How to . improve EMC performances of cable installations.

The major conclusion from the study is that cable trays are an integral part of the ground structure (actually, it forms the ground if correctly done; nothing else is needed), which can perform as a shielding structure for the enclosed electronic systems. The major quality parameter is the connections between the tray parts and the corresponding cabinets. Furthermore, it is found that separation of cables on trays provides little improvement as long as there is no well defined ground structure.

Introduction

The use of cable trays for handling of cable routing is mainly focused on the mechanical and geometrical aspects of finding a proper way of routing cables between different systems and apparatus in an installation. From an EMC point of view, however, we must also consider the electrical characteristics of the configuration. This work was initiated by the Wibe Corporation – a fully owned subsidiary to Schneider Electric – to address the question: is there a difference in the EMC performance between different cable tray types and the electrical connections between mechanical members of the system?

In order to answer those questions, we need to expand the question and consider the actual meaning of EMC performance of trays.

 

The definition of EMC requires a system that can cause interference or be interfered with, i.e. some type of electronics that is performing a function. A cable tray, however, is usually a metal structure that is supporting a set of cables (which in turn do not contain electronics). In order to analyze the EMC performance, we must include the entire system including the electronics sub-systems, the interconnecting cables, the enclosures for the electronics, and finally the metal trays enclosing the cables. The cable trays may thus be a sub-part of a more or less shielding ground structure. The effectiveness of the shield is dependent on the entire quality of the sub-parts and – most of all – their installation.

 

The importance of installation quality is emphasized in the EMC directive 2004/108/EC [1], which states that the installation shall be designed using good engineering practice. However, there is very little information on what this good engineering practice consists of. Specific technical advice may be found in modern literature [2], though. Focusing on cabling systems, information is also found in installation standards such as EN 50174-2 [4]. Technical specifications on this aspect are also found in the EMMA handbook [3]. The intention of this article is to describe the impact on EMC with different installation quality.

 

This article contains the test results from shielding and crosstalk performance testing that was performed on a selection of cable trays and their installation. The purpose of the testing was to

• quantify the shielding and crosstalk performance of different cable tray systems.
• quantify the impact of poor versus optimal installation of trays and cabinets on EMC for a system.

Theoretically, a cable tray installation is usually large compared to the wavelength of most types of disturbances. If the wavelength of the disturbance is comparative to or smaller than the size of the installation, resonance may occur which degrades EMC characteristics.

On the other hand a cable tray system may improve the EMC characteristics if the system is small compared with the wavelength. Thus, a cable tray system may mainly improve the EMC behaviour of an electrical or electronic system from 0 Hz up to at the most 10 – 100 MHz, depending on the system size. In the low frequency range 0 – 100 MHz we find for example transients due to short circuits of the power system, leaky power frequency currents, leaky currents from switching devices, harmonics to the power frequency, lightning related transients, radio frequency emissions and susceptibility, and more.

Do not miss the related article ”Backgound and Comments on the EMC Performance of Cable Trays”

Definition of test items and configurations

A. Test items

The test items were cable tray sections of DEFEM wire trays, Wibe cable trays and Wibe metal trays, as listed in TABLE I. Also a main part of the test set up is a sheet of metal used as a ground reference plane (GRP). Commercial cabinets were used to make the test object similar to typical installations. The paint was removed from the bottom plate, so that low impedance connection was maintained between the measurement equipment, the cabinets and the ground reference plane (GRP). The paint was also removed at the interconnection points for trays, cable terminations, and cable connectors.

B. Cable tray configurations

A metallic sheet (GRP) was placed on the floor underneath the test set up to simulate a building ground system and to stabilize the measurements due to an electrically undefined floor. The cable tray structure was connected to this GRP by one or more wires or plates from the ends depending on test set up. The GRP formed the reference for the measured signals. Different cable tray configurations were used for different measurements. The cable trays were laid on consoles mounted on a wall of plasterboards. Three main configurations – A, B, and C – were used:

• A consists of one 5 m long section connected to GRP with a wire in each end, see Figure 2. This is the so-called typical tray installation. The word “typical” relates to the practice of applying equipotential bonding by use of long wires. In these tests, the wires are rather short in comparison.
• B consists of one 5 meters long section, and two vertical sections closely connected to the GRP. This is the so-called optimal tray installation. The word “optimal” relates to making a low RF impedance connection, in contrast to the equipotential bonding practices as above.
• C consists of one 5 m long tray section connected to two cabinets. This is the so-called relevant total installation. Relevant, in this case, means that we try to resemble a real life installation. The installation configuration C is used with five variants:
• CA: tray and cabinets interconnected with each other with pieces of trays making the shortest possible connections with the lowest interconnect impedance, as shown in Figure 3.
• CB: tray and cabinets connected with two medium long wires (approx. 10 cm).
• CC: tray only connected at one end to a cabinet with one single medium long wire (approx. 10 cm).
• CD: trays connected to the cabinets with very long connections, simulating connection to the equipotential bonding bar in a cabinet, see Figure 4.
• CE: tray and cabinets connected with shortest possible connections (as CA), and the GRP is removed. Reference measurements were performed with test cables on wooden supports.

II. Development of test methods

The characteristics measured and test methods chosen were based on discussions between WIBE and EMC Services, since no specific EMC test methods are available for cabling systems. Several types of tests were developed and performed. These results are documented in [5]. When summarising the results, we found that the following tests were the most descriptive, providing the best information on the EMC performance of the system:

• Common Mode (CM) crosstalk between two cables placed within the configuration.
• Field emission reduction (i.e. shielding effectiveness) from a CM source within the configuration. Field susceptibility was not chosen because it should give the same answer as field emission reduction, i.e. shielding effectiveness works in both directions.

A. Crosstalk setup

The CM coupling between a power cable and a signal cable (not shielded) was measured with a vector network analyser in the frequency range 20 kHz to 200 MHz. The wires of the power cable and the wires of the signal cable were each short-circuited in both ends. One end of each cable was connected to the ground plane (GRP) via resistors of 47 ohm. The output terminal (channel 1 selected) of the network analyser was connected to the power cable and the input terminal (channel 2) was connected to the signal cable, see Figure 5.

The reference terminals were connected to the ground plane via a shielding connection box. Near end cross talk was measured. General cable positions were: a = 25 mm, b = 150 mm, see Figure 6. A view of the configuration C (relevant total installation) is found in Figure 7.

B. Shielding effectiveness setup

The emitted electric field due to injected CM signals in a signal cable was measured to evaluate the effect of different cable tray designs. A signal generator (a tracking generator) connected to a power amplifier drove a CM mode signal in a centered 2-wire cable shorted in both ends, with one end connected to the ground plane (GRP) via a 47 ohm resistor and the other to the amplifier, see Figure 9.

The antenna was connected to the spectrum analyzer (SA) via a pre-amplifier. The vertical emitted electrical field was measured with a rod antenna (as specified in EN 55025 [7]) in the frequency range 100 kHz – 30 MHz at a distance of 3 m perpendicular to the centre of the cable tray.

 

The antenna height above the ground was 1 meter. The vertical and horizontal emitted electrical field was measured with a biconical antenna in the frequency range 30 MHz – 200 MHz at a distance of 3 m perpendicular to the centre of the cable tray, which is similar to the method in EN 55011 [8]. The antenna height above the ground was 1 meter. A view of this setup is given in Figure 8. Generator and antenna measurement set-up for shielding effectiveness measurement, biconical antenna shown (standby rod antenna to the left), measurement on configuration C, tray KHZ.

III. Crosstalk measurements

Some examples from the crosstalk measurements are displayed below. The effect of cable separation is displayed in Figure 10. Here, the typical configuration, A, was used.

The results show that there is practically no difference in the results for different separation distances, nor for the difference in tray type. The result is limited by the relatively high RF (Radio Frequency) impedance in the tray connection to GRP. Crosstalk measurements were made on the mesh tray (Defem) in the relevant total installation C, as presented in Figure 11.

Here, the results vary drastically with installation type. With a single point grounding of the tray, we even obtain an amplification of the crosstalk compared to a wooden board. With improved RF connection, the crosstalk is decreased accordingly. With the removal of the GRP, we found very little change compared to the continuous tray configuration, implicating that the tray system in itself is the primary ground structure for the test setup.

 

In the graph for the KHZ system, we see the similar pattern but with lower reduction, see Figure 12. However, the amplification is somewhat reduced in the single point ground result. Above 10 MHz we find a strong resonance behavior, indicating the general problem of making high frequency measurements with large structures. In addition, a poor quality tray installation – from an EMC point of view – increases the resonant behavior.

IV. Shielding measurements

Some examples from the field attenuation (i.e. shielding) measurements are displayed below for the relevant total installation C. The attenuation is calculated as the quotient between the measured field for the given configuration, and the field from the reference measurement using a wooden board. The results for the W1 tray system in the range 100 kHz – 30 MHz is displayed in Figure 13.

For the single point ground we obtain negative shielding, i.e. amplification of the emitted field. However, with improved RF connection, the field is decreased accordingly. Similar to the crosstalk measurement result, we find only a small change compared to the continuous tray configuration with the removal of the GRP. The results for the W1 tray system in the range 30 – 200 MHz is displayed in Figure 14.

Due to the large system size (compared to the signal wavelength) the measurements show a resonant behavior. Still, it is possible to analyze differences between the configurations in a broader sense. Here, we find very little improvement of the shielding when using the wire ground connections. Only when applying a continuous tray connection, a reduction in the range of 20 dB is obtained. This means that cable tray systems may provide a shielding effect even up to 200 MHz, but only with optimal low RF impedance connection between tray members and cabinets. The corresponding result for the Defem tray system is shown in Figure 15. The results are similar to the ones obtained for the W1 tray, but with slightly reduced values.

The results for the KHZ system showed similar behavior but with lower shielding values overall – but with very low field amplification for the single point ground variant. For frequencies above 30 MHz, the field reduction was marginal.

V. Conclusions

A well designed tray system with optimal performance should provide as low crosstalk as possible. In the measurements on the configuration C, the results show that a poorly implemented tray system amplifies crosstalk between cables. The crosstalk reduction is improved with reduced length of interconnects between trays and cabinets (equaling lower connection impedance). The measurements also showed that the crosstalk is independent of cable separation if the RF ground impedance for the system is poor (= high).

 

The results of the measurements show, that single wire grounding of trays may amplify the field emission at medium frequencies (100 kHz – 30 MHz). Short wire grounding at both ends (configuration CB) gives some improvements, while long wire grounding (configuration CD) has a very small effect. At higher frequencies, only the continuous trays using W1 or Defem gave any field attenuation.

 

When looking at the system configuration and the corresponding results, we find that the cable tray system (including the cabinets at both end, the interconnects, and the cable tray) very much behaves as a shielding enclosure for an electric or electronic system. The behavior is in practice identical to that for a regular shielded system – consisting of shielded enclosures, shielded connectors and cable shields. For this type of system, the shielding effectiveness is limited by the quality of the shielded connector. There is little use of a high quality cable shield if a high quality connector is not specified. The corresponding behavior of a cable tray system has been found in this measurement study. With this in mind, the following conclusions may be drawn.

• Cable trays are an integral part of the ground structure, which can perform as a shielding structure for the enclosed electronic systems. If well designed and implemented, it actually forms the ground system.
• The major quality parameters are the connections between the tray parts and corresponding cabinets. This ground plane structure (trays and cabinets) must be greater than the protected system (i.e. no cables shall be outside the shielding structure).
• Separation of cables on the tray is futile as long as there is no well defined RF ground structure.
• If there is no high frequency connection between trays and cabinets, a solid tray provides an amplification of disturbances in the system. In such a case, a plastic tray is preferred and a coarse grid tray is a good neutral option.
• With long wire grounds, so called equipotential bonding (for electrical safety purposes), the impact of the trays is neutral (neither good nor bad).
• Only in the case of high quality (low RF impedance) connections between the tray parts, the shielding effect of a solid tray may be utilized in its full range.
• If a metallic tray system with good high frequency interconnects between parts is maintained, there is no need for any other ground structure, i.e. the system itself becomes a shielding equipotentializing system. A wooden or plastic external weather protection may then be used.
• The use of trays as an enforced electromagnetic protection of electronic systems is a promising technique. However, the knowledge within the installation industry is very sparse. Moreover, the present installation standards provide little help for understanding the importance of the quality of the interconnections between ground structure parts in a system. The implementation of good engineering practice is therefore still a challenge for the future.

Acknowledgment

The authors would like to thank Wibe Corporation for the funding of the project and fruitful technical discussions on the electromagnetic quality of installations.

 

 

Lennart Hasselgren, Ulf Nilsson
EMC Services Corp.
Mölndal, Sweden

 

 

Do not miss the related article ”Backgound and Comments on the EMC Performance of Cable Trays.”

 

References

[1] Directive 2004/108/EC of the European Parliament and of the Council, of 15 December 2004, on the approximation of the Laws of Member States relating to electromagnetic compatibility.

[2] Tim Williams, Keith Armstrong, EMC for systems and installations, 2000, Newnes, ISBN 0750641673.

[3] Electromagnetic environment handbook – EMMA, ISBN M7773-001271, Swedish Defence Material Administration, Sweden, Part 3 2005.

[4] EN 50174-2 “Information technology – Cabling installation – Part 2: Installation planning and practices inside buildings”, CENELEC, issue 2.

[5] “EMC performance of cable trays – shielding tests”, Report RE- 10273-17181, EMC Services, Sweden, 2008-03-31, (is available from Wibe Corp. Mora, Sweden)

[6] EN 55025 “Radio disturbance characteristics for the protection of receivers used on board vehicles, boats, and on devices – Limits and methods of measurement”, CENELEC, Issue 2.

[7] EN 55011 “Industrial, scientific and medical (ISM) radiofrequency equipment – Radio disturbance characteristics – Limits and methods of measurement”, CENELEC, Issue 4