A cleaning standard for fiber optic connectors could save industry millions of dollars annually. Contamination of connector end faces — in the form of something as simple as dust particles or fingerprints — can change light propagation through the fiber, degrading device performance and causing data errors. To avoid this, it is common practice to clean fiber optic connectors prior to assembly and test, adding to the cost. However, there is no industry standard for connector cleanliness. As a result, the devices can be either overcleaned, which consumes manpower and materials, or undercleaned, which can lead to expensive rework and/or failure in the field. By agreeing to a set of standardized criteria, industry can define a level of cleaning that will assure performance and reliability and, in many cases, will reduce the delays and expense of additional cleaning and inspection. It also is much easier (i.e., less costly) for a supplier to meet a single set of standardized guidelines than to have to meet different specifications for each customer.Cleaning may sound like a simple matter, even mundane, but it adds up to big bucks. The telecom sector alone will spend an estimated $223 million cleaning connectors in 2004. This number is based on an annual volume of connectors and receptacle ports, the cost of the cleaning process, and how often the cleaning process is performed during manufacture and in the field. Connectors and receptacle ports are inspected at each step of optical assembly. Cable manufacturers typically inspect and clean their products at least twice during the manufacturing process. These same cables will be inspected again when received by the OEM or electronics manufacturing services provider, and yet again by the field installer. Once materials leave the supplier, an estimated 10 percent will not meet cleaning criteria at all of the subsequent levels, requiring a three-step process of inspection, cleaning and then reinspection (Figure 1). Figure 1. The contaminated connector (left) shown in this fiberscopic image is cleaned with a cassette (center) and, when reinspected, is found to be clean (right).For connectors, this process costs an estimated $1.43 per connector, which includes three minutes of labor at $25 per hour and cassette cleaning material at 18 cents per operation. For female optical ports on optical modules (i.e., receptacle ports), the process costs about $3.71 per component. The cleaning process of one receptacle port requires an average of $2.25 for material and 3.5 minutes of labor. The telecom sector is expected to use nearly 35 million fiber optic connectors this year; add receptacle ports, and the number becomes significantly larger. If a customer’s cleaning specification is more stringent than it needs to be, more components will fail inspection — even though they may perform satisfactorily. Such cases lead to unnecessary cleaning and reinspection steps that add time and cost to the production cycle.Rework as a result of undercleaning is even more expensive. It costs approximately $20 to reterminate a pigtail device, and rework can cost significantly more if the failure occurs at board- or box-level assembly. Failure in the field can cost hundreds or even thousands of dollars, depending on the complexity of the product.Figure 2. This connector has particles within the critical 25-μm zone (circled in red), which, according to NEMI testing, significantly degraded signal performance. If this connector had been accepted (undercleaning), greater insertion loss and reflectance caused by contamination would have led to functional failure. At 1550 nm, the insertion loss on the connector shown here was 0.4 dB and the reflectance was –44.8 dB. In comparison, the insertion loss of the cleaned connector was 0.23 dB and the reflectance was –50.7 dB.The National Electronics Manufacturing Initiative (NEMI), an industry-led consortium of more than 60 electronics manufacturers, suppliers and related organizations founded to facilitate leadership of the North American electronics manufacturing supply chain, has set up the Fiber Optic Signal Performance Project. After developing data over the past two years to support a standard cleaning specification, the group is working to define what level of cleaning is necessary for performance, without being “too much.” Its research indicates that the effects of contamination and scratches on signal performance depend on the location, number and size of the contamination or scratches. Based on these data, the specification that NEMI is developing will define “zones” and acceptance criteria for scratches and contamination within each zone. The group has found that the most critical zone is within a radius of 25 μm around the core. Particles inside this area, even if not directly on the core, can significantly increase insertion loss (up to 0.5 to 3 dB) and increase reflectance (up to –30 to –20 dB). However, particles outside the critical 25-μm zone do not significantly affect optical signal performance. Figure 3. The contamination shown on this connector is outside of the critical 25-μm zone (circled in black) and did not affect performance. However, under some customers’ cleaning criteria, this connector would have failed initial inspection, requiring cleaning and reinspection (i.e., overcleaning), and leading to additional, unnecessary costs.Scratches on or across the core typically increase reflectance by 4 dB but do not significantly affect insertion loss. Those outside of 25 μm have no significant influence on insertion loss or reflectance. NEMI is collaborating with the International Electrotechnical Commission (IEC), the Telecommunications Industry Association (TIA) and IPC-Association Connecting Electronics Industries to develop a cleanliness standard. Specifications will be jointly submitted to IEC for incorporation into IEC 61300-3-35, and to IPC as a subordinate standard to the existing IPC-0040 standard (Optoelectronic Assembly and Packaging Technology). In addition, the project plans to collaborate with TIA and IPC on the development of cleaning methods and contamination assessment for Level 1 and 2 optical assembly.Meet the authorsTatiana Berdinskikh, engineering adviser for Celestica International Inc. in Toronto, is co-chairwoman of the NEMI Fiber Optic Signal Performance Project; e-mail: tberdins@celestica.com. Heather Tkalec, test engineer for Alcatel