How field-terminated no-epoxy/no-polish connectors can support 10/40-Gbit and analog video systems on par with fusion-spliced connectorized pigtails.
BY DOUG COLEMAN
Field termination of optical-fiber cables continues to be a dominant termination method in the local area network (LAN). But questions have arisen recently on the ability of field-terminated no-epoxy/no-polish connectors to support high-data-rate digital systems and analog video systems. These questions have resulted in a number of businesses and campuses installing factory-terminated fibers through the use of fusion splicers instead of field-terminated fibers.
We have found, however, through system testing that field-terminated singlemode no epoxy/polish connectors can reliably support high data-rate (10 and 40 Gbit) and analog video systems in a manner equivalent to fusion-spliced, factory-manufactured connectorized pigtails.
The ‘no-no’ factor
The no-epoxy/no-polish (NENP) connector incorporates a factory-installed fiber stub that is bonded to the ferrule. The ferrule endface is factory polished and inspected to provide consistent low loss and reflectance performance. You can easily install the connector by inserting a stripped and cleaved field fiber into the back until the fiber is seated against the factory-stubbed fiber (Fig. 1). You activate the mechanical splice by closing the splice with a compressive force to ensure that the cleaved fibers are precisely aligned.
 FIGURE 1. A detailed look at a no-epoxy no-polish connector. |
We evaluated four backbone architecture scenarios to compare the high data-rate (10 and 40 Gbit) and analog video performances of the singlemode NENP SC connector to a spliced singlemode SC connector pigtail.
- Scenario 1 simulated a point-to-point campus and building backbone architecture with a 300-meter (m) distance from the main cross-connect (MC) to the intermediate cross-connect (IC) or telecommunications room (TR).
- Scenario 2 simulated a point-to-point campus backbone architecture with a 3 km distance from the MC to the IC.
- Scenario 3 simulated a mesh campus backbone architecture with a 6 km distance from an IC to the MC to a second IC.
- Scenario 4 simulated a point-to-point campus backbone architecture with a 12 km distance from the MC to the IC.
For each scenario, the singlemode NENP SC connectors had -40 dB reflectance and the singlemode SC pigtails had-55 dB reflectance. We also included singlemode NENP SC connectors with -55 dB reflectance for the analog video system evaluation. Each connector utilized a flat polished ferrule endface.
We installed NENP SC connectors as well as the fusion-spliced pigtails on each backbone cable, and used factory-manufactured SC jumpers with -55 dB reflectance for patch panel cross-connects and interconnects. A score-and-snap cleaver was used to install the NENP connectors, while we installed the factory pigtails using a precision cleaver and a fusion splicer. Each connector was installed by a certified contractor to represent actual field deployable conditions.
Electronics configuration
Ten Gbits/sec testing was performed at the 1310 nm wavelength in accordance with the IEEE-802.3 10GBase-LR physical media dependent using a standards-compliant transceiver. A minimum average receiver power level of -14.4 dBm is specified for a bit error rate (BER) of 10-12. A stringent pseudorandom binary (bit) sequence (PRBS) of 231-1 was used to generate receiver BER waterfall curves for the comparative analysis.
Forty Gbits/sec testing was performed at 1550 nm using four time division multiplex (TDM) 10-Gbits/sec electrical data signals with a PRBS of 231-1 to generate receiver BER waterfall curves for the comparative analysis. Scenarios 3 and 4 included 100 ps and 200 ps of dispersion compensation fiber (DCF), respectively, at the receiver to correct for accumulated chromatic dispersion penalties.
 |
To date, no standard system guidance exists for 40 Gbits/sec. A minimum receiver power level of -8 dBm is expected at a BER of 10-12.
Analog testing was performed across the 55 to 550 MHz frequency range using 77 multiplexed RF channels with transmission at the 1550 nm wavelength. We used a 6 MHz band pass filter (BPF) to isolate channels 2, 9 and 78 to perform the carrier to noise ratio (CNR) comparative analysis using the electronic spectrum analyzer.
Digital transmission
BER waterfall curves are a standard method for comparing optical transmission capability and quality of digital signals. Waterfall curves express the BER response to decreasing received power. For each test scenario, we used a variable optical attenuator (VOA) to attenuate the receive power to derive the BER waterfall curve. BER test results for 10 Gbits and 40 Gbits showed that the NENP connector performance was equivalent to that obtained with spliced factory-terminated pigtails.
Each test scenario easily conformed to a BER of 10-12 at the minimum average receiver power level for 10 Gbits and 40 Gbits. The equivalence in performance is demonstrated in the typical BER waterfall curve results shown at left.
Analog video transmission
RF carrier-to-noise ratio (CNR) analysis is the common metric for evaluating analog video system performance. CATV standards define a minimum 43 dB for video transmission.
We monitored three discrete channels to compare the CNR performance for each scenario. Reference CNR measurements were made for a back-to-back setup with transmission through a short fiber jumper and compared to each tested scenario to derive a delta ( Δ ) CNR.
 |
As with the 10 Gbit and 40 Gbit tests results, the analog video test results for each scenario showed that the NENP connectors give equivalent performance to the spliced factory-terminated pigtails, as shown in the typical CNR results listed below.
Summary
We conducted high data-rate digital and analog video testing to evaluate the system performance difference between backbone singlemode fiber cables terminated with NENP SC connectors and fusion-spliced factory-polished SC connector pigtails. Test lengths simulated typical distances and architecture topologies used in LANs.
Our test results showed the NENP SC connector to have equivalent performance to a spliced SC connector pigtail for 10-Gbit and 40-Gbit digital signals, as well as analog CATV signals.
DOUG COLEMAN is manager of technology and standards at Corning Cable Systems, Hickory, NC (www.corningcablesystems.com).
