Grounding
Grounding
Grounding
Safety, system protection and performance are the three main reasons to earth a
system. Not all electronic equipment needs to be connected to earth to work,
satellites are an example. Sometimes wrong grounding configurations, oriented to
satisfy the special power and performance requirements of electronic loading
equipment, can compromise safety rules generating dangerous situations for
personnel and equipment. Personnel safety, equipment safety and performance
grounding issues have to be analyzed together. In any case, safety rules must not be
violated. All these concepts as well as the relations between ground issues need to
be clarified before passing on to a higher level in the control of the electromagnetic
emissions of the system. Along the present chapter, concepts like safety grounding,
signal or ground reference plane as well as grounding topologies to achieve electric
safety and good performance of the system, are presented. The characteristics of
shield structures, shielded power cables as well as their connections to ground are
also presented. Finally in the last section, some recommendations to design the
CMS detector grounding are included. The grounding design of the detector
constitutes the first of several steps to be followed for the detector integration under
a electromagnetic compatibility plan.
47
48
49
enclosure and racks. All these items are ultimately grounded together at the
grounding electrode of the power service or a separately derived system.
Equipment grounding is required for both personnel safety and power systems
protection. From a personnel safety point of view, properly grounded system
components minimize potential differences that may exist between various system
components under transient and fault conditions. From a system protection
standpoint, properly grounded system components provide a low impedance path
for ground fault currents and promote the timely operation of over-current
protective devices in case of ground faults.
There are two different grounding configurations associated with the connections of
the equipment to the PE point; the standard ground configuration and the insulated
ground configuration.
For the standard equipment grounding configuration, a supplementary ground
conductor provides an additional low impedance ground path in parallel with
the metallic conduit or raceway from the equipment to the power system.
For the isolated grounded configuration, the electronic or electrical load is not
connected to the metallic enclosure and this metallic enclosure is earthed
through an independent conductor to the start point of the safety ground. The
insulated equipment grounding conductors run with the other circuit
conductors feeding electronic load equipment. In an isolated ground
configuration, the additional equipment grounding conductor provides the sole
grounding path from the electronic or electrical load equipment to the power
system or separated derived system.
In either case, the insulated equipment grounding conductor should run in the
same raceway or conduit as the phase and neutral conductors. Grounding
configurations provide equalization of potentials between grounded objects at the
operative frequency. As the frequency increases, other grounding means must be
considered to cover the high frequency range.
Robust design of the electronic load equipment for immunity to disturbances on the
grounding circuit is a good method to get a good grounding [15]. Particularly for
distributed computing and telecommunication electronic loads, using optical
signaling interfaces reduces the sensitivity to disturbances on the ground circuit. As
it is implemented in the CMS detector, the processed data from the detector is
transmitted out of the detector via optical devices, and the slow control signals are
isolated from the detector via opto couplers.
2.2.2.1 Standard equipment grounding
The standard equipment ground configuration [13][14][16] uses one equipment
grounding conductor (PE or G), in green color, running with the phase and neutral
51
metal
raceway
and
conduit.
This
52
receptacle (IGR) only to the equipment grounding terminal or bus of the power
system source or separately derived system. This equipment grounding conductor
extends radially downstream to the chassis of the electronic load equipment
without contacting any grounded metal surfaces such as metal conduits and race
ways, panel boards, and outlet boxes for receptacles. This configuration is shown in
figure 2.3.
53
The application of the isolated ground configuration may provide beneficial effects
to circuits supplying electronic load equipment that do not otherwise connect to
grounded objects. However, if the electronic equipment contains other connections
to grounded objects as in the standard ground configuration, the performance of the
isolated grounding configuration decreases. These connections to ground may be
either intentional or unintentional. Typical examples of these connections are
interconnections of various equipment through grounded shields of data cables and
bonding of equipment chassis to grounded metal equipment racks.
54
The need of an SRS is minimal when all of the inter unit signal and
telecommunication circuits are interfaced to the associated electronic equipment via
optically or isolations transformer coupled means. It is used in CMS detector, where
these interfaces have good common mode voltage breakdown characteristics.
However, the need for an SRS may easily rise to that of a requirement in the event
any of the following three conditions are given:
When the logic AC-DC power supplies used in the associated electronic
equipment are installed with one of the terminals connected to the equipments
metal frame/enclosure. This is typical and recommended practice in the
equipment industries.
When the signal-level circuits and logic AC-DC power supply common
terminals are dielectrically insulated or galvanically isolated from the equipment
ground against recommended practice, and are instead connected to an
insulated ground terminal that is intended for connection to an externally
installed signal ground reference circuit.
When there are actual performance problems occurring with the equipment,
which can be assigned to common mode electrical noise or similar common
mode interference related to the equipments existing grounding system,
whatever its design, or the signal-level
Improved HF grounding for data signaling cables between (non-contiguous) areas
can typically be accomplished by reducing the open loop area enclosed by the cable
and its grounded surroundings. This is typically accomplished via the use of metal
conduit or electrically continuous, solid-bottom, metal cable tray, wire-way, or
similar forms of signal transport ground-plane constructions.
The main advantages of the SRS are:
Low-impedance return path for RF noise currents
Containment of EM (noise) fields between their source (cable, etc.,) and the
plane
Increased filtering effectiveness of contained EM fields
Shielding of adjacent equipment
An SRS may be typically constructed using one the following methods (in
decreasing order of effectiveness), which are well described in [16]:
Solid covering of sheet metal (RP).
Grid of cooper straps (GS).
Grid of cooper and aluminium wire (GS)
Rised flooring substructure (GS).
Hybrid forms of SRS employing mixtures of signal reference grid and signal
reference planes for varied construction and improved overall performance are also
55
useful. They are used where the benefits of each type of SRS are needed for the
collective support of a variety of interconnected electronic load equipment that is
susceptible to CM noise current.
The use of RP may be recommended for some applications where the subject
system operates at a higher frequency than the typical GS signal design cut-off
frequency. Although the RP does not offer a zero impedance path across its surface,
it does offer several orders of magnitude of improvement over any signal or ground
of large-cross section wires that might be used for bonding between units. The
general equation for determining the impedance between two points on a plane is:
2d
Z gp = ( R DC + j Z RF ) 1 + tan -----------------
(2.1)
2
impedance of the ground plane in m . and d the distance between two points.
Accordingly, the impedance between two points on a ground plane should be
approximately equal to the ohms per square value of the material of which it is
composed and of the length such that the distance d is short compared with wave
length of the highest frequency of interest in the design. Finally, the ground
plane must be at least as wide as d. Based on this concept, it may be seen that the
impedance value between points across dimensions d is constant at DC level and
gradually rises as the applied frequency of the current is risen. Finally, isotropic rise
in impedance may be expected at the first resonant point encountering 0.25 on the
plane and at each succeeding odd-order multiples 0.5 of the first resonant point
generally approaching the expected impedance along the projection of Z N . This is
shown in figure 2.4. The most favorable area of operation for the ground plane is in
the range where 1 < 1 20 at the highest frequency of concern.
56
Often the solid form of a ground plane cannot be obtained and a mesh or grid form
must be a substitute for it. GS may be thought of as a plane in which holes, whose
dimensions have been kept below a critical value, have been placed in a repeating
pattern. In general only grounding grids using squared cells are recommended for
use for HF currents. One of the most important issues related to grounding grids
operating in the HF area, as opposed to DC and AC power frequency areas, is the
need for minimizing the reactance in the path of the grounding grid instead of the
resistance.
57
58
59
resistance in the metallic path and for high frequency currents the impedance
increases due to skin effect and inductive effects.
Bonding path installed on the site and intended for use at or near the power main
frequency, such as for ground fault currents, need not consider much more than the
resistance of the material being used together with its heating and fusing
characteristics. Thus simple wire resistance and capacity tables are normally
adequate for the determination of the paths performance with regard to the sizing
and minimizing of ground fault voltage drop across the bonding path. However for
longer ground fault path lengths, wire impedance tables that account for the
slightly increased impedance due to skin effect and reactance at the power
fundamental frequency must be used. When unwanted AC signals are present in
the grounding / bonding path, things are somewhat different. In these cases the
need to minimize the potential developed across the bonding path under high
frequency (HF) voltage and current conditions of all kinds is of major importance.
This need entails the consideration of the path impedances and with the reactance
of the bonding path being of greatest importance.
In general the total impedance of a bonding conductor is given by:
Z = R DC AC + j L
(2.2)
The most important term of the above equation is the inductance L. The magnitude
of the inductance L is calculated based on the concept of partial inductance. The
partial inductance assumes that a current flowing through a segment of unitary
length defines a virtual loop, where the opposite segment is at infinity. Based on
figure 2.8, the partial inductance L pij is the ratio of the magnetic flux penetrating
the surface between segment i and the infinity and the current j that produces it.
60
L =
Lpij
(2.3)
j=1
where the loop contains a total of N segments in which the loop has been divided.
Each segment supports a current I j , and the sign of each term corresponds to the
relative orientation of the currents assigned to segments j. A detailed definition and
analysis about partial and total inductance can be found in [12][20][21][22] and [23].
Based in the definition above mentioned, it is possible to calculate the partial
inductance of a single round bounding.
4l
L H = 0.002 l ln --------- 0.75
d
(2.4)
where l is the length in centimeters and d is the diameter of the wire also in
centimeters. Basically the value of this inductance depends on the length and the
diameter of the bounding, increasing the diameter of round conductors has a
limited effect on the HF impedance presented by the strap. On the other hand, if a
rectangular bonding strap is considered as a flat-strap, the following equation
applies:
b+c
2l
L H = 0.002 l ln ------------ + 0.5 + 0.2235 ------------
b + c
l
(2.5)
where l is the length in centimeters, b is the width of the wire in centimeters and c is
the thickness of the strap. The length to width ratio of a strap may be significantly
altered to benefit a bonding paths performance, whereas altering the diameter of a
round wire cannot produce useful improvements. Basically, it is much better to do
the ground connections with flat-straps than with round conductors. It is also well
known that the inductance increases by winding a conductor into a coil or similar
form where the magnetic lines of force along the conductor may be concentrated
causing an interaction among them. Thus, it must be expected that the best
performance is achieved on a bonding path or with only a minimum and gentle
bends in the winding conductor. Loops and sharp angles must be avoided unless an
inductance is deliberately being created at the point of path discontinuity.
Bounding straps may self-resonate or may resonate with the stray or parasitic
capacitance between the equipment and ground. The bonding strap exhibits high
impedance near resonance and at resonance it looks much like an open circuit.
Hence operation on this area must be avoided. The variation of the impedance with
the frequency and the equivalent circuit of the strap are shown in figure 2.9.
61
(2.6)
Resistance (AC and DC) and inductance are affected in much the same way.
However in the case of the inductance, unless the mutual inductance is also kept to
a minimum between parallel straps, they will not act as independent inductances in
parallel, the mutual inductance in that case must be considered also. Spacing must
therefore be maintained between straps for at least the length of the longest strap
employed if this problem wants to be avoided. A detailed analysis of this concept is
shown in [16]. Grounding straps should be as short as possible to minimize
inductive reactance in the path. The use of at least two bonds widely spaced apart
on the same item of equipment is recommended to further reduce the reactance of
the grounding path. These straps should be of different length such that they will
have different self-resonant frequencies. The straps should never be folded or
coiled, nor bent into curves. Even in equipment, lineups where the equipment is
62
2.4. Shielding
bonded together, the recommended practice is to bond each enclosure to the signal
reference grid with its own strap or two if it becomes easier. Flat foil strips, which
are relatively wide in relation to the length, are the recommended practice.
Connections to the equipment frame or a supplied grounding terminal are critical.
Paint or other surface contact inhibitors should be removed before bonding straps
are directly attached to metal enclosures or cabinet surfaces. Subsequently, the
connections should be properly treated to inhibit rust, corrosion and moisture.
2.4 Shielding
2.4.1 Introduction
The objective of electromagnetic, electric and magnetic shielding is to provide a
significant reduction or elimination of incident fields that can affect sensitive
circuits as well as to prevent the emission of components of the system from
radiating outside the boundaries limited by the shield. The basic approach is to
interpose between the field source and the circuit a barrier of conducting or
magnetic material.
Shielding effectiveness can be defined as the reduction in magnetic, electric or
electromagnetic field magnitude caused by the shield. The effectiveness of a shield
depends on the shield material as well as the characteristics of the incident field (far
or near field), which is defined by the distance between the source and the victim.
So, it is found that techniques for shielding depend on the type of source; whether
the source is a magnetic field, electric field or electromagnetic field source. The
shielding effectiveness (S) in dB, can basically be calculated as the sum of three
components, namely, reflection loss (R), absorption loss (A) and a correction factor
(B) used in special cases to consider multiple reflections in the shield.
S = A+R+B
(2.7)
63
Each component has a different expression and value depending on whether the
incident wave is magnetic, electric or electromagnetic field. Although analyzed in
detail in [12][24][25][26] some issues are presented in this section. The presence of
holes and joints can decrease the effectiveness of a shield and its analysis is well
detailed in [12] and [25]. The present study does not provide further analysis on the
subject though it should always be taken into account during the design phase.
Other important point analyzed in this section is the shielding of cables. Special
attention is paid to the connection of the shield at either one or two ends of a cable
and the implications that these connections have in the performance of the system.
(2.8)
r f
R = 168 10 log -----------
(2.9)
where t is the thickness of the shield in m., f the frequency, r the relative
permeability and r the conductivity. This equation is in close relation to the skin
depth of the material. The reflection losses decrease with the frequency, whereas the
absorption losses increase due to the skin effect. Based on these equations, it is
possible to state that the reflection loss is the primary contributor to the shielding
effectiveness at low frequencies. However, at higher frequencies the absorption loss
is the primary contributor to the shielding effectiveness at high frequencies.
64
2.4. Shielding
-
R = 322 10 log ---------------------3 2
r f r
(2.10)
where r is the distance between the source and the shield. When the distance is
undetermined, the equation showed for the far-field reflection could be used
instead, as electric near-field reflection losses are always lower than, or equal to, the
far-field reflection losses, hence introducing a certain safety margin. Based on those
equations, conclusions for fields with dominant electric component are very similar
to those attained for electromagnetic fields. The reflection loss is predominant at
low frequencies, while absortion loss is predominant at higher frequencies.
r f r
(2.11)
Based on this equation, the reflection loss decreases for decreasing frequencies, and
is lower than the reflection loss for the plane wave reflection. So, reflection losses
65
are usually negligible for lower frequencies and absorption losses are small for low
frequencies too. This fact forces the use of different shielding techniques against low
frequency magnetic fields [27][28][29].
Basically, there exist two different ways for shielding against low frequency
magnetic fields [12][24].
Deviation of the magnetic flux with high permeability material.
The shorted tuned method, which consists in the generation of opposing fluxes
that cancel the magnetic field in the area of interest.
To deviate the magnetic flux, it is recommended the use of magnetic material
instead of conductor material because it increases the absorption losses, hence
improving the attenuation of the magnetic field (as it is the primary shielding
mechanism at low frequency against magnetic fields). However, when a magnetic
material is considered as material shield, two properties of this material, which
introduce some limitations, have to be taken account:
The permeability of a magnetic material decreases by increasing the frequency.
The permeability of a magnetic material decreases by increasing the magnetic
field strength.
The former depends only on the material, the latter depends on the material and the
section of the magnetic circuit. For example mu-metal material has a permeability of
over 10000 from DC up to 1 kHz, however at 20 kHz the permeability is not larger
than cold-rolled steel. The working conditions of the shield is an important point as
it has a serious impact in the selection of the shielding material.
As a summary a magnetic material such as steel or mu-metal makes a better
magnetic field shield at low frequencies than does a good conductor such as
aluminium or copper. However at high frequencies, good conductors provide better
magnetic shielding. For non-magnetic material the shield effectiveness increases
with the frequency, therefore, it is recommended to calculate the attenuation for the
lowest frequency of interest. On the other hand, for magnetic materials the shield
effectiveness may reduce due to the decrease of the permeability with the frequency.
66
2.4. Shielding
A shield is made of non magnetic materials, which as it has been shown in the
previous section, has very bad properties to attenuate magnetic fields, and specially
low frequency magnetic fields. A shield should never be considered as a good
shield against magnetic fields unless a current is allowed to circulate through it,
which can create an opposite field that tends to cancel the perturbing field. These
currents can only circulate if both sides of the shield are connected.
When a shield is not connected to ground or is connected only in one side, the
shield attenuates only the electric component of the incident field. In this
configuration, the shield can be considered as a capacitive voltage divider as it is
shown in figure 2.10. Connecting the shield to ground is required to define the
capacitive voltage divider. The shield side closest to the electronic system should be
grounded in this case.
shield current generates a magnetic field opposite to the perturbing magnetic field,
attenuating its effect on the central conductors of the shielded cable. This effect does
not have any adverse effects on the cable shield attenuation to electric field
components.
The attenuation to electric field and magnetic fields is quantified in shielded cables
by two characteristic parameters as the transfer admittance and the transfer
impedance, respectively. The transfer impedance of a shield is defined as the ratio
between the voltage per unit length generated by the circuit formed by the shield
and the conductors inside the shield and the shield current. The transfer impedance
gives the open-circuit voltage developed between the internal conductors and the
shield for one ampere of shield current, in a cable of 1 meter long [25]. On the other
hand, the transfer admittance of a shield is defined as the ratio between the current
per unit length in the conductors inside the shield and the voltage between the
shields and the external structure. The transfer admittance gives the short-circuit
current induced in the internal conductors (when the internal conductors are
shorted to shield) for one volt between the shield and the external structure, in a
cable of 1 meter long [25]. Both the surface transfer impedance and the surface
transfer admittance are characteristic parameters of shielded cables and are studied
more in detail in chapter 4.
Cable shields grounded at both ends can carry unwanted shield currents due to
potential difference between the two grounded ends of the shield. In general, these
currents will be related to the power systems fundamental and harmonic
frequencies. DC and low frequency currents in the shield, as described above, can
be eliminated or significantly attenuated by placing a blocking device between the
shield and its ground connection point at one end. For example, a series-connected,
back-to-back arranged stack of rectifier diodes or capacitors can be used. When high
frequency capacitors are used at one end of the shield, the capacitor blocks the low
frequency currents but allows high frequency current to flow through the shield to
increase the shield attenuation to magnetic fields at high frequency. Back-to-back
diodes or surge diodes can be used when the connection between the shield and the
ground point has to be established if the potential between these two structures is
higher than the voltage drop of the semiconductor devices [14].
The cables shield also offers a return path to common currents flowing through the
central conductors. The return path is particularly important in shielded power
cables to return the common mode currents generated at the output terminals of
switching power supplies. If a regular cable without shield is used to connect these
power supplies to the load, common mode current will establish between the
conductors and the signal reference system (SRS), existing a large radiation loop for
these currents. In order to be effective, shields must be grounded via
low-impedance paths at the frequencies of interest. Long shields need to be
grounded at multiple locations along their length [12][16][17]. To be effective,
shields should be grounded using especial connectors that bonds the total
68
periphery of the shield to the chassis or frame where the shield has to be
continued.Pig tail connection of the shield to the chassis reduces the attenuation of
the shield a high frequencies.
69
The main grounding structure of the detector is the own metallic structure of the
detector together with the concrete reinforcement structure at ground level. This is
the ground reference system of the detector (SRS). All safety and protective grounds
have to be connected to these structures. Both are shown in figure 1.3 in green color.
At CMS detector we can distinguish two different areas, where the grounding
design can follow different strategies. Those areas are the DC-area around the
detector itself and the 400Hz distribution area. The first area includes very low
voltage units, in the range of 2.5V to 10V with current capabilities up to 100A, and
high voltage units ranging between 600V and 10KV with very tinny current
capability. The 400Hz distribution area is composed by the three phase 400V
power distribution system with a possible installed apparent power of 500-750VA.
As it was described briefly in chapter 1, the DC-area is composed by all the
front-end electronics (FEE) of the sub-detectors. They can be arranged in millions of
channels where the sensitive detector is close integrated to the front-end amplifier
and signal processor or can be composed by amplifiers that process tinny signals
coming from chambers distributed in a vast area of the detector (muon system). In
both cases, the metallic structures, boxes or frames containing the electronics and
the detection devices have to be connected to the metallic structure of the detector
or SRS. The general practice is to use the metallic frame or box as shield or screen
for the sensitive electronics, forcing to connect internally the common point of the
electronics to the box or frame. Then, when the FEE is considered as a load, each
unit has one of the power terminals connected to the metallic box or frame, which is
different to the scheme represented in figures 2.2 and 2.3. Additionally, due to the
voltage levels and current capabilities associated to the equipment in this area, an
independent conductor for PE is not included and each FEE is directly grounded to
the SRS. This topology is similar to the multi-point grounding connection depicted
in figure 2.7, but the PE o green cable is not included.
Power supplies (PS) for the FEE system, in general, are located outside of the
detector, in the CMS case, they will be placed in two areas, one around 20-40 mts
away from the FEE and the other about 100-120 m. To avoid the return current from
the FEE uses as path the SRS structure of the detector, the outputs of the power
supplies are generally floating respect to their frame or case. This frame or case has
to be grounded locally to the SRS. This topology is possible in the case of low
voltage distributions due to the extreme low voltage involved (V<45V). For high
voltage power supplies, the output terminals are not completely floating as the low
voltage PS and usually resistors or back to back diodes are connected between the
common terminal and the PS case. This topology assumes that all the power
supplies include a transformer with screen that separates electrically the input
terminals from the output terminal, which is in general observed for all practical
applications.
70
Considering the system at the input terminals of the power supplies as individual
loads, each one is now in close agreement with the diagram depicted in figures. 2.2
and 2.3. The AC or DC distribution system at this side is perfectly defined by
standards [25] and no violations of safety rules are allowed. In the 400Hz
distribution area, the PEs run in parallel with the power cables and they are
connected together at the secondary star of power distribution transformer. This is
shown in figure 2.11, where the proposed grounding topology of the HCAL
sub-detector along the power supply is shown.
Conduit
Conduit
GND
Generator
PE
PE
Distribution
Transformer
GND - detector
Distribution
Panel
FEE
Conduit
Shield
PE
Distribution
Panel
AC\DC converter
GND - detector
GND - detector
71
through the connection between the FEE and the metallic case holding it. This point
is critical due to the front-end electronics in general uses the frame or case as
screening. From safety point of view, this connection is better at the input power
terminal to avoid that the fault current uses as path the FEE where the fault current
capability is probably not enough. From performance point of view, it is convenient
to connect the FEE to the screen as close as possible to the sensitive amplifier or
detector to reduce the parasitic currents coupled to the amplifier. A trade-off
between performance and safety occurs in the selection of the connection point
between the FEE and the metallic case, prevailing the design based on electrical
safety.
The brief description of grounding design based on electrical safety considers a low
frequency range, covering from DC to a few KHz. Grounding can affect the
performance of the equipment installed and it is necessary to evaluate its impact
considering the high frequency behavior of the system. As it was described in
chapter 1, data signals from the FEE installed into the detector are transmitted to the
counting room via optical links. Slow control signals are electrically isolated via
opto couplers. The signal transmission via optical links simplifies considerably the
grounding configuration and allows focusing the grounding issues of the detector
to the power distribution system. This configuration allows considering the only
mechanism of interference among FEE units installed into the detector is
conduction and radiation through the power cables.
The electronic load equipment installed on the detector is multipoint grounded to
the signal reference structure. Basically, reference terminal or common points of the
digital and analog electronics are connected at a convenient point at the board level
to optimize performance. Electronics, hardware and power parts are grounded
separately at module level and then connected together at a single point, which
represents the main ground per sub-detector. As it was pointed before these
connection is a trade-off between safety and performance.
The noise generated by the PS units can be galvanically coupled to the FEE. This
noise is more severe in switched mode power supplies due the high frequency
switching of semiconductor devices. This noise has to be filtered by EMI filters
included at the input and output of the power supplies. Additionally, EMI filters
can be included at the input terminals of the FEE. Although, CM currents can flow
through the power cables, the return path for these currents will be the SRS if no
shield is used in the cable, increasing the capability of the cables to radiate. Shields
reduce the radiation area for these currents and reduce the noise current flowing
through the ground of the detector.
The shield of power cables are connected to one end at low frequency (FEE side)
and to both ends at high frequency. This can be achieved grounding the shield to the
PS side through a capacitor. Cables shields connected at both side at high frequency
adds screening to the conductor from external electromagnetic fields, avoiding the
72
induced current to flow trough the sensitive parts of the front-end electronics. All
auxiliary equipment as cooling system, etc. should be bonded to the signal reference
system. The equipment should be isolated from the FEE to avoid antenna behavior.
Each piece of electronic equipment has to be connected to the SRS. The bonding
connections should be as short as possible with no sharp folds or bends. Flexible
and multiple straps are preferred instead of round conductors. The main goal of
these connections is to decrease as much as possible the ground connection
impedance as well as to increase the resonant frequency of the strap. As a result, the
performance of the FEE is not affected. As example of strap effects, the ground
connection between the read-out box (RBX) to a reference copper plane has been
studied. The HCAL-RBX is connected during the FEE immunity test via a strap as it
is shown in figure 2.12 and the inductance measured. Results are shown in table 2.1
Strap
RBX
Copper plane
mA Strap
mV Strap
At 5 MHz
3.384
22.4
0.209 H
At 5 MHz
1.298
8.40
0.205 H
At 10 MHz
0.987
12.2
0.165 H
At 10 MHz
1.196
13.90
0.187 H
Average
0.191 H
73
capacitances associated to the strap and to the RBX are estimated. These values
correspond to 380 pF and 275 pF. The problem associated to the strap resonance
frequency may be easily solved by decreasing the length of the straps and adding
more connections to send the resonance frequency beyond the frequency of interest
of the FEE. Also these straps or ground connections should be connected to
opposing corners of the equipment and to the nearest, but separate points, on the
signal reference grids. This is shown in figure 2.13.
Strap 1
Strap 2
Strap 3
2.6 Summary
Grounding is necessary for safety, protection and performance reasons. It is based
on standards and no violation of safety rules can be done to improve the
performance of the electronics. The metal parts of equipment, enclosures and racks,
which are susceptible to be energized by electrical currents must be grounded for
personnel safety reasons and equipment protection. No operational currents must
flow in this ground (structures and cable). The ground conductor or protective earth
must run in parallel with the phase and neutral conductors.
There are mainly two different ways of grounding the electronic equipment, they
are: the isolated ground and the standard ground. The main difference between
both grounding configurations is that in the isolated ground configuration, the
metallic parts holding the load are grounded independently from the rest of the
metallic parts of the system, whereas in the standard ground, the metallic box
holding the load is directly connected to the general structure of the ground system.
The standard ground configuration only needs a simple cable as ground conductor,
which is called protected earth. However, the other configuration needs one
74
2.6. Summary
additional cable, which must fulfill with the typical requirements of the ground
conductors, and is used to ground independently some areas of the system.
In cases when two or more components of an interconnected system are installed in
an area where there is a physical space between them and across their separation
data input/output cables and inter-unit power circuit cables (DC, AC or both) are
routed, there are indirect bounding problems ranging from DC to several tens of
MHz or higher that can compromise the system performance. To make compatible
the grounding with the system performance, electronic components are locally
grounded to ground planes or grids known as signal reference surfaces. The main
advantage of these structures is to provide a low impedance path for the RF noise
current and to contain electromagnetic fields between their source and the plane.
The connection layout to the equipment to the reference structures can be made in
different ways but the two most important ones are: the single point ground
connection and multi-point ground connection. The former use only one point
where the components of the system can be grounded. For large systems, this
connection is not recommended due to the large amount of cables needed, the large
magnitude of the impedances associated with the ground connection and the poor
CM rejection due to the parasitic components associated with the system. In the
multi-point ground connection each component is grounded at two points; one at
load level and other at the safety ground point. This configuration allows a much
better control of the noise of the system. Attention should be paid to the generation
of low frequency ground loops, which could deteriorate the system performance.
The connection of the components of the system to the ground is made by straps.
These straps must be as short as possible to minimize inductive reactance of the
path. The use of at least two bond straps widely separated on the same equipment
is suggested to reduce the reactance of grounding path. These straps should have
different lengths.
Shielding of different components of the detectors is necessary to attenuate the
electromagnetic, electric or magnetic field that can couple into the system causing
performance deterioration or malfunction. The nature of the incident field as well as
the frequency of interest is of primary importance to select the shielding method. As
an example, magnetic materials are use to attenuate magnetic fields, whereas
conductor materials are recommended to attenuate electric and electromagnetic
fields.
The selection of shielded cables as well as the connection of the shield to the
equipment is very important. The cabless shields are made of conducting
materials, so they only can be used to attenuate electric or electromagnetic fields.
The attenuation of magnetic fields only can be performed if a current flows through
the shield, implying the connection of the shield to ground at both ends.
CMS detector can be separated in two different parts, the AC-400 Hz distribution
area and the DC distribution area. For grounding purposes, the former is a standard
75
ground system used in power distribution in which the protective earth wire runs
in parallel with the system. The DC distribution area is characterized by equipment
powered by extra low voltage/high current and high voltage/very low current. The
ground system is a sub-class of the standard configuration where there is not a
protective earth wire running in parallel with the power distribution cables. The
detector structure is used as signal reference surface and the front-end equipment is
locally grounded to this structure using straps. In this area it is necessary to take
special care about the grounding because it is not standard topology. Ground
connections have to be established to provide, in this order, personnel safety,
equipment safety (prevent fire hazard) and optimal system performance
2.7 References
[1] F. Szonsco, "Earthing of High Energy Physics Detector systems", CERN internal
document - Pre-print, 2002.
[2] F.Szonsco, "Assessment of EMC Parameters of LHC Front End Electronics", Proc. 5th
Workshop on Electronics for LHC experiments, LEB 1999, Snowmass, Colorado,
USA, pp 20-24, September 1999.
[3] P. Van der Laan, M. Van Houten, A. Van Deursen, "A grounding philosophy", Proc.
IEEE Symposium on Electromagnetic compatibility, CD-ROM Symposia Records
1955-1995.
[4] "European Standards", 93/68/CEE - (73/23/CEE), Ed. 1993
[5]" CERN code C1-Electrical Code, Ed. 1990
[6] William T. Rhoades, "Congruence of low voltage power main transient designs", Proc.
IEEE National Symposium on Electromagnetic compatibility, Denver, USA, pp
285-293, May 1989
[7] Edward C. Cantwell, "Effective grounding- The key to personnel and proper
equipment operation", Proc. IEEE Symposium on Electromagnetic compatibility, pp
194-199, 1980.
[8] A. Chouvelon, W Weingarten, "Grounding as seen by TIS", TIS-GS/TM/98-01,
CERN-1998.
[9] Stanley A. Erickson, "Spacecraft electromagnetic environment prediction", Proc. IEEE
Symposium on Electromagnetic compatibility, pp 106-115, 1978.
[10] Jasper J. Goedbloed, "Electromagnetic compatibility", ISBN-0-13-249293-8, 1990
[11] Tim Williams, "EMC for product designers, Paraninfo", ISBN-0-12-55710-0, 1997.
76
2.7. References
77
78