelength gives a worse result, this most likely indicates a macro-bend in the fibre somewhere along the route. Often this is a bend in the switch – one that is easily found with the VFL visual fault locator. A clear light leakage will be seen at the bend location. However, it may be that the fibre bend is a consequence of a cable bend somewhere along the route. If this is the case, the transmission method will not give an answer about the exact location of the damage. Verification with an OTDR is necessary.
In the opposite case, the measurement for 1310 nm gives a worse result (and the difference is greater than that due to the attenuation of the fibre), this most likely indicates a problem with the positioning of the fibre, or to be more precise, the fibre cores. As a rule, this will be a problem somewhere at the connector(s), but it could also be a matter of a poorly made splice. Of course, without additional diagnostics using an OTDR, the possible location of the fault can only be done by trial and error.
It is worth considering why a wavelength of 1550 nm will highlight fibre bends and 1310 nm will highlight inferior fibre splices. To determine this, it is necessary to look at the structure of an optical fibre and introduce the definition of the fibre's MFD (Mode Field Diameter).
Optical fibre structure. Light waves propagate in the core and some in the fibre sheath.
The structure of a typical optical fibre comprises a core and a surrounding sheath. These have different indexes of refraction (the core slightly larger), so that light introduced into the core at the right angle is completely internally reflected and propagates from the transmitter to the receiver. The physical diameter of the core is, of course, constant and can be, for example, 8.2 µm, regardless of the wavelength it carries. However, light waves do not propagate only in the core. Some of them are also transmitted in the sheath, and the area of the core and sheath that is responsible for the propagation of the light waves is the aforementioned MFD also referred to as the effective core area. It is the diameter of the MFD that is quoted by fibre manufacturers as the basic parameter of the fibre. The physical diameter of the core is of secondary importance. An example MFD value for a Corning SMF-28e+ fibre complying with ITU-T recommendation G.652.D, is 9.2 µm at 1310 nm and 10.4 µm at 1550 nm.
The fact that the MFD is different for different wavelengths may affect the measurements as described above. The larger diameter for 1550 nm means that the signal for this wavelength runs closer to the sheath border. Exceeding the minimum bending radius of the fibre will therefore result in greater attenuation for this wavelength, as some of the signal will 'escape' from the sheath more quickly. Conversely, the smaller MFD area for 1310 nm means that it will be more sensitive to the offset of the cores in relation to each other.