An engineering team examines potentially dangerous electrical failures of insulated flag-type connectors in portable electrical appliances.
By Scott G. Davis, Andrew Diamond, Will Gans, Peter Hinze, and Harri Kytomaa
 |
In over ten years of manufacturing, a problematic pattern began to emerge for five different models of portable electrical appliances. The appliances were returned for service to the manufacturer with electrical failures. Initially, the exact cause was unknown despite preliminary investigations by the manufacturer. When the recurring failures became associated with a fire hazard, however, the manufacturer immediately assembled an engineering team to look in to the cause of the failures.
To begin the investigation, the team inspected twenty-seven service-returned appliances, revealing a consistent pattern of failure. Specifically, the failures occurred on the insulated crimp connectors on the power cord or neutral wires, and were characterized by discoloration of the crimp connector insulation (see Fig. 1). These observations were consistent with overheating of the connector at the junction between the wire and the crimp connector. In a few cases, the connector insulation and surrounding material ignited, resulting in fire damage to the connector and unit. Determining the cause of the connector failures required a detailed investigation of the crimp connections.
 FIGURE 1. Insulated crimp connectors showed varying levels of heat damage (left) and burn damage (right). |
The appliance design called for 16 AWG and 18 AWG stranded multiconductor wiring, rated to 105°C. The 16 AWG wires carried the full electrical load (nominally 1,500 W) and were comprised of the power cord and neutral wires. All power wiring was fastened to the appliance with crimped, quick-connect flag-type connectors. The manufacturer outsourced assembly of all wires and crimp connections to overseas suppliers.
The appliance power cord used insulated crimp connectors exclusively, while the internal neutral wires had either insulated or uninsulated crimp connectors, depending on model design. Electrical measurements conducted on new units supplied by the manufacturer indicated a maximum load current of 12.5 A on the 16 AWG power cord, neutral wires, and associated connectors. Under normal operation, the appliance was designed to operate under cyclic conditions (meaning the appliance shut off when a given set point was reached); however, the appliance continuously operated at maximum load in certain extreme cases when the set point was not attained.
Investigation
The team examined twenty crimped connectors from the service-returned units under a microscope to qualitatively determine the crimp quality. The twenty connectors chosen were representative of all of the 16 AWG connectors found in the appliance (power cord and internal neutral wire, insulated and uninsulated connectors) at various degrees of failure. In addition to the twenty connectors from the returned units, the engineers examined four unused connectors. These unused crimp connections were randomly chosen from batches that were pre-assembled by the overseas suppliers and recently delivered to the appliance manufacturer for installation.
To prepare the connectors for microscopic examination, they were encapsulated in an epoxy resin and cured for approximately ten hours. With the epoxy resin set, the connector was sectioned roughly in half at the area of interest, revealing the cross section of the crimp (see Fig. 2). Each cut surface was polished to a 0.05-µm finish. After polishing, the crimp cross sections were examined and photographed under an optical microscope with a 6X objective lens.
Observations
All observed failures occurred in the crimps of the insulated power cord or neutral wire connectors, which carry the maximum current load. These failed connectors show evidence of prolonged overheating, and in certain cases, fire damage. Furthermore, the failed connectors consistently exhibited other notable defects that were only detectable using the cross-sectional analysis, such as limited deformation of the crimped connector, limited deformation of the conductor wire strands inside the crimp, and significant void fraction (limited contact surface) of the conductor wire strands within the crimped connector (see Fig. 3).
The sectioning also revealed that some connectors consisted of stranded wire that was pre-soldered before crimping, which was not specified by the manufacturer. Wire that was pre-soldered resulted in minimal deformation of the crimp, and reduced contact area between the conductor and the connector. Another notable observation revealed that failure did not appear to be wire specific, because failure was observed in both the stranded wire and the pre-soldered stranded wire configurations. Failures did, however, appear to be connector specific with no failures observed in uninsulated quick-connect flag-type connectors. More specifically, the crimp quality in uninsulated connectors was observed to be superior to insulated connectors. Uninsulated crimp connectors exhibited increased deformation and proportionately increased surface contact and a lower void fraction as compared to their insulated counterparts (see Fig. 4). The curvature of the crimped connector is indicative of a greater crimping efficacy, as demonstrated by the degree of deformation of both the connector and the conductor strands within.
In addition to the failed connectors, unused insulated connectors from the power cord and neutral wire assemblies as received from the overseas suppliers were sectioned and examined. These connectors exhibited the same poor crimp quality and high variability between crimps seen in the failed service-returned connectors. The poor crimping was characterized by limited contact between the conductor wires and crimped connector.
Analysis and conclusions
Failure of the appliances was due to improperly formed crimps in the insulated connectors. The uninsulated connectors did not experience failure. The crimp quality, as qualitatively evaluated by optical microscopy of the crimp cross section, was markedly higher for the uninsulated connectors than the insulated connectors. Poor crimping in the failed connectors led to a decrease in contact area between the connector and the wire conductors, which in turn was accompanied by an increase in the associated contact resistance. This caused additional heat generation at the crimp, and, over time, failure of the connector and the unit. In certain cases this overheating led to ignition of the connector and unit, resulting in a fire hazard.
Contact resistance varies as a function of time and depends on the size and shape of the mating contacts, vibration, temperature, moisture, and other factors.1 Contacts exposed to long-term high temperatures experienced the buildup of an oxide layer, which progressively increases contact resistance, further increasing the temperature. Units that exhibited inadequate crimping did not necessarily fail, as appliance failure was strongly dependent on the cycle use of the product.
Recognizing that price competition often results in the outsourcing of electrical components, this study showed how important it is not only to use reputable suppliers, but also to consider incorporating destructive testing methods in electrical component evaluation. Destructive testing is particularly helpful due to the inherent difficulties in visually evaluating the crimp quality of insulated quick connectors. Examina-tion of randomly sampled crimp connections using the microscopic inspection technique as employed in this study can provide valuable information regarding the quality of the supplier's assembly process. These measures help to ensure the reliability of crimp connections and greatly reduce the potential risk of product failure and, in some cases, fire hazards.
REFERENCE
- Medora, N. K., Electronic Failure Analysis Handbook, New York, NY: McGraw-Hill (1999).
SCOTT G. DAVIS, ANDREW DIAMOND, WILL GANS, PETER HINZE, and HARRI KYTOMAA are with Exponent Failure Analysis Associates, 21 Strathmore Road, Natick, MA 01760. Tel: (508) 652-8557; email: daviss@exponent.com.
