Challenges of Testing Optical Fiber and Copper Networks

Man Test Equipment

To increase testing accuracy, it is recommended to use the same light source for testing as the source used in the client’s network. For example, if the network operates using VCSELs, the optical fiber cabling link should be tested using a VCSEL test adapter.

It is human nature to always seek the easiest path, whether in real estate, studies or on the job. True professionals learn early that “there is no gain without pain” and that success lies in intelligence, hard work and sincerity. As an information technology systems (ITS) consultant, I have found that no matter how well you write specifications or draft design drawings, the biggest challenge is always testing and commissioning. This article will not discuss challenges related to testing methods for alien crosstalk or debate on array technology testing of 40 and 100 gigabit per second (Gb/s) links. Instead, it will look at the practical challenges of testing optical fiber and copper infrastructures from a consultant’s perspective.

Testing 1 and 10 Gb/s Multimode Optical Fiber Networks

Contractors often debate when they test OM3 and OM4 multimode optical fiber networks using a light emitting diode (LED)-based testing module and get passing results. The challenge as a consultant is to educate professionals and ensure that the testing is done per the application’s requirement.

When we choose OM3 and OM4 optical fiber for an infrastructure, the driver that comes to mind is the vertical cavity surface emitting laser (VCSEL). Introduced as a cost effective multimode optical fiber transmitter for gigabit Ethernet (GbE) and Fibre Channel applications, the VCSEL is currently used for 10 Gb/s transmission. We are already witnessing networks of 40 and 100 Gb/s with array technologies (e.g., multifiber push-on [MPO] and multifiber termination push-on [MTP] connectors). VCSELs work at the 850 nanometer (nm) and 1300 nm center wavelengths with a modulation frequency as high as 10 gigahertz (GHz), which is much higher than an LED.

When we look at data rates of 100 megabits per second (Mb/s) in optical fiber technology, the driver that comes to mind is the LED, which also works at 850 nm (i.e., 800-900 nm center wavelength) and 1300 nm (i.e., 1250-1350 nm center wavelength) but at a very low modulation frequency with a maximum of only 600 megahertz (MHz).

If a test report indicates a length limit of 2000 meters (m [6562 feet (ft)]), that usually confirms that the link has been tested for 10/100 Mb/s. Per industry standards, one cannot achieve these distances using 1 or 10 Gb/s transmission over multimode optical fiber cable using LEDs. Earlier revisions of the TIA-568 standard allowed multimode optical fiber backbone channels up to 2000 m (6562 ft) long. TIA-568-C.0 has removed these distance allowances that were often confusing to designers and users since they were inconsistent with actual supported distances. The TIA-568-C.0 Annex D includes an applications chart that lists supported distances and cabling loss allowances (i.e., loss budgets). For example, the supportable distances for 10GBASE-SR, the most commonly implemented form of 10 GbE.

Sometimes the submitted LED-based test result may indicate a 2000 MHz-km, which is the modal bandwidth for OM3 optical fiber; not the actual tested bandwidth. LEDs cannot test up to that limit. For OM3 and OM4 optical fiber networks, testing should be conducted using a VCSEL driver rather than an LED driver. VCSEL drivers can truly test the actual 2000 and 4700 MHz-km bandwidth with a 10 Gb/s transmission rate. The limit for the LED is 200 to 600 MHz-km with a maximum transmission rate of 622 Mb/s.

Therefore, when choosing the type of optical fiber, it is important to know the applications that are to be supported by the optical fiber channels and the application bandwidth requirements for each of the optical fiber types being considered. High speed LAN applications (i.e., 1 GbE and 10 GbE) require the use of a VCSEL to deliver the light source. Because a VCSEL illuminates a smaller number of modes in the optical fiber than LED, the bandwidth statement for this laser-enhanced optical fiber are higher than for the LED. To increase testing accuracy, it is recommended to use the same light source for testing as the source used in the client’s network. For example, if the network operates using VCSELs, the optical fiber cabling link should be tested using a VCSEL test adapter.

Testing 1 and 10 Gb/s Copper Networks

There are two basic testing configurations in the twisted-pair copper world—channel and permanent link. The contractor’s favorite choice tends to be channel testing due the following reasons:

  • Easy to pass
  • Per TIA standards, 20 percent more headroom due to stranded cords being the weakest link › Good patch cords can pass high-loss or failed permanent links
  • Good patch cords can pass high-loss or failed permanent links
  • The same testing patch cords are moved around to all permanent links for testing › The final patched cable to the computer is never tested
  • Channel testing is void the moment the test patch cord is removed from the jack
  • The RJ-45 plug of each patch cord in the channel adapter is not taken into testing consideration
  • Horizontal channel length of 100 m (328 ft), including patch cords
Channel testing typically provides inaccurate data that often ends up being maintained in a client’s permanent record book especially when technicians use the same quality patch cords to test all links.

On the contrary, permanent link testing validates all links with test leads provided by the equipment manufacturer, which are not taken into testing consideration. It is very important to keep accurate records of data links that are behind walls and in ceilings and raised floors. It is worth spending time to achieve accuracy of testing links that are part of the building infrastructure. Some advantages of permanent link testing include the following:

  • Validates links from patch panel to the outlet
  • Mated pair (e.g., patch panel to outlet), including RJ-45 plug and IDC at each end, and the horizontal link are taken into testing consideration
  • 90 m (295 ft) of horizontal link length, excluding patch cords
  • Stringent testing parameters are sometimes difficult to pass if not installed per standards and best practices
  • No allowance for additional headroom, as no stranded conductor is used
  • Covers all installed links that may have been planned for future expansion and cannot be completed as channels
Ultimately, a copper network’s performance relies on the performance of the channel, including complete end-to-end link with patch cords. However, the advantage of permanent link testing is the assurance that the fixed cabling infrastructure meets its intended standardized transmission performance level. One cannot go wrong when selecting a permanent link testing configuration for validating networks–it is the method that I recommend to clients. Following is a checklist that can be used when reviewing test results for compliance:

1. Test equipment has the latest software version

2. Test equipment has the latest test limit version

3. Calibration of test equipment is accurate

4. Test results are submitted in PDF and native format

5. Test result cable ID in compliance

6. Permanent link testing performed on copper

7. Test result cable type (copper and optical fiber) are in compliance

8. Test results based on LED for OM1/OM2 optical fiber

9. Test results based on VCSEL for OM3/OM4 optical fiber

10. Test results based on lasers for OS1 optical fiber

11. Multimode optical fiber testing done at 850nm and 1300nm wavelengths

12. Singlemode optical fiber testing done at 1310nm and 1550 nm wavelengths

13. Bi-directional testing

14. Accurate quantity of adapters and splices

Redistributed with permission from BICSI News Magazine–May/June 2013 Issue.

The articles, opinions and ideas expressed herein are the sole responsibility of the contributing authors and do not necessarily reflect the opinion of BICSI, its members or its staff. BICSI is not liable in any way, manner or form for the articles, opinions and ideas, and readers are urged to exercise professional caution in undertaking any of the recommendations or suggestions made by authors.