Practical Papers, Articles and Application Notes


In this issue you will find one practical paper that should interest members of the EMC community. It is entitled “Electromagnetic Emissions of Integrated Circuits and PCBs” by Jerzy F. Kolodziejski and Juliusz Szczesny. In this paper, the authors discuss the sources of electromagnetic emissions from integrated circuits and PCBs that appear as conducted and radiated electromagnetic disturbances over a broad frequency range. This paper was originally presented as, “Use of power supply currents for electromagnetic compatibility investigation of digital integrated circuits” at the XVI International Wroclaw Symposium and Exhibition on Electromagnetic Compatibility, June 25-28, 2002 and has been revised and reproduced here by permission. In addition to this paper, I wholeheartedly recommend that you read the paper by Keen, Tesche and Butler on the Clemson EMC course that appears in the Education Section of this Newsletter. Had it not appeared there, I would have been pleased to have it printed in this section. Our profession would benefit tremendously if all of our EMC courses were of as high quality as this one.
The purpose of this section is to disseminate practical information to the EMC community. In some cases the material is entirely original. In others, the material is not new but has been made either more understandable or accessible to the community. In others, the material has been previously presented at a conference but has been deemed especially worthy of wider dissemination. Readers wishing to share such information with colleagues in the EMC community are encouraged to submit papers or application notes for this section of the Newsletter. See page 3 for my e-mail, FAX and real mail address. While all material will be reviewed prior to acceptance, the criteria are different from those of Transactions papers. Specifically, while it is not necessary that the paper be archival, it is necessary that the paper be useful and of interest to readers of the Newsletter.
Comments from readers concerning these papers are welcome, either as a letter (or e-mail) to the Associate Editor or directly to the authors.


Electromagnetic Emissions of Integrated Circuits and PCBs

There are at least a few reasons for the growing interest of electronic components manufacturers and users in the electromagnetic compatibility (EMC) of these components. Among them of great importance are: increase of packaging density of electronic circuits, lowering the levels of useful signals in respect to the background fluctuations (noise) level, and the rise of speed and functional complexity – above all in integrated circuits (ICs) and particularly digital ICs.
Talking about the EMC properties, one should take into account the two sides of the problem – electromagnetic emission (EME) and electromagnetic immunity (or susceptibility) of the considered objects. The first question will be discussed shortly in this paper. Illustrative examples will relate to digital integrated circuits and to the printed circuit boards (PCBs) with ICs. Both conducted and radiated emissions will be considered.

Electromagnetic Emission of Printed Circuit Boards
Electric circuits on PCBs, which can produce EME, are composed of ICs, cooperating active and passive components, connecting traces, power, control and signal lines as well as I/O ports and attached cables. The EME is caused by functional activity of ICs and other active components and by the flow of time varying electric currents.

Fig. 1. Differential mode electromagnetic disturbances generated by current loops inside the PCB circuits

Electromagnetic disturbances generated by current loops inside the circuits on PCBs, Fig. 1, have the characteristics of differential-mode signals in the systems with forward and return conductors. The current loop can be considered as a small frame antenna that creates in its close vicinity an electromagnetic field with a dominant magnetic component. The dimensions of the small loop antenna should be smaller than a quarter wavelength of the radiated signal (e.g. < 100 cm at 75 MHz and <10 cm at 750 MHz). If the loop area is around several square centimetres and the loop current is a few tens of milliamps, then at the frequencies higher than 30 MHz, the EME levels could exceed the limits in European Standard EN 55022 for class B equipment (for residential area). In this situation, it would be strongly recommended to lower the frequency of useful signals, reduce current values and limit the area of current loops. If these means could be impractical, e.g. for the reasons of other requirements, the application of shielding should be considered. Common-mode disturbances are caused by unwanted voltage drops, appearing in the systems with cables and conductors used as the forward path of the signals, and with the reference (ground) plane, which serves as the return path.

Fig. 2 Common mode disturbances generated in the systems with cables and conductors, which are the forward path of the signals, the ground becomes the returning path

 

The voltage drops may be related to the voltage difference between the reference plane (local ground) and module or equipment ground, Fig. 2. The attached cables and conductors, working as the rod antennas, mostly radiate high frequency common-mode disturbances. It is usually assumed that the common-mode impedance of cables or longer traces on the PCB is approximately 150 W. Frequencies of emitting disturbances may depend on the cable resonance, appearing above few tens of MHz. It is always difficult to foresee the paths of common-mode currents within the circuit, since some weakly controlled parasitic elements, mainly capacitances, often participate in the current flow to the ground. Significantly lower values of common-mode currents, in comparison to differential-mode currents, can produce EME that exceeds the limits given in EN 55022.

Integrated Circuits as the Source of Electromagnetic Emissions
Complex and fast ICs such as microprocessors, microcontrollers, digital signal processors and various ASICs, often have large transients in the supply currents. It is then justified to treat the question of EME of the integrated circuits in the same manner as in the case of PCBs. Emission levels are higher for larger currents and faster changes of currents and voltages. They also depend on the circuit design and technology, ground and supply systems, pin assignments, packaging types, applied shielding, etc.

Fig. 3. Current loops around the conductors inside IC, which become the source of the common mode (asymmetrical) radiation


Current and voltage transients related to high frequency switching and short rise/fall times of useful signals cause EME of digital ICs. Transients from digital logic spread out as the conductive signals that may couple onto the power, control and signal lines and through the IC pins reach the interconnected components on the PCB. The simultaneous switching of multiple outputs of complex IC can produce the fluctuations of current and voltage on the power bus, known under the name of simultaneous switching noise – SSN.
Current loops around the conductors inside IC, although smaller than the loops on a PCB, cause the differential (symmetrical) mode radiation, passing from chip to out-of-package. Unwanted voltage drops in the circuit and the presence of attached interconnecting conductors acting as the antennas of various effectiveness, are the source of common-mode (asymmetrical) radiation. As in the case of the PCB, this type of emission can be related to the fluctuations of voltage difference between the PCB reference plane (PCB ground) and the reference potential USS and supply voltage UDD inside IC. These fluctuations are called ground and supply bounce, Fig. 3.
The energy carried by emitted electromagnetic waves may be sufficiently high to induce interfering voltages on the connecting conductors and cables located around functioning IC. The observed power levels radiated by some microcontrollers in the frequency band of 1-1000 MHz were in the range of –110 dBm (0.01 pW) to – 75 dBm (about 32 pW). It is rather obvious that the levels of radiated power can be different for different types of circuits. It is less obvious that the difference between the circuits of the same type but coming from various manufacturers or production lines may be as high as a few tens of decibels. The highest emission levels are appearing in the vicinity of quartz, clock generator and output buffers of the circuit. This is the reason for a practical recommendation not to install in the close proximity of these networks other electronic components that are particularly susceptible.
Electromagnetic emissions are usually tested in the broad frequency range from a few kHz up to some GHz. Regarding the complexity of the emission mechanism, different models and techniques are adopted as the base of measurement methods.

Table 1. Basic Features of EME Standard Test Methods


Measurement Methods of EME
The International Electrotechnical Commission (IEC) proposed as the standards five - and practically four - measurement methods of emission of ICs (elaborated by Working Group 9 of Technical Subcommittee SC 47A). These are the following methods: TEM cell, surface scan (magnetic loop), 1 W/150 W (resistive probes), workbench Faraday cage (WBFC) and magnetic probe. Short characteristics of the methods are given in Table 1.
These methods together allow evaluation of ICs emission levels in the frequency range up to 1 GHz, comparison of various types of ICs, and guide the manufacturer’s efforts to design new circuits with reduced emission generation. Simplified illustrations of typical measurement arrangements are given in Fig. 4.

 

Fig. 4. Simplified illustrations of typical measurement arrangements: a) TEM or GTEM cell, b) loop probe for measuring near magnetic fields, c) resistive probe, d) Faraday Cage, e) magnetic probe


The radiation level can be measured inside a TEM cell (Fig. 4a), or a GTEM cell for higher frequencies. The IC is mounted on its test board that is clamped to a mating port (wall port) cut in the top or bottom of a TEM cell. The RF voltage appearing at the input of the connected spectrum analyser is related to the electromagnetic potential of the IC and its application (test) circuit.
Electric and magnetic components of near electromagnetic field (roughly at the distance from radiation source less than 1/6 of the radiated wavelength) can be measured with the help of magnetic and electric probes (Fig. 4b). The probes are mechanically scanned over the surface of an IC or other elements mounted on a PCB. The shift of probes can be done by hand, with the use of a microscope stage or fully automatically under the control of a computer program. Automatic equipment provides a two-dimensional picture of the RF current circulating within the tested PCB. Finally, it can be easily converted into a coloured map showing the field strength distribution.
RF emission from small electronic modules or ICs occurs mainly via the attached conductors and cables. In two of the measurement methods of conducted emissions (resistive probes, Fig. 4c and magnetic probe, Fig. 4e), RF current inside the circuit is measured to allow indirect estimation of the emission level. The variable RF component of current on the supply lines is dominant and reflects the whole IC activity in respect to the generation of the electromagnetic disturbances. The contribution of some I/O ports also should be taken into account. Thus, spectral characteristics of the supply current IDD (or ISS) of the IC as well as the selected parameters of its waveform, defined in the time domain, can be easily correlated to the emission level. For IC pins that are intended to be connected directly to long (> 10 cm) PCB traces or cables, RF voltage measurement using a 150 W coupling network is recommended in IEC 61967-4.
Measurement of common-mode conducted disturbances, produced by an electronic module or by an IC mounted on its test board, can be performed inside the workbench Faraday cage (Fig. 4d). This is a metallic box of 500x300x150 mm, equipped with adequate connectors, filters and matching elements. RF voltage at the selected IC port is measured across the used common-mode impedance.

Fig. 5. The distributions of EM fields over the package of 8-bit Intel 80C31 type of microcontroller: a) magnetic component, b) electric component


Illustrative Examples
Digital circuits usually generate the strongest electromagnetic disturbances. Many experimental results of emission measurements were obtained for 8-, 16- and 32-bit microcontrollers.
The distribution of magnetic and electric components of the EM field, measured over the package of 8-bit Intel 80C31 and 8-bit Dallas Semiconductor DS 87C520 types of microcontrollers are, respectively, given in Fig. 5 and 6. A DS 87C520 device is pin and instruction-set compatible to the popular 8051 circuit and can run at clock rates up to 33 MHz (over 2 times faster than 80C31). Both circuits were mounted on test boards and worked in the close programming loop that allowed activation of all inherent blocks. Measurements were performed at the established set of points over the packages with the help of E and H miniature probes (made by Langer EM-Technik, Germany). Signals from the probes were amplified in the broadband amplifier and sent to the input of a LeCroy 9370 type oscilloscope. Results showed represent the average peak-to-peak values of the voltages taken from over 100 measurements. As shown in Figures 5b and 6b, the distribution of the E component is similar in both cases. The highest strength appears in the vicinity of ICs I/O ports (3rd column, 1-3 rows). As shown in Figure 5a, the H component for the 80C31 circuit has its maximum roughly in the same place as the E component. As shown in Figure 6a for DS 87C520, the H component distribution is more even, but on average slightly higher. The difference is seen in the vicinity of I/O port 3 and XTAL terminals (1st column, 9-12 rows).

Fig. 6. The distribution of EM fields over the package of 8-bit Dallas Semiconductor DS 87C520 type of microcontroller: a) magnetic component, b) electric component


Even more impressive are the results presented in Fig. 7. Changes in the design and technology introduced by the manufacturer (Hitachi) resulted in the substantial reduction (as much as 15 dB) of magnetic field of the H8S series of microcontrollers in comparison to the former H8 family. Both circuits are 16-bit microcontrollers with fclock = 16 MHz.

Fig. 7. The distribution of magnetic component of EM fields over the package of Hitachi 16-bit microcontrollers: a) H8/3048 type, b) H8S/2148 type


Examples of the H component distribution over an electronic module on a PCB, taken with the help of an automatic EMC-Scanner made by Detectus AB Sweden and an HP 11940A type magnetic probe, are given in Fig. 8. The four consecutive pictures were taken at four different peak frequencies: 96, 112, 212 and 631.9 MHz. The change of localization of the strongest field sources can be easily noticed; they are related to ICs at lower test frequencies and move to some output buffers and I/O ports and connectors for higher frequencies.
The above conclusion about the frequency range of the EME of ICs well correlate with the measurement results obtained, e.g. by application of resistive or magnetic probe methods. In Fig. 9 the spectral characteristic of an EME of an ST 6210 microcontroller is presented. The curve, being the envelope of individual spectral values of the emission, was measured with the help of a magnetic probe MP-10L (NEC Corp.) that was placed over the IDD supply line of the circuit (according to IEC 61967-6).

Fig. 8. Distributions of magnetic component of EM field over electronic module (400x300 mm) obtained on Detectus AB scanning system; 10 kHz bandwidth, peak frequency: a) 96 MHz, b) 112 MHz, c) 212 MHz, d) 631.9 MHz (Reproduced with the kind permission of Detectus AB, Malung, Sweden)

Fig. 9. Spectral characteristic of ISS current of ST 6210 type microcontroller


Power supply currents provide valuable information not only on the conducted emission levels of ICs, but also on their immunity to external electromagnetic disturbances. On the basis of measurement and analysis of currents, one can draw conclusions about the disturbance type and its intensity. Disturbing electromagnetic signals are revealed in the supply currents much earlier than in any changes in circuit output signals. Fig. 10 and Fig. 11 show, respectively, the output voltage and a part of the spectral characteristic of the supply current of an 80C31 type microcontroller, when a sinusoidal 10 MHz disturbing signal was coupled to the input terminal of the circuit. It is easy to note the appearance of a fundamental component of the disturbing signal V1 inside the spectral characteristic of the IC. Note that there is no evidence of the presence of a disturbing signal on the IC output voltage. (The visible small fluctuations are caused by the clock signal.) The situation would, of course, change if the disturbing signal levels were significantly higher and sufficient to cause the switching of the IC output.

Fig. 10. Output voltage of 80C31 type microcontoller - 10 MHz sinusoidal signal was coupled to the input port

Fig. 11. Selected part of the spectral characteristic of 80C31 type microcontroller at the conditions as above, V(I) - voltage at the output of magnetic probe)


Summary
Integrated circuits and various electronic components inside electronic modules mounted on PCBs are the source of electromagnetic emissions that appear as conducted and radiated electromagnetic disturbances over a broad frequency range. There is a need to measure the emission levels and quantitatively characterise emission sources. Among the elaborated and accessible measurement methods worth noticing are the methods related to the measurement of power supply currents, since they can reflect the inherent functional activity of the circuits. EMC

References
[1] IEC 61967 Integrated circuits. Measurement of electromagnetic emissions, 150 kHz to 1 GHz. Parts 1-6
[2] J.F.Kolodziejski, J. Szczesny: Electromagnetic disturbances produced by digital integrated circuits (in Polish), Elektronizacja nr 10, 2002, pp. 21-23


Jerzy F. Kolodziejski received his M Sc degree from the Wroclaw Technical University in 1959, his Ph D (E.Eng.) from the Warsaw Technical University in 1965, and his D Sc from the Institute of Electron Technology Warsaw in 1992. He has been employed at the Institute of Electron Technology since 1970 and as a Full Professor since 2001. For several years, he has been involved in the measurement methods of semiconductor devices and their reliability testing. His current interest is in EMC problems, mainly on the IC and PCB level. He is a nominated expert of the Polish Standardization Committee and currently is an active member of WG9 SC 47A of the IEC. He is an author or co-author of a many publications on EMC problems, including one book (in Polish).

 

 

Juliusz Szczesny received his M Sc degree in physics from the Nicolaus Copernicus University in Torun, in 1973, and his Ph D from the Institute of Electron Technology Warsaw in 1987. From 1974 to 1999, he worked on microwave measurement methods and the design of discrete semiconductor devices. Since 2000, he has been working on EMC compatibility problems, mainly on ICs and PCBs. He is a member of WG 2 SC 47E of the IEC as an expert in discrete semiconductor devices. He is an author or co-author of several publications on microwave semiconductor devices and EMC problems.


If you would like to contact the IEEE Webmaster
© Copyright 2004, IEEE. Terms & Conditions. Privacy & Security

return to contents
IEEE logo