EMI EMC PT 4

Download as ppt, pdf, or txt
Download as ppt, pdf, or txt
You are on page 1of 62
At a glance
Powered by AI
The key takeaways are that EMC design considerations need to be made at the primary, secondary and tertiary levels which include PCB design, cables/connectors, and shielding/gaskets respectively. Factors like component selection, trace impedance, layer stackup, crosstalk, power distribution and decoupling need to be addressed during PCB design for EMC compliance.

The three levels at which EMC control measures can be incorporated are the primary level (PCB, circuit design etc.), secondary level (cables, connectors etc.), and tertiary level (shielding).

Digital circuits are generators of EMI since they use high frequency square waves rich in harmonics that get distributed throughout the system, while analog circuits are much quieter as high frequency square waves are not normally present. The harmonic content of digital signals decreases at specific rates with increasing frequency.

EMC Design of PCBs

Contents

Introduction
Component selection and mounting
PCB trace impedance
PCB layer stackup
Crosstalk control
Power distribution
Decoupling

Contents

Zoning
Grounding
VIAs connection
Termination

Introduction

Introduction

Introduction
EMC Control Measures
can be incorporated at
three levels.
Primary level (PCB, Circuit
etc)
Secondary level (Cables,
connector etc)
Tertiary level (Shielding)

Introduction

Introduction
Primary Level
Circuit design measures, Board layout
grounding

Secondary Level
Cables & Connectors

Tertiary Level
Apertures ,Gaskets etc

Component Selection & Mounting


Digital circuits
Digital circuits are generator of EMI,high
frequency square-waves are rich in harmonics
are distributed throughout the system.

Analog circuits
Analog circuits are much quieter because high
frequency square waves are not normally the
feature.

Component Selection & Mounting


Fourier transform
Basic to understanding of why switching circuit cause
interference is the concept of time domain /frequency
domain transform
Switching waveform can be represented as
trapezoidal with defined rise and fall time .The
harmonic amplitude content of a trapezoid decreases
from the fundamental at a rate of 20dB per decade
until a breakpoint is reached at 1/ptr, after which it
decreases at 40dB/decade

Component Selection & Mounting

Component Selection & Mounting


It is advisable to use the slowest logic
family that will do the job; dont use fast
logic when it is unnecessary. Treat with
caution any proposal to substitute devices
from a faster logic family, such as
replacing 74HC parts with 74AC.

Component Selection & Mounting


Harmonics contents of different logic families

Component Selection & Mounting


Some IC manufacturers are addressing the problem of RF
emissions at the chip level.
By careful attention to the internal switching regime of VLSI
devices, noise currents appearing at the pins can be
minimized.
The transition times can be optimized rather than minimized
for a given application .
Revised package design and smaller packages can allow
the decoupling capacitor to be placed as close as possible
to the chip, without the internal lead frames inductance
negating its effect;

Component Selection & Mounting


From the point of
view of immunity, a
slow logic family will
respond less readily
to fast transient
interference

Component Selection & Mounting


The high frequency spectral content of the
component signal are gauged by the signal
speed :

where f0 is the fundamental frequency (inverse


of repetition rate) of its signal, I0 is the
magnitude of its drive current, and tr is its
rise/falltime

Component Selection & Mounting


In a digital system we know that the key
component of a signal that dictates its highfrequency spectral content is its rise/falltime. In
fact we can say that a digital signal has a
bandwidth that is essentially the reciprocal of
its risetime, BW 1/tr . Hence a clock having a
500 ps rise/ falltime has significant spectral
content up to 2 GHz.

Component Selection & Mounting


In the initial layout of the PCB, a spreadsheet
should be compiled of every part to be placed on
the PCB according to its speed.
This priority list should be given to the layout
personnel with instructions to apply the greatest
care to placement of the highest-speed
components at the top of the list and to place the
components first according to the highest speed
and next to the next lowest speed, and so on.

Component Selection & Mounting


Numerous automated layout software provide for
autorouting so that the schematic is laid out and parts
connected in a rather automatic manner. This can and
usually does create severe EMC problems unless it is
used judiciously.
It is best to place the highest-speed components
manually and to manually route their connection lands
with the previous concepts kept in mind

Component Selection & Mounting


As a general principle, recall that the higher the spectral
content of a signal, the greater the ability of that signal to
couple inadvertently to other conductors and parts.

Hence the highest speed components and the lands they


connect should be positioned on the PCB well away
from off board connectors and lands as shown below.

Component Selection & Mounting

Component Selection & Mounting

Anticipating that the rise/ falltimes of this clock may need to be


increased to reduce its high-frequency content, pads should be
placed so that
(a) a series resistor can be inserted in the lands and
(b) a capacitor can be inserted across the lands forming a lowpass
filter.

Component Selection & Mounting

In the beginning, the capacitor pads would not be populated, and a


0-V surface-mount resistor could be placed in series with the clock
lands.
Later in the design, if problems are uncovered necessitating
increasing the clock rise/falltimes, then the capacitor can be added
and the value of the series resistor can be changed.
Why is this significant? The answer is that no time-consuming
repeated PCB layout would be required; only the bill of materials
(parts list) need be changed. This is a virtually cost-free solution.

Component Selection & Mounting


Once again, it is a good practice to place all the
components, or at least the higher-speed ones, by hand
and not allow automatic software to do this.
Perhaps in the future EMC layout rules will be
incorporated into these layout tools, but today it seems
better to place these components by hand.

Component Selection & Mounting

Illustration of the unintentional coupling of signals between chip


bonding wires, causing supposedly quiet module pins to have highfrequency spectral content. A processor pin, the reset pin, was
thought to be quiet and hence was connected to a long PCB land.
Unsuspected coupling caused it to carry the clock signal, thereby
creating large radiated emissions. A series inductor was added to
prevent this from occurring.

PCB Trace Impedance

Track impedance is dominated by


inductance at frequencies higher than a
few kHz

PCB Trace Impedance


Where L is the inductance and di/dt is the rate of
change of current. For example, consider the
case where the power-supply wiring has an
inductance of 50 nH. If the transient current,
when a logic gate switches, is 50 mA and
thegate switches in 1 ns, the noise voltage
generated across the power-supply wiring when
this one gate changes state will be 2.5 V

PCB Trace Impedance


Trace inductance of a connection can be reduced in two
ways:
minimizing the length of the conductor, and if possible increasing
its width
running its return path parallel and close to it.

The inductance of a PCB track is primarily a function of


its length, and only secondarily a function of its width.
For a single conductor of diameter d and length l inches,
following equation gives the self-inductance

PCB Trace Impedance

PCB Trace Impedance

Two conductors with current in opposite directions (e.g., a signal and


its ground return trace) have a total loop inductance Lt, equal to

where L1 and L2 are the partial self-inductances of the individual


conductors and M is the partial mutual-inductance between them. I
To minimize the total loop inductance, the partial mutualinductance between the conductors should be maximized.
Therefore, the two conductors should be placed as close together
as possible

PCB Trace Impedance


Another method for decreasing the inductance of a circuit
is to provide alternative paths for current flow. These
paths must be electrically, but not necessarily physically,
parallel. If two equal inductances are paralleled, then the
equivalent inductance will be one half that of one inductor.
On many low-cost products, inner plane boards cannot be
afforded. Hence the important effects of having a ground
plane (and a power plane that can act like a ground plane
in returning currents) are not present

PCB Trace Impedance


On low-cost products where ground plane cannot be
afforded, a simple way of reducing the inductance of
grounding system is with a gridded ground system.

PCB Trace Impedance

The limiting case of a gridded ground is when an infinite number of


parallel paths are provided and the ground conductor is continuous,
and it is then known as a ground plane
The main EMC purpose of a ground plane is to provide a lowimpedance ground and power return path to minimize induced ground
noise

PCB layer stackup


PCB layer stackup (the ordering of the layers and the
layer spacing) is an important factor in determining the
EMC performance of a product
The following four factors are important with respect to
board stackup:

The number of layers


The number and types of planes (power and/or ground)
The ordering or sequence of the layers
The spacing between the layers

PCB layer stackup


In deciding on the number of layers, the following should
be considered:
The number of signals to be routed and PCB cost.
Clock frequency.
Will the product have to meet Class A or Class B emission
requirements?
Will the PCB be in a shielded or unshielded enclosure?
The EMC engineering expertise of the design team.

PCB layer stackup


One- and Two-Layer Boards:
The main EMC concern here is to keep loop areas as small as
possible.
One- or two-layer boards should only be considered when clock
frequencies are less than 10 MHz
On one- or two-layer boards all critical signals should be routed
first, in order to guarantee optimum routing. Critical signals should
be routed as short as possible, with an adjacent ground return
trace
On two-layer digital boards, the ground and power should be
routed so as toform a grid (see Section 10.5.3). It is not unusual to
see a 10- to 12-dB decrease in emissions when a ground grid is
added to a two-layer digital board that previously did not have one

PCB layer stackup


Multilayer Boards:
Multilayer boards (four or more layers) using ground and/or power
planes provide significant reduction in radiated emissions over
two-layer
A oftenused rule of thumb is that a four-layer board will produce
20 dB or more, less radiation than a two-layer boardall other
factors being equal
Boards containing planes are much better than those without
planes for the following reasons:
The planes allow signals to be routed in a microstrip (or stripline)
configuration. These configurations are controlled impedance
transmission lines that produce much less radiation than the
random traces used on one- and two-layer boards

PCB layer stackup


Multilayer Boards:
When the return current is on the adjacent plane, the loop area is
reduced.
The ground plane significantly decreases the ground impedance
and hence the ground noise

PCB layer stackup


Multilayer Boards Objective :
A signal layer should always be adjacent to a plane.
Signal layers should be tightly coupled (close) to their adjacent
planes.
Power and ground planes should be closely coupled together.*
High-speed signals should be routed on buried layers located
between planes. The planes can then act as shields and contain
the radiation from the high-speed traces.
Multiple-ground planes are very advantageous, because they will
lower the ground (reference plane) impedance of the board and
reduce the common-mode radiation.
When critical signals are routed on more than one layer, they
should be confined to two layers adjacent to the same plane. As
discussed, this objective has usually been ignored.

PCB layer stackup


Four layer Board :

A common four-layer board configuration, consisting of two signal layers


and two planes, is shown in Fig.
It consists of four uniformly spaced layers with internal power and ground
planes. The two external trace layers usually have orthogonal routed

PCB layer stackup


Four layer Board :

If the power plane shown in Figs. 16-12 or 16-13 is split to


accommodate different dc voltages, it is important to restrict routing
on the bottom signal layer so that traces do not cross the split in the
plane.
If some traces must cross the split, stitching capacitors should be
located close to where the traces cross the split to provide a lower
impedance return current path.

PCB layer stackup


Six layer Board :

Most six-layer boards consist of four signal routing layers and two
planes.
From an EMC perspective, a six-layer board is preferred over a fourlayer board because it is easy to shield high-frequency signals by
placing them on buried layers between planes, or to provide for
orthogonally routed signal layers that are referenced to the same
plane.

PCB layer stackup


Six layer Board :

PCB layer stackup


Six layer Board :

PCB layer stackup


Eight layer Board :
Most eight-layer board stackups are used to improve
EMC performance rather than add additional routing
layers.
Most eight-layer boards (and all the ones that we will
concentrate on here) consist of four signal-routing layers
and four planes.

PCB layer stackup


Eight layer Board :

Crosstalk control
Crosstalk is one of the several important PCB design that must be
considered during any design cycle.
Crosstalk refers to the unintended electromagnetic coupling between
traces, wires to traces, cable assemblies.
Crosstalk between cables, wires and traces affects intrasystem
performance.
Crosstalk is generally considered to be functionality concern (signal
quality)

Crosstalk control
To cope with this problem, it is essential to have
suitable circuit models to simulate complicated
structures.
Simplified but accurate models are of great help
in understanding the crosstalk mechanism.
Lumped model of two coupled lines that can be
easily implemented in SPICE-like circuit
simulators. By virtue of its simplicity, this is the
first approach for inexperienced users when
dealing with crosstalk and using a commercial
circuit simulator that usually does not offer
efficient coupled line models in its library.

Crosstalk control

Crosstalk control

Crosstalk control

Crosstalk control

Crosstalk control
Near end and Far end Crosstalk
Near-End Crosstalk (NEXT), defined as the
voltage VNE at point D of the victim line
near to the source;
Far-End Crosstalk (FEXT), defined as the
voltage VFE at point E of the victim line far
from the source.

Crosstalk control

Since the overall coupling effect is


the sum of capacitive and
inductive crosstalk, near-end
crosstalk is of greater magnitude,
and hence, of greater concern for
a PCB designer.

Crosstalk control
Techniques to Reduce Crosstalk
1. Use wide traces.
2. Locate traces near to the reference plane.
3. Increase the spacing between traces.
4. Match the lines.
5. Reduce the coupling length.
6. Provide other traces for shielding.
7. Avoid locating critical traces near to the edge of the PCB.
8. Ensure a solid ground plane without cuts.

Power Distribution & Decoupling


The characteristics of an ideal direct current (dc) power distribution
system are as follows:
To supply a constant dc voltage to the load
Not to propagate any ac noise generated by the load
To have a 0 O ac impedance between power and ground

Power Distribution & Decoupling


There are two primary purposes for using decoupling between power
and ground-reference planes.
1.Functionality:

Power Distribution & Decoupling


Role of decoupling capacitor:

Power Distribution & Decoupling

Effective power supply decoupling has become increasingly more


difficult to achieve as the result of increasing clock frequencies and
faster rise times. Ineffective decoupling can lead to excessive power
bus noise as well as to excessive radiated emission.
It is important to understand that decoupling is not the process of
placing a capacitor adjacent to an IC to supply the transient switching
current rather it is the process of placing an LC network adjacent to
an IC to supply the transient switching current .

Power Distribution & Decoupling


2. To reduce the noise injected into the power and ground reference
plane

Power Distribution & Decoupling

All decoupling capacitors have inductance in series with them.


Therefore, the decoupling network is a series resonant circuit. The

inductance comes from three sources, as follows:


The capacitor itself
The interconnecting PCB traces and vias
The lead frame inside the IC

Power Distribution & Decoupling

The internal inductance of a surface mount technology (SMT)


capacitor itself is typically 12 nH, the interconnecting PCB traces
and vias add 5 to 20 nH or more according to the layout, and the
internal lead frame of the IC may have 3 to 15 nH of inductance
according to the type of IC package. The inductances of the
interconnecting PCB traces, however, are usually the only
parameters under the system designers control.
From the above, we observe that the total inductance can vary from a
low of about 10 nH to a high of 40 nH. Typically it is in the range of 15
to 30 nH. It is this inductance that limits the effectiveness of the
decoupling network. It is very important to remember this factwe
are placing an LC network between the power and ground, not a
capacitor

You might also like