Some people in the automation industry don't pay enough attention to encoders , let alone encoder signal cables. "Twisted pair shielding," "the thicker the better," "ground, ground, and ground again, ground any empty space" seem so simple, but these vague or even wrong understandings of cables are the root cause of the interference in the field. Often, it starts with a misunderstanding of a simple cable and its connection.
In fact, encoder signal cables are quite technically complex, incorporating fundamental physics and materials science knowledge. The more fundamental the concept, the closer it is to physical principles, requiring an understanding of these principles combined with laboratory experiments and field verification, rather than relying solely on product manuals. Heidenhain encoders have always had strict requirements for their signal cables. I cannot provide a definitive answer here, as I lack sufficient laboratory facilities; I can only deduce this from the principles of electromagnetic wave signals. This is based on a custom encoder cable I commissioned twelve years ago from a foreign cable manufacturer. This manufacturer already had cables, but based on my understanding of Heidenhain cables, I requested a custom design specifically for encoders, valuing their laboratory facilities to test and provide the parameters I required, ensuring standardized experimental parameters. We have used this cable for 12 years, in numerous projects of varying sizes, and many knowledgeable peers have used and approved it. This foreign company has also successfully transitioned into the industrial control field, producing robot cables.
The saying goes, "Know yourself and know your enemy, and you will never be defeated." In this article, we discuss how we need to understand what encoder signals are, the characteristics of cables, and where interference might originate, in order to analyze and develop countermeasures based on the complex interference environment on-site. Here, I've selected ten questions in the industry where there's a lack of clarity or even some misconceptions about encoder signal cables. I hope this will spark discussion and reference among those truly practicing in the field, and I welcome comments and debates.
1. What are encoder signals?
As I introduced in my previous article, there are many types of encoder signals. This article will only discuss the most commonly used incremental pulse signals and digital serial signals (such as SSI signals), with electronic switching frequencies below 800kHz. Other types of bus signal cables, single-cable technology cables, and industrial Ethernet cables will not be discussed in this article. (I haven't fully grasped them yet either.)
The encoder signal discussed here is a square wave. However, the conductive characteristics of cables are calculated and designed based on electromagnetic waves. A square wave is not a single-frequency electromagnetic wave. According to Fourier decomposition, a square wave is a superposition and combination of electromagnetic waves of many frequencies. The diagram below illustrates that a square wave is composed of at least N=19 electromagnetic waves of different frequencies. (N=1 represents only one frequency of electromagnetic wave).
Therefore, the signal transmitted on the encoder cable is a combination of electromagnetic waves ranging from lower to higher frequencies.
Frequency characteristics of electromagnetic waves:
Extremely low frequency ELF below 3kHz
Very Low Frequency (VLF3-30kHz)
Low frequency LF30-300KHZ
MF300-3MHz (Medium Frequency)
High frequency HF3-30MHz
VHF 30-300MHz (TV channels 1-12)
UHF 300-3GHz (TV channel 13 and above)
Ultra-high frequency SHF3G-30GHZ
Light is also an electromagnetic wave. Many characteristics of high-frequency electromagnetic waves are the same as those of light that we are familiar with.
When electromagnetic waves pass through the interface of different media, they undergo refraction, reflection, diffraction, scattering, and absorption. Electromagnetic waves propagate within a conductor medium both along the direction of the conductor and along its diameter. They are refracted at the outer surface of the conductor, radiating into another medium (similar to light encountering water; this could involve entering another conductor medium or spatial radiation), and reflected back into the conductor medium for continued propagation (similar to light reflecting off water). The repeated reflections of various wavelengths create a chaotic scattering effect (similar to fog). Impurities within the conductor absorb electromagnetic wave energy, generating heat and causing wave absorption (similar to the principle of an induction cooker or microwave oven). Electromagnetic waves propagate both within conductors and outwards from the conductor interface. At low frequencies, electromagnetic waves are primarily transmitted through tangible conductive bodies. The reason is that in low-frequency electrical oscillations, the mutual change between magnetism and electricity is relatively slow, and almost all of its energy returns to the original circuit without enough energy to be radiated outwards. When the frequency of electromagnetic waves is high, the proportion of outward radiation gradually increases. In high-frequency electromagnetic oscillations, the mutual change between magnetism and electricity is very fast, and it is impossible for all the energy to return to the original oscillation circuit. Therefore, electrical energy and magnetic energy propagate into space in the form of electromagnetic waves with the periodic changes of electric and magnetic fields. Energy can be transferred outwards without a medium, which is a kind of radiation.
When higher-frequency electromagnetic waves reach the interface of a conductor, part of them refract and radiate away from the conductor, while the other part is reflected back into the conductor as if it hits a mirror. This reflected high-frequency electromagnetic wave superimposes with the next wave moving outward, resulting in a concentrated movement on the conductor surface. Therefore, high-frequency electromagnetic waves exhibit both the "skin effect" (moving along the conductor surface) and the "radiation effect" (leaving the conductor surface). Higher-frequency electromagnetic waves are also more easily absorbed by impurities in the conductor and attenuate rapidly.
When the surface of a conductor carrying high-frequency electromagnetic waves is a sharp interface, the shape of the tip, even after reflection, may still point outwards, increasing the chances of multiple radiations. Waves moving along the surface are more likely to radiate outwards, while fewer waves are reflected back. This is the tip radiation effect of high-frequency electromagnetic waves. This is also the principle behind antennas for transmitting high-frequency radio waves.
An LC high-frequency oscillator circuit is an oscillator circuit that uses an inductor (L) and a capacitor (C) to form a frequency-selective network, used to generate high-frequency sinusoidal signals. The two ends of the inductor (L) are either a spiral coil or a long metal wire, while the two ends of the capacitor (C) are either two non-contact metal conductors or a sharp, diverging metal tip connected to ground. A sharp, diverging metal tip connected to ground more easily forms an oscillating antenna. Under certain energy, frequency, and circuit open configurations, the LC high-frequency oscillator circuit can transmit electromagnetic waves into space (transmitting antenna) or receive electromagnetic waves from space (receiving antenna).
In summary: Electromagnetic waves are a form of energy, sinusoidal waves. High-frequency signals are conducted along the surface of a conductor. High-frequency signals also have a proportion that radiates into space. When a metal conductor has a long section with sharp corners, high-frequency signals are more likely to radiate at the tip or receive external high-frequency electromagnetic waves (interference).
Second, what is the encoder signal frequency?
The encoder signal frequency we often talk about is the frequency of a square wave, also known as the electronic switching frequency (switching characteristics from 0 to 1). However, the frequency transmitted in cables should be calculated based on electromagnetic wave frequencies, that is, electromagnetic wave frequencies composed of more than 19 electromagnetic wave frequencies. Each electromagnetic wave frequency has a multiple coefficient relationship with the square wave frequency. Even if the square wave frequency is not high, the signal still contains some high-frequency electromagnetic wave components, which will become apparent in long-distance transmission or in the presence of high-frequency interference.
The rising and falling edges of a digital square wave signal follow the high-frequency characteristics of electromagnetic waves, while the smooth high and low levels follow the low-frequency characteristics. The frequency of a digital signal is not the same as its electromagnetic wave frequency; higher-frequency electromagnetic waves exist within it. If our digital signal is only below 300kHz, due to the steep rising and falling edges of the square wave, there are still many high-frequency electromagnetic waves present, which is the main reason for the occurrence and reception of various electromagnetic interferences. Similarly, any switching voltage, rising sharply from a low level to a high level, is also a surge combination of electromagnetic waves of various frequencies.
Third, what are the electrical characteristics of the encoder cable?
Since encoder cables transmit electromagnetic waves of various frequencies, the calculations are not based on Ohm's law for DC (or low-frequency) current and voltage, but rather on the capacitance and inductance of AC signal electromagnetic waves. Furthermore, as previously mentioned, higher-frequency electromagnetic waves have the characteristic of propagating along the surface of metallic conductors.
Inter-line capacitance: With sufficiently high-frequency electromagnetic waves and longer cable lengths (more noticeable after 30 meters), the larger surface area between the two core wires within the cable creates inter-line capacitance. High-frequency electromagnetic waves travel through capacitance. Inter-line capacitance also forms between signal core wires and power lines, as well as between signal core wires and the shielding layer. High-frequency electromagnetic waves traveling from one signal core wire to another through this capacitance are called "crosstalk interference," such as crosstalk between phases A and B. Electromagnetic noise traveling on the shielding layer travels through inter-line capacitance to the signal core wire; this is called "external high-frequency electromagnetic interference." Inter-line capacitance is measured in power volts (pF) per meter in the laboratory.
The interference with the signal is caused by the high-frequency signal traveling through the inter-line capacitance.
Inter-line capacitance can change drastically as the cable insulation ages, and some long-used, aging cables may suddenly start losing signals.
Inductive characteristics of a conductor: A sufficiently long conductor will exhibit inductance due to the surrounding electric field that conducts electromagnetic waves. This inductance results in a delay in the electromagnetic wave signal, measured in nanoseconds (ns/m). When the surrounding electromagnetic field is disturbed, this inductive delay effect becomes more pronounced, leading to more uncertain signal delays. Because different frequencies of electromagnetic waves have different effects on inductance, the superimposed square wave signal is already distorted.
This inductive characteristic will be more pronounced if the encoder signal line is wound. Therefore, encoder cables should be routed as straight as possible.
Fourth, what does "encoder twisted pair" mean?
Twisted pair cables are used to transmit signals that are in opposite phases as a pair. For example, in an incremental encoder, the inverse signals are paired, with A+ and A- as one pair and B+ and B- as another pair, and each is transmitted on a twisted pair cable.
The purpose of the "pair" is to combine A+ and A- into a superposition of positive and negative electromagnetic waves from the outside. After being superimposed by the paired twisted pairs, it appears as a single DC field without the transmission of electromagnetic field changes. The electromagnetic field formed by DC transmission is not sensitive to inter-line capacitance, meaning there is no crosstalk interference or external high-frequency electromagnetic interference.
Before twisting:
Transmission after A+ and A- are twisted together:
After the A+ and A- groups of signals are twisted together with the B+ and B- groups of signals, it is equivalent to two DC pairs. The inter-line capacitance has no effect on the almost DC electric field, and the crosstalk between the two groups of signals is reduced.
"Twisting" refers to the pitch of the twisting rotation, which is related to the inductance characteristics of the cable. It is necessary to preset the main frequency range of the signal that the cable will travel through, and obtain the optimal cable inductance characteristics (minimum signal delay and distortion) through laboratory testing.
From both electrical theory and laboratory practice perspectives, the primary function of twisted-pair cables is to simultaneously transmit paired, anti-phase signals over a cable. For single-phase signals, such as only ABZ signals, the significance of twisted-pair cables is minimal. Some "experts" insist on using "twisted-pair shielded cables" even without considering whether the encoder transmits both A+ and A- signals simultaneously, or even if only ABZ signals are transmitted. I don't know their basis for this claim. Feel free to leave a comment to discuss.
5. What is the function of the encoder shielding layer?
Electrostatic shielding: Electrostatic shielding is a closed metal enclosure made of a metal with good conductivity, such as copper, and connected to a ground wire. The device or signal that needs to be shielded is placed inside, so that the external electrostatic interference field does not affect the internal circuit and signal.
Electromagnetic shielding: Electromagnetic shielding also uses a highly conductive metallic material to form a shielding layer, surrounding the circuit to be protected. It shields against interference not from electric fields, but from high-frequency (above 40kHz) magnetic fields. When the high-frequency magnetic field generated by the interference source encounters the highly conductive electromagnetic shielding layer, several reactions can occur: reflection, refraction, absorption, multiple refractions followed by reflection, or it may penetrate and interfere with signals on the internal core wires.
High-frequency electromagnetic waves, similar to optical waves, are mostly reflected back when they encounter an aluminum foil shielding layer or a very dense shielding layer. Lower-frequency electromagnetic waves may partially refract into the shielding layer's metal conductor, with some reflecting again at the conductor's surface. After multiple reflections, they are absorbed, forming an electromagnetic wave "fog" on the outer surface of the shielding layer. This dissipates the energy of high-frequency interference, protecting the circuitry inside the electromagnetic shielding layer from the effects of high-frequency interference magnetic fields. If it's a single layer of aluminum foil, or the shielding layer isn't dense enough, some high-frequency interference will still penetrate the shielding layer and interfere with internal signals. Aluminum foil only reflects very high-frequency electromagnetic waves 100%; many longer-wavelength, lower-frequency electromagnetic waves may still refract through the aluminum foil and enter the interior. Therefore, aluminum foil is generally used as a shielding layer for very high-frequency Ethernet signals. For ordinary incremental pulse signals with lower frequencies, aluminum foil is not effective. For encoder signals, a dense, highly shielding metal mesh shielding layer is preferable.
The shielding layer is not intended for use as a conductor, much less as an equipotential bonding conductor. If grounding at both ends is required, interference signals may be introduced due to the connection. If equipotential bonding is necessary, a thick copper conductor should be used outside the shielding layer to connect the two ends at the same potential, rather than using the shielding layer as an equipotential bonding conductor.
6. Is a thicker signal cable always better? What does a professional encoder cable look like?
This is a serious misunderstanding.
Encoder signals are high-frequency electromagnetic waves, and high-frequency signals travel along the surface of a conductor. They are not like current, voltage, or DC signals, nor are they low-frequency signals that rely on DC resistance. Therefore, the core wire doesn't need to be very thick. While thicker wire does increase the outer surface area, the cross-sectional area is quadratically related to the wire diameter, while the outer surface area of a conductor is linearly related to its diameter. Simply thickening the core wire wastes copper. The thickness of the core wire in a dedicated encoder cable affects the surface area and the design of distributed capacitance. It's not arbitrarily thickened; doubling the cross-sectional area only increases the outer surface area by 40%. A thicker signal core wire can actually alter the inter-wire capacitance design, making it more prone to crosstalk interference or external high-frequency electromagnetic interference. If thicker wire seems better, there must be a matching thickness between the signal and power wires, with capacitive and inductive characteristics specifically designed for the signal frequency.
Some automation project designs requiring encoder signal lines of 0.75 square millimeters or more are misleading. Such thick lines, without considering the cable's capacitance and inductance characteristics, will create significant inter-line capacitance over long distances. This will cause crosstalk between phases A and B at high frequencies due to capacitance, and the signal line will also form capacitance with the shielding layer, introducing leaky high-frequency interference. However, the 0V power supply wire within the cable should be as thick as possible. A thicker 0V wire can form a larger inter-line capacitance with the shielding layer, absorbing and diverting high-frequency interference signals away from the 0V wire. Unused wires within the cable can be combined into a single 0V wire.
For reference, Heidenhain's encoder cables from Germany have signal wires of only 0.14 square millimeters and 0V wires of 0.5 square millimeters.
The data provided for Heidenhain encoder cables is limited. Judging solely from the appearance of the signal cable, the initial impression is that the signal wires are very thin (only 0.14 square millimeters!). This contradicts the claims of some "experts" that "encoder wires should be as thick as possible."
(How much current are you planning to draw?)
The shielding layer has many fine and dense metal wires, achieving almost 100% shielding. Looking at the parameter table above, and combining it with my previous explanation: 4 pairs of 0.14 square millimeter signal wires and 4 pairs of 0.5 square millimeter power wires, the thickness ratio of the signal wires to the power wires is 1:3.5. This means the capacitance ratio formed by the cable surface area is 1:3.5. If high-frequency interference penetrates the shielding layer, it will be diverted to the power wires because the capacitance of the shielding layer is greater than that of the power wires. Heidenhain requires a spare power supply; if a spare power wire is not used, two power wires can be combined to achieve a thicker cable. The cable's capacitance characteristics are a distributed capacitance of 90pF/m and an inductive characteristic with a delay time of 6ns/m. It can transmit up to 150 meters. If two power wires are combined, the transmission distance is even greater, generally exceeding 200 meters.
Over a decade ago, during the synchronous lifting of a submarine at eight lifting points at a naval maintenance base, I used Heidenhain encoders and cables, the longest of which approached 300 meters. The frequency converter was from ABB, and at the time, only ABB frequency converters had push-pull interfaces with inverted signals (A+/A-, B+/B-). Only with such inverted signals, combined with dedicated encoder signal cables, could a stable and high-fidelity signal be guaranteed even after 200 meters. The longest distance reached in that particular project was 300 meters.
Let's take a look at another encoder cable that I customized more than ten years ago (possibly an advertisement):
1. High-density fine-wire shielding layer, with nearly 100% shielding coverage.
2. The signal core wire is 0.079 ultra-fine high-purity soft copper wire, 30 strands twisted. Under the same cross-sectional area of the cable, the multiple strands of fine wire achieve the maximum peripheral surface area, and the high-frequency signal travels on the conductor surface.
3. The power cord is made of 105 strands of 0.079mm ultra-fine high-purity soft copper wire. "Thick" refers to the number of strands, not that it is actually a single thick wire. The capacitance ratio (outer diameter area) of the power cord to the signal cord is 3.5:1.
4. The inter-line capacitance measured by the factory's laboratory is approximately 97 to 100 pF/m.
5. The delay measured by the factory's laboratory is 7 ns/m.
6. This cable is designed for encoder signal transmission over 30 meters and has a pulse frequency (electronic switching frequency) of less than 800KHz.
Are double shielding layers useful?
Double shielding solves several problems. First, the two shielding layers are thick enough to block more electromagnetic interference. Second, it addresses the issue of messy grounding at both ends; the outermost layer can be grounded at both ends, while the inner layer can be floating or grounded at one end. In practice, if equipotential bonding is required at both ends, a thicker copper conductor can be run outside the signal cable. Third, the inner and outer shielding layers act as capacitors, absorbing transmitted high-frequency interference signals. This is similar to how the thicker 0V wire in professional cables forms inter-wire capacitance with the shielding layers. If you understand how effective double shielding is, you may no longer need double-shielded cables.
7. What are some examples of bad encoder cables?
1. Directly replacing the encoder-specific cable with Category 5 eight-core civilian network cable. This is due to being misled by technical support from those who only offer simple explanations of "twisted pair shielded cable." While some Category 5 eight-core communication cables do indeed have twisted pair shielding, that twisting is designed for ordinary network signals with low civilian requirements (allowing for interruptions and resuming). The shielding layer is a thin aluminum foil, which only reflects very high frequencies and does not shield lower electromagnetic waves. Its capacitive and inductive characteristics are incompatible with the encoder's frequency range.
Ordinary Ethernet signals have checksums and fault-tolerant designs, allowing for the recovery of a small number of lost signals. However, encoder pulses and SSI signals cannot tolerate any signal loss.
2. It only emphasizes the DC or AC low-frequency current transmission characteristics of the cable with a sufficiently large cross-sectional area, but neglects the high-frequency design of the signal cable. The number of core wires in the cable is not large, and the cross-sectional area is only large. High-frequency signals pass through the surface. If the high-frequency capacitance and inductance of such a cable are not designed, the high-frequency electromagnetic wave part will be lost and the waveform will be distorted.
3. The shielding layer is not dense enough. Some substandard cables have shielding layers that are either thin aluminum foil or sparse aluminum wires, which do not provide shielding, reflection and absorption of interference.
4. Too many broken and reconnected wires, burrs on the wire ends, and aging. Ideally, the encoder signal should reach the receiver via a single cable, with as few intermediate connections as possible. Too many connections indicate more reflective surfaces and transmitters, increasing internal reflections and external transmissions, leading to more signal glitches and distortion. Uneven waveforms are also more susceptible to interference and signal loss.
As cables age, the performance of the insulation layer changes, the capacitance characteristics change, or leakage occurs. If the cable is frequently moved, or if there are broken or burrs on the core wires, the inter-wire capacitance will be broken down.
5. Inferior and low-priced cable materials, including copper, insulation, filler wire, and shielding layer. This requires assessing whether the supplier has accumulated years of experience with the ISO9000 quality assurance system. For industrial systems, it generally takes more than three years for a supplier with guaranteed quality to establish a stable presence.
8. Where does the interference come from? What exactly is interference?
1. All kinds of non-sinusoidal electromagnetic waves and rising and falling switching signals contain electromagnetic waves of various frequency bands, including high-frequency electromagnetic waves that are more easily radiated outward.
2. A capacitor can be formed between any two open, non-contact metal conductors, or between a sharp metal tip and the ground. A long metal wire or coil can form an inductor. Capacitive and inductive components can form a circuit that oscillates at a specific frequency. This circuit can send or receive radiated signals, especially higher frequency electromagnetic waves, which are more likely to form such radiation and reception.
3. Digital square wave signal generators contain many high-frequency electromagnetic wave components. These high-frequency waves are easily affected by high-frequency electromagnetic waves of the same frequency from space, altering their shape. During transmission through wires, these high-frequency electromagnetic waves are also more prone to attenuation and outward radiation, resulting in the loss of this high-frequency energy. Similarly, this high-frequency component, due to its easy outward radiation, also becomes a source of interference for other digital square wave signals. Signals of various waveforms generated and transmitted by various circuits can be decomposed into high-frequency components using Fourier transform, which are easily attenuated by high-frequency interference from space, and may even interfere with other signal waveforms.
Besides the fact that the digital square wave signal itself is a high-frequency interference source, there are also electrostatic interference, low-frequency and magnetic field interference, and other high-frequency interferences.
4. Electrostatic interference
Static electricity refers to non-flowing charges or potential differences. When two media have different charges or potential differences, and energy accumulates to a certain level, or when the distance is close, or when there is electric field distortion at a sharp end, or when charge conduction from a layer of dust causes breakdown, this static electricity will undergo a breakdown discharge—either dying in silence or exploding in silence. This is a brief reconstruction of the electromagnetic field due to a sudden change in the spatial electromagnetic field layout, releasing energy. The most typical example is lightning. In practical industrial control automation, typical examples include the opening and closing of contact switches (mechanical relays and electromagnetic brakes/solenoid valves), charge accumulation in dry non-metallic media, charge accumulation in dusty environments, sharp corners and burrs on metallic conductors, and the electrostatic difference caused by the non-equipotentiality of various parts of equipment during the transition from power outage to power-on. Static electricity exists in large quantities and can occur at any time, resulting in various micro-discharges. These short-duration discharges can interfere with digital signals, ranging from a small spike in a waveform to damage output and receiving devices.
In industrial control systems, NPN devices are still widely used. The common terminal of these devices is at a high level, while most equipment uses 0V as its common ground. These two "common terminals" pre-establish two non-equipotential electrostatic potential differences, which makes electrostatic interference more likely. The use of NPN encoders and various types of switches should be avoided.
To combat electrostatic interference: equipotential bonding, no metal sharp corners or burrs, dust-free, isolation, and antistatic treatment using non-metallic media.
5. Low-frequency and magnetic field interference
Low-frequency and magnetic field interference mainly originates from power sources, motors, and various coils. Industrial power is 50Hz three-phase or two-phase AC. In areas with significant power, the conduction of AC power (through straight cables and various conductor coils) generates changes in the surrounding electromagnetic field, as well as electromagnetic wave reflection, beat superposition, and harmonics. The instantaneous three-phase imbalance during motor rotation contributes to the external magnetic field, and low-frequency leakage within switching power supplies and frequency converters also plays a role. Low-frequency interference primarily occurs through close-range inductive coupling and direct conduction via dielectric materials (metallic conductors). Low-frequency interference disrupts the energy portion of digital signals, clipping the entire waveform.
To combat low-frequency interference: maintain distance, use magnetic rings or ferromagnetic materials to absorb low-frequency energy, use metal-sealed isolation to protect components, and block low-frequency conduction paths. (Do not connect shielding wires and grounding wires arbitrarily just because you see them).
6. Other high-frequency interference
According to the Fourier transform principle, all changing waveforms can be separated into electromagnetic waves of various frequencies. Except for perfect sine and cosine waveforms, which have only one frequency, all other waveforms contain high-frequency harmonics. These high frequencies not only conduct through the surface of the conductor but also radiate outwards, especially from areas with sharp metal corners and wire burrs. High-frequency interference exists in signal sources, signal cables, and signal receiving units. Areas with sharp metal corners, open terminal sharp corners, and burrs on metal cable ends are more likely to absorb external high-frequency interference. High-frequency interference disrupts the graphic outline of digital signals; when the high-frequency energy is high, it can instantly heat up and damage devices (like induction cookers or microwave ovens).
High-frequency radiation waves are reflected by bright metal surfaces, but absorbed by fine and dense metal wires; this is the function of cable shielding.
To combat high-frequency interference: The design features dense metal shielding that absorbs energy without sharp metal edges or burrs. Changing the capacitance or inductance values in the design alters the LC frequency to prevent high-frequency self-oscillation.
9. Discussion on how to "ground"?
PE protective grounding is essential for personal safety; there should be no high or low potential differences on site that could pose a safety hazard. Electromagnetic compatibility (EMC) tests have also confirmed that a unified grounding connection between AC, DC, earth, and the enclosure provides the best signal interference immunity.
There are often three categories of "grounding" connected to the ground wire, serving as equipotential bonding.
1. A unified grounding system for metal conductors such as the earth, cabinet shell, and machine base.
2. The neutral wire and ground wire of the three-phase balanced AC power supply are connected.
3. The power supply of the DC control system is connected to ground at 0V and the same potential. (Sometimes this part is left floating, or connected to ground via a capacitor.)
The above three types of "grounding" should each have independent routes and follow the shortest conductive path to connect to the earth, and should not be mixed or connected in series. Furthermore, there should be a certain distance between them before grounding. For example, signal lines should be a certain distance away from power lines to prevent near-end induction.
The "conduction" mentioned above needs to take into account how the high-frequency signal (interference) travels.
The three types of grounding should be distinguished, and should not be mixed or connected in series to prevent interference caused by connection.
The three types of "grounding" should be kept at a certain distance before being properly connected to the grounding plate to prevent inductive interference with the signal.
However, EMC requirements are based on laboratory testing, while on-site testing often falls far short of these laboratory requirements. Therefore, a "standard answer" cannot be provided.
1. Grounding for components such as earth, cabinet casing, and machine base can be affected by factors such as poor conductivity, poor contact due to long-term aging, and altered electrical characteristics due to dust accumulation in humid air, all of which reduce grounding effectiveness. Each casing should have a copper busbar or a relatively thick copper wire for grounding, and should ideally be connected in parallel rather than in series to take the shortest route to earth.
2. Three-phase imbalance in AC power supply leads to induced virtual voltage on the neutral wire. The short-duration electromagnetic waves traveling through the ground wire are delayed and may take a "shortcut" via surrounding capacitive loops, interfering with the signal. This type of interference arises from three-phase imbalance in the AC power supply, uneven winding of nearby motors (the imbalance only becomes apparent during startup and acceleration), or uneven magnetic field distribution in permanent magnet motors, causing strong AC energy to interfere with the signal. This type of interference is more pronounced in areas with high-power motors nearby.
3. Without isolation between various devices on a 0V power supply, high-frequency electromagnetic waves from internal components related to DC power supplies, such as switching power supplies, frequency converters, solenoid valves, and relays, are briefly exposed and travel directly on the 0V line. Due to the delay of high-frequency electromagnetic waves on conductors, a drop in voltage is created. This cannot be predicted by simply connecting a DC resistor to the voltage.
The interference induced on the 0V and power supply is mostly caused by short-duration high-frequency interference from the frequency converter (internal AC-DC-AC inverter), switching power supply (also internal DC-AC-DC inverter), solenoid valve, electromagnetic brake, large contactor, etc. Sometimes linear power supply or power isolation is selected.
Therefore, we need to determine what equipotential bonding is (how the AC circuit "conducts" under high-frequency conditions) and how to ground it. It's not about being obsessive about grounding and simply "grounding" any open wire. We need to understand what type of ground it's connected to, its routing, whether it passes through interference zones to induce interference, or whether direct connection introduces interference. Such haphazard grounding might actually introduce interference.
4. Is the encoder housing and shielding wire grounded?
In principle, the encoder housing and shielding wire belong to Class 1 protective grounding (shielding grounding). Many encoder housings and shielding wires are connected by contact, while others are connected by capacitors for AC (high-frequency) grounding. Grounding is achieved through the encoder cable shielding layer, and some are even grounded together with the 0V power supply. However, in many cases, the encoder shaft is connected to the motor, and the connection between the encoder shaft and housing is through the balls of the ball bearings, which are in frictional contact during rotation. Sometimes, due to the influence of three-phase power imbalance, poor conductivity between the shaft and housing through the ball bearings, micro-discharges on the balls, and the high-frequency conduction delay of the housing and shielding layer conductors, a short-duration electromagnetic wave drop can easily form between the shaft end, housing, and 0V power supply. Improper mixing of all three types of grounding can introduce interference. Therefore, some encoder housings choose to leave the shielding wire suspended from the housing, allowing users to choose how to ground the housing and shielding layer according to various electrical conditions at the site.
Is "direct conduction" the only way to ground the encoder cable shield? Sometimes connecting a capacitor is also "high-frequency grounding," and sometimes one end of the shield is left floating or even both ends are left floating, which is also high-frequency capacitor grounding. When the encoder wire is longer than 30 meters, the shield and the power line inside the cable form an inter-line capacitance. High-frequency interference is poured into 0V and grounded through the inter-line capacitance. Leaving one end floating can prevent interference from being directly introduced from the interference source.
When encoder signal transmission over long distances, requiring an additional 100-meter signal cable, I generally recommend grounding the shielding layer at the receiving end and leaving it suspended near the encoder. With sufficient cable length, the shielding layer and the cable's 0V power supply form a capacitive ground, preventing direct conduction between the encoder housing, cable shielding layer, and the receiving end's 0V. Due to the long distance, there will inevitably be a time delay in the conduction of high-frequency interference. Directly mixing and grounding these three components would create an electromagnetic wave drop along the line, introducing interference that travels along the shielding layer and affects the internal signal.
5. The encoder has A+/A- and B+/B- signals. However, the receiver only has ABZ and no inverting interface. Is the twisted-pair shielded cable still useful?
For longer distances (greater than 30 meters), a twisted-pair shielded encoder cable is required. The encoder's A+/A- and B+/B- terminals must be paired and run on twisted-pair cables, and the encoder must have both A+/A- and B+/B- terminals. If the receiver does not have A- and B- interfaces, AB- can be left floating at the receiver (voltage and inter-line capacitance will still exist on the lines), or a terminating resistor (120 ohms) can be connected to the 0V of the receiver.
10. Example: How can an encoder signal be transmitted 200 meters?
To better understand the characteristics of the high-frequency portion of a square wave, below is a typical oscilloscope graph of an incremental square wave signal after passing through a 200-meter cable:
In this process, the steep rising and falling edges disappear, replaced by sloping rising and falling edges. Additionally, a small, protruding peak appears in the middle of the originally flat square wave.
We already know that high-frequency electromagnetic waves are more easily attenuated and radiated outwards in conductors. In the incremental AB phase square wave signal in the conductor, the steep rising and falling edges actually contain a large portion of the high-frequency components. During the transmission process of 200 meters of conductor, these components have been attenuated or radiated outwards. Therefore, what remains is more like the rising and falling (slope) of a low-frequency electromagnetic wave signal. The rising edge time of the A phase signal shows a small peak in the B phase signal, which is the crosstalk interference of the high-frequency components of the A phase into the B phase signal.
GI58N (A+/A-, B+/B-, Z+/Z-) encoder + 200-meter long-distance dedicated encoder signal cable (model: Jingpu F600K0208), waveform at the receiving end after 200 meters at high speed.
Due to the high frequency and the fact that the signal is slightly distorted after being transmitted over a distance of 200 meters, the superposition of reflected waves and crosstalk are very small, the square wave pattern is still clearly distinguishable, the bottom is less than 0.7V, the signal quality meets the requirements of the incremental pulse receiver, and the receiver counts accurately.
Please note: For encoder signals longer than 30 meters, an encoder output with A+/A- and B+/B- inverted signals must be used, including 5V differential or 10-30V push-pull type with inverted signals. Furthermore, the signal cable transmission must be A+/A- and B+/B- paired twisted-pair transmission. At the receiving end, a receiver with A- and B- interfaces should be used whenever possible. If it is an ABZ single-phase encoder, it is recommended to leave A- and B- terminals floating at the receiving end, or to connect a 120-ohm terminating resistor to 0V.
Friendly Remarks:
The above discussion focuses on incremental pulse signals and SSI serial signal cables (including Endat and Biss signals).
There is no standard answer to how to solve signal interference problems on-site, and there is no single article that can explain it all.
For those who prefer to quickly save answers, please focus on the explanation of principles in the text. You should analyze different phenomena on-site, find answers in practice, and try various anti-interference solutions.
