Connecting hardware forms a significant part of an installation, but often it gets only little attention compared with cables when a LAN is planned. However, it should be realized that the connecting hardware is in many respects the weakest point of the transmission line. Even a very fine cable performance may be destroyed at this point, both because of bad connecting hardware performance, missing plug intermateability, installer failures and last but not least a sometimes poor high frequency crosstalk performance of the RJ 45 connector interface (if applicable). Consequently the transmission performance of the installation is often depending on understanding of performance characteristics of the connecting hardware and correct termination procedures of the cables.

The significant issues to consider when selecting connecting hardware for copper cabling is connector interface performance, EMC performance and termination of cable screens (if applicable). The considerations for selection of cabling type for a LAN are not included in the present discussions but are found in the similar section for cables.

Only connecting hardware specific issues are included in the below brief introduction which should aid selection of the optimum connecting hardware for an installation and to avoid the most frequent connecting hardware related installation problems.

Connector Interface Performance

The RJ 45 connector forms the predominant interface for generic cabling. It is used both for wall outlets, where it is the only specified Cat. 5 interface, and traditionally also for patch panels. However, the RJ 45 connector was developed for very low frequency applications and its high frequency performance is not surprisingly very poor. This creates problems for the connecting hardware producers which have to compensate for the bad crosstalk performance in their circuit design.

The user should of course trust his supplier that he has solved the compensation challenges. However, a short description of the fundamental problem and its consequences will help to understand performance capabilities and restrictions.

The basic construction of the RJ 45 connector is long parallel contact pins without any screen separation between the pairs. The pairs 3-6 and 4-5 even have a crossover to allow the designed contact sequence being, contacts 1-2 for one pair, contacts 3-6 for a second pair, contacts 4-5 for the third pair and contacts 7-8 for the fourth pair. Using common sense one would expect (and usually find) fine crosstalk performance between the non-neighbour pairs 1-2 and 7-8, medium to low performance between neighbour pairs and low performance between crossover pairs 3-6 and 4-5. This makes compensation necessary, primarily between pairs 3-6 and 4-5. The extent of compensation is increasing with growing demands to crosstalk performance. This has the following consequences when more than the Cat. 5 performance is required:

Connecting hardware may have critical far end crosstalk (FEXT) performance in spite of a fine near end crosstalk (NEXT) quality. FEXT is becoming important for future high speed full duplex applications. FEXT should therefore be evaluated for all connecting hardware intended for high speed applications, especially Cat. 5+ and Cat. 6.

The RJ 45 plug performance is critical for Cat. 5+ and especially Cat. 6 connecting hardware. The end user should carefully evaluate any limitations specified by the connecting hardware or system supplier which might only allow one source or very expensive patch cord products. Generally, all manufactures of Cat. 6 patch cords have different performance, and intermateability of products is most likely not possible.

Documentation for any claimed Cat. 6 connecting hardware having RJ 45 interface should be critically evaluated. The presently proposed limits are extremely difficult to pass.

Cat. 7 connecting hardware with RJ 45 interface can even in the best case only pass requirements for the outer pairs (1-2 and 7-8).

Cats

The number of connecting hardware ratings have been extended by three new categories and standardisation of these new performance ratings is in rapid progress. Presently the four cable categories Cat. 5, Cat. 5+ (Enhanced Cat. 5), Cat. 6 and Cat. 7 are included in the 3P verification programme.
Discussion of the different categories of connecting hardware is found in the section about standardisation.

When evaluating the need for a specific category of connecting hardware for an installation it should be noted that any higher category will always cover all categories having a lower rating. The only exception concerns the interface, where a higher category may require another interface to allow full four pair transmission (presently only applicable for category 7).

The Cat. 7 interface has not yet been standardised, but will possibly be a new connector design which is targeted for standardisation in 1999.

Termination of Horisontal Cable Screens

Unfortunately much screened connecting hardware is based on a technically incorrect, but commercially attractive solution for termination of the horisontal cable screen to the connecting hardware screen, i.e. through a cable drain wire. Whether or not this failure is present can easily be verified when removing the cover of the connecting hardware. A drain wire will in this case be terminated to a connecting block or other fixed contact on the circuit board of the connecting hardware.

You should never purchase this type of connecting hardware without solid evidence about EMC performance, for instance passing of transfer impedance requirements. 3P has found EMC problems for all such types of screen terminations.

EMC Performance

EMC Performance of unscreened connecting hardware is mainly determined by the balance of the circuit. Many manufactures have not focused on this issue and consequently the EMC performance of most unscreened connecting hardware is 4 - 20 dB (3 - 100 times) worse than for the connected cables.

The EMC performance of the connecting hardware, and not only the cables, should be evaluated, especially if high speed applications and/or high EMC loads of the environments are expected.

EMC performance of screened connecting hardware is also determined by the balance, but more important by the overall screening. Screening performance requirements (transfer impedance) must of course be passed for screened connecting hardware. However, if data are not available or are not trusted the following rough guidelines can be used to conclude evidently bad EMC performance:

DC resistance through the connecting hardware, i.e. from patch cable screen to horisontal cable screen, must be low and stable. This means that the same milliohm level must be measured for all ports, also by repeated measurements after handling.

No slots should be present, for instance between metal parts. Overlapping metal should always be used between different screen parts. Contrary to this, small holes (approx. 1 cm in max. dimension) are normally not critical.

The screen of the jack should have connection directly to the external screen of the connecting hardware. No passing through circuit parts should normally be allowed for screen terminations.

Correctly terminated screened connecting hardware can be expected to have EMC performance min. 15 dB (30 times) better than unscreened. Badly terminated screened connecting hardware often has EMC performance equal to the one of unscreened. However, as exceptions to these guidelines may be significant, the possible improvement of screen terminations should be evaluated with the supplier if EMC performance is critical for running the application in question.