Business & GA, Commercial

Avionics System Design: Living with Connectors

By Walter Shawlee 2 | August 1, 2000

Connector choices and tooling have a rather extraordinary impact on system reliability, because all connections to and from your unit have to travel through them. As a result, designers must stay focused on how these parts affect their work all during the product design cycle.

Historically, military approvals and designated connector families dictated many choices. But many other types of connectors are entering common use–though not always in a painless manner. At the same time, many manufacturers see shrinking military markets; some higher end, mil-spec connectors parts are losing second sources or easy availability. Also, many "cost-reduced" versions of existing styles are now commonly sold. These should be labeled as "performance reduced" parts as they may get deftly substituted by the purchasing department, with some unintended consequences downstream.

At the unit connection level, the bayonet shell, circular cable connector (PT06/KPT06/MS3126/MS3116, MIL-C-26482/I), initially pioneered by Bendix, has long been around. It owes much of its longevity to an outstanding fundamental design, rich selection of inserts, many suppliers and low cost. This has proliferated into many derivative designs, with MIL-C-38999 family parts (both threaded and bayonet) being the latest and most sophisticated, offering "scoop-proof" designs to prevent damage to male contact inserts. This family has gold pins; it can accept coax and high power inserts, and now economical PC pins from Air Electro that allow direct board attachment, plus a choice of solder or crimp technology at the low end.

From older parts to newer ones, these designs fit the bill very well for everything from low-level signal to high-power applications (up to 12/10 -gauge wire), wherever a cable-end connector is used. A key feature of this family is that correct, full mating is clear (clicks into place), and mis-mating, even by a confirmed three-star doofus, is pretty close to impossible. Field repair or rework of crimp versions often is quite simple with the correct tooling, and individual contacts can be replaced easily, even in close quarters. Full environmental sealing or hermetic sealing also is available, allowing these parts to be used in many of the aircraft’s environmentally hostile areas.

Some wiring problems can occur due to difficulty in correctly identifying pins in the spiral insert layout. Suppliers include Cannon, Amphenol and FCI, among others.

The earlier, circular threaded connectors (MS-3102/3106, MIL-C-5015) still remain popular for many instrument connectors and devices with higher circuit currents like inverters, regulators and generators. These solderable, silver-plated contacts are suitable for high currents. The connectors come with many large-diameter/high-current contact inserts, as well as inexpensive two-, three- and four-contact inserts for simple instruments. Later derivative connectors have crimp contacts for easier termination. Good quality parts are available from Cannon, Bendix, Amphenol, DDK and others.

Keep in mind, these plugs do not have a good user indication of correct mating, and their silver contacts are not suitable for low-level/low-voltage signals. Some care also is required to prevent contact corrosion and oxidation with silver contacts.

Plastic Alternatives

All-plastic alternatives to metal circular connectors have become common in the commercial world, and a few have migrated to the aircraft industry. This cost-reduced design often is used with stamped contacts and should be employed with care. They can be used in some instances as an effective subassembly disconnect or cable splice in a non-critical area, but have only medium mechanical strength without strong cable clamping or retention capability. Correct crimping also requires the exact vendor specific tool and correctly sized contacts.

D-subminiature connectors, originally from Cannon, also have existed a long time, (40-plus years) and are cost-effective for some circumstances. They are basically a shrink version of earlier DPX connectors, used for larger module and tray assemblies. With gold pins, PC pins, and the ability to accept coax and high-power inserts, D-subminiatures often are applied both internally, in larger units for module connections, and externally for airframe connections.

The main problem with these connectors, however, is the poor locking mechanism. It has improved little in 40 years, other than the new one-handed lock from Positronics, which was an outstanding bit of engineering. Many a technician has been heard cursing as he or she drops the jackscrews from a D-sub connector down into some darkened part of an aircraft, or accidentally undoes the jackposts and sends loose conductive hardware drifting aimlessly into the bowels of the equipment.

Many new versions of this connector come from the cost-driven computer industry and have minimal gold plating, as well as formed/stamped contacts that withstand only a few mating cycles. Real care is needed in selecting the right version of D-sub connector to avoid long-term problems. Not surprisingly, the mil-style versions work great for years, and the incredibly low priced consumer versions are quite poor, and hard to terminate. High quality parts, well suited for aircraft use are available from Positronics, Cannon, AMP and Amphenol, among others.

The D-subminiature connector design holds interesting characteristics: the linear distributed layout and good access to pins, particularly in two-row designs. This can provide useful control of signal routing and coupling to good effect. ARINC 600 connectors also offer this characteristic. Circular connectors, especially when dense, can suffer from much cross-talk and unwanted coupling of low-level analog signals.

There has been confusion over D-sub part numbers. Even some computer-aided design (CAD) libraries are incorrect, making matters worse at the design level. The DB-25 is a 25-pin connector, but other sizes are not "DB" parts; they are DE-9, DA-15, DC-37 and DD-50, just for reference.

Double Density Designs

Double density designs also exist, like your VGA monitor connector. But these do not follow the original standard scheme, and have different layouts and numbers from different vendors, adding new steps to the misprocurement dance. Standard parts have 20-gauge contacts, as a default, and double density versions take 22-gauge contacts. Keep in mind the double density 15-pin connector has to fit 15 wires through a tiny backshell the same size as a DE-9 connector, sometimes not such an easy trick in the airframe.

The original rack and panel style of Cannon’s DPX connectors for tray-mounted boxes are now largely converted to ARINC 600 style connectors in current designs, with generally good performance. Mounting equipment in a tray assembly with mating connectors attached can be awkward. The degree of awkwardness depends on how well the equipment is secured to the tray and how well the tray’s overall design facilitates the connector’s correct alignment and engagement. This is no simple problem. The large sizes of units compared to the small target connector and pins easily can illustrate how things can go wrong quickly–and with considerable destructive force.

Connector damage through mechanical forces remains a problem both with loose assemblies and trays. High-end tray makers like Box-Mount and Barry take considerable pains to control this, but not all tray designs, guides and locks are equal.

On a service level, trays with integral connectors have the bad habit of being secured so that the rear connectors’ repair often is problematic, sometimes verging on plain impossible and unreachable. I recently saw two tray-mounted autopilot computers in a helicopter that looked as if the tail boom had been built around them, making field service a flat-out nightmare. Surprisingly, no placard exists advising that articulated tentacles are needed for effective service.

Most high-cost interconnect repairs that I’ve seen at the airframe level stemmed from tray-mounted unit problems–something to keep in mind when planning installations and shelf design. Of course, trays can be used for effective mounting without having integral connectors. Box connections can be made via front cable connectors–often a very attractive and less costly alternative to a captive connector built into the tray back.

Coax connectors have their own strange quirks. One little known oddity is the very shallow zone of engagement in many units. This may give the impression that they are fully mated mechanically, while still being unconnected electrically.

Especially notorious are SMA connectors. They require a wrench for correct mating, and must be fully threaded and "wrench tight" to achieve electrical connection. This requirement can be missed, as the zone just prior to engagement may result in enough capacitive coupling to give the impression of functionality. But a scan with a network analyzer would show massive resonances and ripple due to the series air gap capacitor, all of which are highly vibration sensitive–an ideal recipe for disaster in an aircraft system. Our lab tests also showed the same phenomena with TNC connectors (although not as severe), but type N connectors achieve connectivity virtually when first coupled. By the way, a quick clue to mating technique: all coax connectors with hex wrench flats require a wrench, while coax connectors with round-knurled barrels are intended to be tightened by hand.

Another interesting bit of connector trivia: almost all connectors have separate mechanisms for electrical and mechanical mating. If you look inside a type N, BNC or TNC, you will see a grounding contact cylinder, not related to the outer mechanical lock. Only SMA connectors, of the commonly used types, do not have this; they depend on the threaded outer shell for both ground and connectivity–which is why they can exhibit so many odd characteristics when improperly mated.

Most people report consistently better connector performance with crimp style coax connectors–and this mirrors our own experience. However, you must have the correct dies for both the connector ground and center contact, and you should always perform a pull test on the finished crimp connections to assure they are tight. Rigid lines and some microwave fittings often still must be done via soldered assembly.

While many internal systems use IDCs (insulation displacement connectors) on flat cables, some risks can be encountered during their assembly. If the cable is pulled with any tension across the connector, and no cable strain relief is applied, the inner knife contacts will shear the cable, resulting in numerous internal failures. Use of a tester such as the Cirrus cable test fixture may help control these problems, but care in assembly is important to avoid damage, and strain reliefs are vital to prevent damage later.

Teflon ribbon cable does exist, for those looking to source something other than PVC insulated ribbon wiring, as do fully mil-qualified connectors at slightly increased pricing. Vendors include AMP, Berg, 3M and Samtech, among others.

Grounding

Grounding is a final issue regarding your overall connector plan. A common and somewhat untidy practice has been to tie shields and other ground connections to the cable clamps of circular connectors and other types with a convenient screw. This is problematic, as actual resistance to ground can be significant and may take place through several aluminum shell junctions, paint and grease, all of which will be subject to corrosion and moisture.

Always provide an actual pin in the connector for grounding shields, or direct the ground wires to a local clean airframe ground close to the cable. This will give much better long-term performance, as well as providing the correct grounding for the lines, whether or not the box is present, which can be an important issue.

My choices generally tend to be for a bayonet circular connector, if budget permits and circuitry allows, or a D-sub with improved locking hardware, if many contacts in a controlled geometry are needed. Your own designs and concerns may lead you to the same choices or elsewhere. But keep in mind, everything in your system eventually attaches via these choices.

Walter welcomes reader comments and may be reached by e-mail at walter2@sphere.bc.ca .

Receive the latest avionics news right to your inbox