View AP2421_305704.PDF datasheet online --- IC-ON-LINE

Datasheet File OCR Text:

Microcontrollers ApNote AP2421
: Additional file APXXXX01.EXE available
EMC by Simulation
Reduction of Electromagnetic Radiation by Detection and Optimization of Critical Networks on a PCB during the Layout-process Oliver Winkler / Siemens HL MC AT and ZT PP 5
Semiconductor Group
04.97, Rel 01
EMC by Simulation
1 2 2.1 2.2 2.3 3 3.1 3.2 3.3 3.4 4 4.1 4.2 5 5.1 5.2 5.3 5.4 6 6.1 6.2 6.3 7 8
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Electromagnetic Compliance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Signal Integrity, Emission and Susceptibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The earlier - the better . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Power of EMC-Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Theory of Electromagnetic Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The reason of radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Near Field and Far Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Differential and Common Mode Radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Modelling for simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 4 4 5 7 7 8 8 9
IBIS - Input/Output-Buffer-Information-Specification . . . . . . . . . . . . . . . . . . . . . . . . 10 IBIS characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 Usage of IBIS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 Simulating the Electromagnetic Scanning of PCB . . . . . . . . . . . . . . . . . . . . . . . . . . Simulation input . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Results of the Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Simulation of a PCB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Usage of EMC-Simulation-tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Measure to reduce emission . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Improvement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Simulation Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 12 12 13 16 17 17 17 19
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 Attachment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
AP2421 ApNote - Revision History Actual Revision : Rel. 01 Page of Page of actual Rev. prev.Rel. Previous Revision : none (Original Version) Subjects (changes since last release)
Semiconductor Group
2 of 21
AP2421 04.97
EMC by Simulation
1
Introduction
Unwanted electromagnetic radiation from electronic devices has become a serious problem worldwide raising the spectre of legal liability to say nothing of threat human life. All electronic devices emit electromagnetic radiation when operating. Since today's devices are more sensitive to electromagnetic fields, their normal operation can be disrupted by emissions from adjacent equipment. To get under control this unwanted electromagnetic effects the attitude towards how to make devices electromagnetic compliant has to be changed, first. It has to be changed from costly "damage" control to EMC measures that have already been integrated in the PCB-design process [1]. To realize this there are different methods, of course. In the following it is shown some experience made by EMC-simulation tools.
Semiconductor Group
3 of 21
AP2421 04.97
EMC by Simulation
2 2.1
Electromagnetic Compliance Signal Integrity, Emission and Susceptibility
EMC consists of three major disciplines (also see Figure 1): Signal integrity, emission and susceptibility. Signal integrity contains all problems of crosstalk, reflection, delay and all its effects on rise and fall times. Signal integrity effects on the stable functionality of a electronic device. If any
Susceptibility
Signal Integrity
Emission
Fig. 1: Three Disciplines of EMC
of these effects mentioned above occur, the device would not work correctly. Signal integrity is relevant for high speed boards working at a clock frequency starting at 50 . . . 100 MHz. Emission means all kinds of electromagnetic fields produced by an electric device. Emission is a problem mostly caused by digital devices. At least by electromagnetic susceptibility a device is influenced by electromagnetic fields of another device. But it is mostly the analog part of a device which is influenced by susceptibility. 2.2 The earlier -- the better
Electromagnetic compliance (EMC) must not simply be added to a device in a post-PCB-design process. This is due to the fact that the most post-design measures do not fight the reasons but the results of electromagnetic incompliance. As Figure 2 shows the later measures to improve EMC are made the smaller the choice of measures becomes and -- what is most important -- the more expensive the measures and thus the product itself becomes. In other words, EMC has to be designed with the PCB or even the whole device. Since plenty of EMC-simulation-software was developed in recent years, now, some became tools that help a designer of an electronic device to make his product electromagnetic compliant. By Semiconductor Group 4 of 21 AP2421 04.97
EMC by Simulation
Design
Test
Production
Costs
EMC-Facilities Development Process
Fig. 2: EMC-Facilities and their Costs as a Function of the Development Process simulation of critical parts of a PCB-design the designer gets spontaneous feedback of his attempt to increase EMC by routing a net in a certain way. So, the optimization of a PCB-design can be done completely in the layout phase of a PCB-design -- without building up a physical prototype but a virtual prototype. Verification and optimization within a virtual prototype shortens iteration loops and reduces time for measurements in a test-chamber (Figure 3). To summarize, an EMC-simulation-tool can definitely not tell if a device will or will not pass the test standard given by law. An EMC-simulation-tool will help to design a device as good as possible; it will help to improve the quality (EMC) of the device; and it will help to save time and costs. 2.3 The Power of EMC-Simulation
To be honest with EMC-Simulation tools, there are many things which can not be realized by simulation. But the further development continues more and more features become implemented. Figure 4 shows an overview of the simulation tools' productive usage for each discipline. It is easy to see that tools for signal integrity are already integrated in the design process. There are two kinds of tools for the computation of emission. Tools for calculating emission of a PCB are in the second half of 1996 being implemented successfully in several design environments. Although there are already tools for calculating emission of whole devices (i.e. PCB, enclosure and cables) available it will last at least until May/June 1997 to guarantee their efficient use. Of course, the use of a tool for PCB emission will be helpful and almost a soft requirement to reach this goal. And to have a future look, tool which allow the simulation of electromagnetic susceptibility will be available and useable in 1998.
Semiconductor Group
5 of 21
AP2421 04.97
EMC by Simulation
Without EMC Simulation:
AAAAAAAAA Layout AAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAA Prototyping AAAA AAAA AAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAA Testing AAAAAAAA AAAA A AAAAAAAAAAAAAAAAAAAAAAAAA A EMC Measurement AAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAA Redesign A AAAAAAAA AAAAAAAA AAAAAAAAA Prototyping AAAAAAAAA AAAAAAAAAAAAAAAAA Testing A AAAAAAAAA A AAAAAAAAAAAAAAAAAAAAAAAA EMC Measurement AAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAA AAAAAAAAA AAAAAAAAA AAAAAAAAA A A x times A A A A A A A A A A A A Design Entry AAAAAAAAA A AAAAAAAAA A Layout AAAAAAAAAAAA A AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA A A AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA EMC Simulation AAAAAAAAAAAAAAAAAAAA AAAA AA AAAAAA A AAAA Redesign A AAAAAA A AAAAAA AAAAA Prototyping AAAA A AAAAAAAAA A AAAAAAAA Testing AAAAAAAAA A AAAAAAAA A AAAAAAAA AAAAAAAAAAAAAAAAAEMC Measurement A AAAAAAAAAAAAAAAAA A AAAAAAAAAAAAAAAAAAAAAAAA A AA A x times AA A AAAAAAAA AAAAAAAA AA A AA A AA A AA A AA A Time Saving AA A AA A AA A
Design Entry
With EMC Simulation:
Fig. 3: PCB Development Process with and without EMC-Simulation
Si
gn
I al
nt
r eg
ity Em io iss n
of
Bs C P Em io iss n
of
D
e vic e
s lity bi ti
time
Su
e sc
p
1994
7/1996
5/1997*
1998* * at least
Fig. 4: Roadmap for integration of EMC-simulation tools into product design process
Semiconductor Group
6 of 21
AP2421 04.97
EMC by Simulation
3 3.1
Theory of Electromagnetic Simulation The reason of radiation
Having a closer look at a certain net on a PCB this net can be divided into several trace segments which are connected by several nodes. Figure 5 shows that each net can be modelled by a source,
Driver Trace
Ground, Power
itrace
Receiver
i back
Fig. 5: Loop antenna structures
some traces and a sink. Remember that in most cases the traces are located right over a ground plane. Current now flows from the source to the sink. Since both source and sink are connected with the ground plane there are currents within the ground plane backwards, from the sink to the source. It is well-known that these currents flow right beyond the traces. So it is easy to see that the current behaves like current in a loop antenna structure. And so these loop antenna structures have the same effects like a magnetic loop antenna: They cause magnetic radiation.
Semiconductor Group
7 of 21
AP2421 04.97
EMC by Simulation
3.2
Near Field and Far Field
Figure 6 shows the difference between the near field and the far field. The far field gives information
Sources
Near Field
Far Field
Fig. 6: Difference between the Near Field and the Far Field
about the total energy radiated by all antenna structures and regards all sources as one point source. Actually most measurements done in a test chamber are measurements of the far field. Although the final measurement is of the type far field measurement, during the layout phase it is necessary to get some knowledge about the near field. This is due to the major goal to detect and optimise critical nets on a PCB. The near field contains information about the several sources which created the near field. By inspecting the near field of a PCB an area right on the PCB can be determined containing the source of the radiation. By having this detailed information about the intensity and the location of radiation's source it is obvious which net has to be optimised in order to bring radiation down. 3.3 Differential and Common Mode Radiation
There is also a distinction whether the radiation is differential mode or common mode. Differential mode radiation comes from the nets i.e. from the currents in the nets (so-called differential mode currents) on the PCB themselves. Common mode radiation results from parasite effects like induced currents (so-called common mode currents) in mass areas, power areas and cables. Although the common mode currents are not as high as the differential mode currents most of a device's radiation is common mode radiation. Especially for far field analysis common mode radiation has to be taken into account.
Semiconductor Group
8 of 21
AP2421 04.97
EMC by Simulation
3.4
Modelling for simulation
As already explained all nets going to be simulated are divided into several trace segments. Each trace has got at least one source and one sink. Whereas the trace segments' geometric (length, width and height) and material data (mostly copper) can be derived from the layout tool the information about current sources and sinks have to come from an internal library containing driver and receiver models. These library models consist of voltage-current-curves (v-i-curves) which characterize the pulldown or pullup behaviour of the driver/receiver. One way to obtain these models and their curves is by converting an IBIS-model (Input/Output-Buffer-InformationSpecification) provided by the manufacturer of the component.
Semiconductor Group
9 of 21
AP2421 04.97
EMC by Simulation
4
IBIS -- Input/Output-Buffer-Information-Specification
Since component models e.g. spice models may show any information about technology details many component manufacturers refused to publish any simulation models. But simulation becomes more and more important, so a neutral standard has been created, called IBIS (Input/Output-BufferInformation-Specification). 4.1 IBIS characteristics
The IBIS standard is neither fixed on any simulator type nor it is limited on any component manufacturer nor it is fixed on any hardware platform. Figure 7 shows how one IBIS-ASCII-file replaces plenty of simulator specific libraries. The lower part of Figure 9 shows that the IBIS model
New: Using IBIS Models
Old: Separate Library Models Simulation Library Files
IBIS File
Incases Quad Quantic etc. Hyperlynx etc.
Incases
Quantic
Quad
Hyperlynx etc.
Fig. 7: IBIS provides plenty of simulators
has to be provided by the components' manufacturer. It can also be seen, that the IBIS file is accessed by the simulator via a converter provided by the software manufacturer. The IBIS model is of the type behavioural and it contains model name, model type (e.g. input, output, i/o, tri-state, open-sink, open-source) and capacities. Further, there are data like v-i-curves for pullup and pulldown, curves for clamping diodes, package information and pin assignments. So, it is only a description of the components input and the output buffers, no information at all about its inside. For more information contact http://www.eia.org/eig/ibis/ibis.htm
Semiconductor Group
10 of 21
AP2421 04.97
EMC by Simulation
4.2
Useage of IBIS
An IBIS model can be used for signal integrity analysis like crosstalk, delay and reflection and even for EMC-simulation of PCB. However, effects like Vss-Bouncing and Vcc-Bouncing and even simultaneous switching noise can not be simulated by the IBIS model (Figure 8). These are effects
IBIS Model - Static I/O Behaviour
Reflection
Crosstalk
Radiation PCB
Vss Bounce
Vcc Bounce
Simultaneous Switching Noise
Fig. 8: Use of IBIS-models
of the component itself that are not regarded in a simulation process, yet. But the more exact a simulation should become the more exact a simulation model has to be. Therefore, some additional information should be added to the IBIS-standard which would bring simulation and at least design and development of products one step further.
Semiconductor Group
11 of 21
AP2421 04.97
EMC by Simulation
5
Simulating the Electromagnetic Scanning of PCB
The following describes both the design flow and the experience made with the EmcScan Simulator from Quantic Laboratories. 5.1 Simulation input
Figure 9 shows the whole scenario of an EMC-simulation process. After the layout has been generated fully or partially its data will be passed to the simulator. They contain geometric data of the nets, data about the components and also data about the board's layers and its materials. The extraction can be done within the layout tool's framework.
PCB-design system Schematic Layout Framework integration Results Developer Weighting / Pulse Train EMC simulation tool EmcScan Redesign measures
Component manufacturer IBIS model Converter
Fig. 9: Overview of an EMC-simulation scenario
Another important input is the weighting of the nets. This weighting symbolizing the traffic on the nets defines a certain percentage of pulses for each net. It is obviously just a worst case estimation but including the library driver models (generated out of IBIS-models) exact computation becomes possible. 5.2 Results of the Simulation
The simulation computes several results. They can be accessed in the following manners:
* * * * *
Amplitude of the magnetic field vs. frequency Amplitude of currents on the traces vs. frequency Colour map of the PCB showing the magnetic field in certain areas for a selected frequency Colour map of the PCB showing the current-density on the traces for a selected frequency Quality factor to compare several designs (near future development)
Semiconductor Group
12 of 21
AP2421 04.97
EMC by Simulation
5.3
Simulation of a PCB
The simulation consists of the four steps shown in Figure 10. In the first step, the transmissionlinesimulation, all selected nets are divided in several trace segments. For each trace segment current waveforms and voltage waveforms (Figure 11) are calculated regarding both crosstalk and
Transmission Line Analysis
Fast Fourier Transformation
EmcScan
Near Field Calculation Method of Moments
Discussion of Results
Fig. 10: Flowchart of the EmcScan simulation
reflection. Based on these results the Fast-Fourier-Transformation (FFT) processes the spectral analysis of each trace segment's current. Figure 12 shows the FFT result for one trace segment: the amplitude of current as a function of discrete frequency steps. Using these results and the theory of loop antenna structures mentioned above the radiated magnetic field is computed by the method of moments in step three. Again, each trace segment is taken into account. Figure 13 shows the result of the field calculation 5 mm above the PCB. Due to the method of modelling loop antenna structures producing magnetic radiation, the field shown in Figure 13 is of the type magnetic field. Kept in mind to do a worst case simulation this result is a summary of the highest field amplitudes. It gives an overview what frequencies may cause EMC-problems. Once such a potential problem has marked out, the locations of the radiating nets can be found by exploring a colour map (Figure 14) of the selected frequency. Red areas show locations of high radiation. A real hardware scanning of the PCB would produce a similar colour plot. However, the big advantage of a software scanning is that it takes place within the PCB-design process. Since the currents on the traces are responsible for the radiation the critical nets for this certain frequency can be fixed easily by exploring a second map (Figure 15) showing the traces coloured by the current density.
Semiconductor Group
13 of 21
AP2421 04.97
EMC by Simulation
Fig. 11: Voltage (left) and current (right) waveforms
Fig. 12: Result of the Fast-Fourier-Transformation FFT
Semiconductor Group
14 of 21
AP2421 04.97
EMC by Simulation
Fig. 13: Totally radiated Near Field of a PCB
Fig. 14: Colour Map at a certain frequency. Red dots show areas with high magnetic radiation
Semiconductor Group
15 of 21
AP2421 04.97
EMC by Simulation
Fig. 15: Currents on the traces. Red traces show high current density
5.4
Measures
As already described in the previous, radiation directly depends on the current intensity. Thus, to reduce radiation, it is necessary to reduce current density in critical traces or nets.
To achieve this following measures can be taken: adding components (e.g. a serial resistor) rerouting the trace / net changing the layer changing the placement of components on board. Since the simulation takes place after or -- what would be much better -- during the layout phase the above mentioned measures are easy to implement.
* * * *
Semiconductor Group
16 of 21
AP2421 04.97
EMC by Simulation
6
Usage of EMC-Simulation-tools
All results of the previous chapter were taken from an industrial product design process. Figure 13 shows a high magnetic emission at a frequency of 16 MHz. The so-called "hot spot" in the upper left of Figure 14 shows the source of the emission. In detail, the trace producing the radiation can be pointed out in Figure 15. 6.1 Measure to reduce emission
By extracting the critical network the reason of the high current density and the high radiation in the trace segment between (1) and (2) becomes evident. In Figure 16 the driver of the net is located at
(3a)
(2) (1) (4)
(3b)
Fig. 16: Critical Network that has to be optimised
(1). In (2) the net splits to the receiver pins (3a) and (3b) what causes a high current flow and therefore a high radiation of the trace segment. A serial resistor (4) placed as near as possible to the driver pin (1) reduces the current in the trace segment. Of course, a resistor may cause signal integrity problems in the within the net. By trying out several types of resistors the best value can be found. Figure 17 shows the voltage waveforms with a resistor of 0 (no resistor), 80 (best solution) and 200 . 6.2 Improvement
Figure 18 shows how a 80 resistor reduces the radiation of the net. On the left side there is the amplitude of radiation vs. frequency of the net without any resistor. On the right side there is the amplitude of radiation vs. frequency of the net with an 80 resistor. Radiation is about 10 dBA lower. There is also an obvious decrease of radiation in a higher frequency range up to 25 dBA.
Semiconductor Group
17 of 21
AP2421 04.97
EMC by Simulation
Fig. 17: Voltage Waveforms for Several Resistor Values -- 0 , 80 , 200
Fig. 18: Emission of the Original (left, 0 ) and Redesigned (right, 80 ) Trace Further, a look at the far field of the trace would show a more evident difference between the original and the redesigned trace, since higher frequencies become more critical while they turn from the near to the far field. Semiconductor Group 18 of 21 AP2421 04.97
EMC by Simulation
6.3
Simulation Time
The time a simulation-run lasts depends on the system that is used. On e.g. a Sun Ultra 2 a simulation of 16 nets takes about one hour of time. Actually, it can be said that the simulation of a whole complex PCB can be done in an overnight run. The simulation of one net, during the optimising process, can be done within five minutes.
Semiconductor Group
19 of 21
AP2421 04.97
EMC by Simulation
7
Conclusion
It has to be pointed out that simulation will never replace a final measurement since all laws and standards require measurement for verification. However, by integrating simulation in the development process and by creating and optimising not a physical but a virtual prototype positive effects on the whole product can be achieved. In a short term the number of measurements decreases. Simultaneously the PCB-designers' EMCknow-how increases. In a mid-term simulation results in a improvement of quality and image, it shortens the development process, reduces the number of measurements and required physical prototypes. It also gives a feedback if a measure taken has been successful. So the product can be designed in a most efficient way. In a few words, simulation saves time and money. It is quite an investment but it is more an innovation.
Semiconductor Group
20 of 21
AP2421 04.97
EMC by Simulation
8 8.1
Attachment List of EMC-Software-Manufacturers
The following list names several manufacturers of EMC-Simulation software to compute emission of a electronic device. Not all of them have the full functionality described in the previous.
* * * *
HyperLynx, Inc, Redmond, USA http://www.hyperlynx.com Incases Engineering GmbH, Paderborn, Germany http://www.pad.incases.com Quad Design Technology, Camarillo,USA http://www.quaddesign.com Quantic Laboratories, Inc., Winnipeg, Canada http://www.quantic-emc.com http://www.megasolution.de References John Lavery, PhD, Electromagnetic Services Inc. Quantic Lab., EmcScan Manual Tim Williams: EMC for Product Designers Newnes: Oxford, Boston, Johannesburg, 1996 Anton Kohling: CE-Konformitatskennzeichnung Publicis MCD Verlag / VDE Verlag GmbH Assistance Otto-Hahn-Ring 6 81739 Munich Germany
8.2
[1] [2] [3]
8.3
Siemens AG ZT PP 5
Peter Kaiser
Phone: +49 (0)89 / 6 36 - 4 78 89 E-mail: Peter.M.Kaiser@mchp.siemens.de Markus Klamet Phone: +49 (0)89 / 6 36 - 4 32 70 E-mail: Markus.Klamet@mchp.siemens.de
Semiconductor Group
21 of 21
AP2421 04.97

▲Up To Search▲

Price & Availability of AP2421

	To Download AP2421 Datasheet File
If you can't view the Datasheet, Please click here to try to view without PDF Reader .