EMC Design Technology

Spectrum utilization and potential interference

Figure 14 shows the frequency range commonly used in daily life, including AC power frequency, audio frequency, frequency bands occupied by long, medium and short wave radios, 900MHz and 1.8GHz commonly used in FM and TV broadcasting, and cellular phones. But the actual spectrum is much more crowded than this, and the frequency band above 9KHz is almost used for specific occasions. The frequencies shown in this figure will soon extend to 10GHz (or even 100GHz) as microwave technology is widely used in everyday life.
Figure 15 overlays on Figure 14 some of the less familiar spectrums emitted by common electrical and electronic equipment.
Pictures related to this topic are as follows:
insert image description here

Pictures related to this topic are as follows:
insert image description here

Figure 15. Spectrum after superimposing the interference we generate. AC power rectifier devices can emit switching noise from the fundamental frequency to fairly high harmonic frequencies, depending on the power of these devices. Power supplies (linear or switching mode) around 5 kVA typically fail to meet conducted emission limits below a few MHz due to switching noise from their 50 or 60 Hz bridge rectification. The noise generated by the thyristor DC motor drive device and the AC phase-shift control system is similar. These noises can easily interfere with medium- and long-wave and some short-wave broadcasts.
The operating fundamental frequency of the switching power supply is generally between 2kHz and 500kHz. It is common for switching power supplies to still have strong emissions at frequencies 1000 times their operating frequency. Figure 15 shows the emission spectrum of a switching power supply with a frequency of 70 kHz commonly used in personal computers. This will interfere with broadcast communications, including FM broadcasts. A typical emission spectrum generated by a 16MHz clocked microprocessor or microcontroller is also shown in Figure 15. Emissions from these devices typically exceed emission limits at frequencies of 200MHz or higher. Currently, as personal computers use clock frequencies of 400MHz or even above 1GHz, digital technology is bound to interfere with the high-end spectrum.
All of the above phenomena occur because all conductors are antennas. They convert the transmitted electrical energy into an electromagnetic field, which then leaks out into the wider environment. At the same time, they can also convert the electromagnetic fields around them into conductive electrical signals. This is a universal truth. Therefore, conductors are the main cause of radiated emissions from signals and the cause of contamination of signals by external fields (sensitivity and immunity).

2.2 Conductor leakage and antenna effect

The electric field (E) is created by the voltage on the conductor and the magnetic field (M) is created by the current flowing in the loop. Various electrical signals on conductors can generate magnetic and electric fields, so all conductors can leak electrical signals on them to the external environment, and also introduce external fields into the signal.
At wavelengths (λ) much greater than 1/6 of the frequency of interest, the electric and magnetic fields combine into a complete electromagnetic field (plane wave) containing both the electric and magnetic fields. For example: For 30MHz, the turning point of the plane wave is 1.5m; for 300MHz, the turning point of the plane wave is 150m; for 900MHz, the turning point of the plane wave is 50m. So as frequency increases, it is not enough to think of conductors as transmitters and receivers of electric or magnetic fields, as shown in Figure 16.
Pictures related to this topic are as follows:
insert image description here

Another effect of increasing frequency is that resonance occurs when the wavelength (λ) is compared to the length of the conductor. At this time, the signal signal can be almost 100% converted into an electromagnetic field (or vice versa). For example, a standard dipole antenna is just a piece of wire, but when its length is 1/4 the wavelength of the signal, it is an excellent converter for converting a signal into a field. While this is a simple fact, it is important for technicians working with cables and connectors to realize that all conductors are resonant antennas. Obviously, we want them to be very inefficient antennas. If we assume that the conductor is a dipole antenna (good for our purposes), we can use Figure 17 to help our analysis.
Pictures related to this topic are as follows:
insert image description here

Figure 17 Cable Length and Antenna Efficiency
The vertical axis of Figure 17 represents the conductor length (unit: m). For the convenience of observation, the spectrum of Figure 15 is reproduced. The diagonal line on the far right gives the length of the conductor as a function of frequency when the conductor is an ideal antenna.
Clearly, even very short conductors can create emission and immunity problems in commonly used frequency bands. It can be seen that at 100MHz, a 1-meter-long conductor is a very effective antenna, and at 1GHz, a 100mm conductor is a very good antenna. This simple fact is the main reason why EMC is called "Black Art".
In previous years, the frequencies that were widely used in everyday life were low, and typical cables were not very effective antennas, which is why the wiring "conventions" tended to become obsolete.
In Figure 17, the slanted line in the middle indicates the length of the conductor that may cause problems even though the conductor does not become an efficient antenna. The diagonal line on the left indicates the case where the length of the conductor is extremely short, and its antenna effect is negligible (except for particularly strict products). How many times have you heard someone say, "No problem, I'm grounded"? It's a common joke among EMC industry folks that RF is colorblind. So the yellow/green wire that transmits RF signals (yellow/green for safety grounds in US standards) cannot be imagined as a good ground, and all conductors used for grounding are also antennas.

2.3 All cables are affected by their inherent resistance, capacitance and inductance

Ignoring the role of fields and antennas for the time being, let's look at a few simple examples below. These examples illustrate that, in the usual frequency range, small deviations from the ideal can also cause problems with the signal carried on the conductor.

  • A wire with a diameter of 1mm has more than 50 times the resistance at 160MHz of the DC state, which is a result of the skin effect, forcing 67%
    of the current to flow within the outermost 5 microns thickness of the conductor at this frequency .
  • A wire with a length of 25 mm and a diameter of 1 mm has a parasitic capacitance of about 1 pF. This may sound trivial, but at 176MHz presents a loading effect of about 1kO. If this 25 mm long wire is driven in free space by an ideal square wave signal with a peak-to-peak voltage of 5V and a frequency of 16MHz, then at the eleventh harmonic of 16MHz, driving the wire alone would require 0.45mA of current.
  • The pins in the connector are about 10mm long and 1mm in diameter, and this conductor has a self-inductance of about 10nH. This also sounds trivial, but when a 16MHz square wave signal is passed through it to the motherboard bus, with a drive current of 40mA, the voltage drop across the connector pins is around 40mV, enough to cause serious signal integrity and / or EMC issues.
  • A 1-meter-long wire has an inductance of around 1µH, which when used in a building's grounding network can prevent a surge protection device from working properly.
  • The self-inductance of the 100 mm long ground wire of the filter can reach 100nH, which will cause the filter to fail when the frequency exceeds 5MHz.
  • A 4-meter shielded cable with shield terminated in a 25mm "pigtail" will render the cable shield ineffective at frequencies above 30MHz.
    Empirical data: For wires under 2 mm in diameter, the parasitic capacitance and inductance are: 1pF/inch and 1 nH/mm (sorry for no uniform units, but this is easier to remember). Its simple arithmetic relationship is as follows: The
    related pictures of this topic are as follows:
    insert image description here

2.4 Avoid using conductors

The above analysis shows that: as the frequency increases, the problems of the cable become more and more. It is increasingly difficult to use it to transmit the signal intact and prevent it from leaking.
Cables are starting to present more and more problems even for low frequency signals such as audio. Since all semiconductor devices have crystal detector characteristics in the frequency band up to several hundred MHz (even low speed op amps like the LM324), cable antenna effects can unwittingly contaminate the audio signal.
Therefore, from the perspective of the most economical means of meeting EMC requirements, it is best to avoid metallic cables and connectors altogether. Non-metallic wires can be used for communication, and many similar products are already available, including:

  • Optical fiber (more suitable for non-metallic wires)

  • Wireless communication (eg: Bluetooth; local area network)

  • Infrared (eg: IrDA)

  • Free Space Microwave and Laser Communications (eg, between two buildings)
    2.4.1 Cost/Benefit Analysis of Nonconductive Products
    Many designers believe that cost savings can only be achieved by using traditional cables and wires. But when considering the cost of a complete project, the reliability and electromagnetic compatibility of the product or system, installation and many other factors, it is often found that the total cost of fiber optic or wireless communication is lower. Of course, it was too late by then.
    For signal cables and connectors, except for the simplest electronic products, the price of raw materials is not necessarily related to the sales price. A proper cost/benefit analysis is necessary for signal integrity, EMC compatibility, danger of overcharging, risk of high return rates, quality complaints, unsalable products, etc.
    Design engineers are reluctant to consider the business risks of the products they design, but they are the only ones who decide whether the product is competitive (usually the requirements are made by marketers). However, if electronic engineers blindly consider only the functional parameters of the product and the price of raw materials, then their company will lose its competitive advantage and will also be exposed to unpredictable business risks.
    **

  • Mr. Bai Jilong has been engaged in the electronics industry for 15 years. He has developed more than 100 products so far, and most of them have been mass-produced.

  • It took 5 years since 2018 to record thousands of practical-level electronic engineer series courses, from components to core modules to complete products

  • Lao Bai's original intention is "May the world's engineers not take detours" Among them, there are courses on MOS tubes and IGBTs in detail**

Guess you like

Origin blog.csdn.net/m0_49011926/article/details/124200217