The effect of friction and wear behavior on the signal transmission of brush-ring system during electrical contact sliding

doi:10.21203/rs.3.rs-2219207/v1

Download PDF

Research Article

The effect of friction and wear behavior on the signal transmission of brush-ring system during electrical contact sliding

https://doi.org/10.21203/rs.3.rs-2219207/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Copper-graphite brush and slip-ring systems are often used to transmit electrical energy and signal between moving and stationary parts. At present, the research on the brush-ring system mainly focuses on power transmission, while on signal transmission is relatively lesser. In this study, the effects of friction and wear behavior on signal waveform distortion were in-situ analyzed using a custom-designed tribotester, which can specially synchronize the slip-ring rotation period with the input signal waveform. Results were analyzed comprehensively from tribological and electrical aspects to find out the key factors, including friction coefficient, friction temperature, contact resistance, surface morphology, roughness, wear particles, and compositions. It was found that the distortion of signal waveform is mainly affected by the friction film. During the sliding electrical contact motion, the increase of friction film will increase the contact resistance, which in turn increases the signal waveform distortion. At the same time, under the normal load and shear force, the friction film moves along the sliding direction, hence causing the signal phase angle to shift. From a tribological point of view, the friction coefficient and temperature decrease with increasing the friction film. So, for the brush-ring system, the formation of friction film is good for the friction stability but bad for signal transmission.

Sliding electrical contact

Copper-graphite brush

slip ring

Signal waveform distortion

Friction film

Contact resistance

The current-carrying friction pairs are typically regarded as the “aorta” of power transmission (PT) and the “neural network” of signal transmission (ST), known as the single point of failure component [1]. If it fails, that will cause the entire system failure. Therefore, a good sliding electrical contact component can guarantee power and signal transmission [2]. In recent years, with the development of electrical contact engineering, copper-graphite composite has become one of the widely used materials [3]. Usually, the material composed of copper, resin and graphite, so has good lubricity, excellent electrical conductivity, thermal conductivity and ductility, and is often used in hydro-generator, wind turbine blade pitch control, rotating radar antennae, medical CT scanners, and armored vehicle turrets [4–6].

In the field of PT, a large number of researches have been carried out and many practical application results were obtained and provided much valuable information for the design of electricity transmission engineering [7]. Different from PT, the ST does not have the problem of resistance heat or arcing corrosion, but any slight changes in the contact interface will have a significant impact on signal integrity [3]. Usually, signal integrity refers to the signal quality in the circuit system [8], where the signal should be transmitted from the source to the receiver without distortion. Nowadays, with the increase of signal rate, the ST can not only consider the realization of the logic but also how to make the device receive the correct signal waveform. Therefore, the research of signal integrity is also about how to optimize the signal transmission path so that the receiver can obtain the correct signal waveform.

Transferred data integrity is an important characteristic for the brush-ring system [9]. To explore the critical parameters, several special sliding electrical contact tribotesters were designed, which can be divided into three types. First, the signal with pulse type was supplied by DC power and measured by PC data acquisition system, where the surface morphology were examined by a surface profilometer to analyze the loss of electrical signal at the sliding interface [3, 4]. Second, signal noise was measured by using a digital recorder which bandwidth is much higher than the rotational frequency used in the experimental, where the occurrence of contact interruptions was checked by using an oscilloscope to understand the relation between different wear regimes, electrical noise and life time in sliding electrical contacts [10–13]. Third, a particular device equipped with two brushes per circuit is used to transmit high-speed data to evaluate contact noise or variation in contact resistance [14]. Results show that during the sliding motion the signal noise or distortion were produced by the change of contact, which depends on some parameters, including changes in the surface film, electromagnetic interference, operating parameters (i.e. normal load, sliding speed or vibration), and physical parameters (i.e. geometry shape, surface roughness, or wear particles). However, due to the input signals are pulse or square wave type, or the signal frequency is much higher than that of the slip ring, fast Fourier transform is usually used to transform the output noisy signal from continuous time domain to fixed frequency domain to analyze the change of noise amplitude. Therefore, it is difficult to establish the relationship between the signal waveform distortion and the friction and wear characteristics of the brush-ring system.

This study aims to propose a new test device to evaluate the effect of friction and wear behavior on the signal distortion of brush-ring system. Owing to the input signal waveform is synchronized with the rotation frequency of the slip ring, the change of signal waveform distortion during the whole lifetime can be obtained through the analysis of time domain. Further, the friction and wear characteristics of the brush-ring surface were in-situ online analyzed, so the correlation between the tribology behavior and the signal waveform distortion could be established directly.

2.1 Testing devise

A dedicated tribotester with an online analysis device was designed and built to investigate the friction and wear behavior on the ST of brush-ring system. As shown in Fig. 1. The device consists of a main shaft, one side of which is equipped with a resistance sensor and applied a constant load, and the other side is fixed with a brass-ring sample and stuck an angle scale. The shaft is driven by a three-phase motor (FUJI Electric MLH8075M-0.2KW-6P). Owing to the constant load, as the shaft rotates at a constant speed, a periodic sinusoidal signal will be generated from the sensor (see section 2.2 for detail). The sensor and the slip ring are fixed on the shaft, when the shaft rotates one cycle, a complete sinusoidal signal will be transmitted synchronously from the ring to the brush. Since the rotation period of the slip ring is the same as that of the sinusoidal waveform, the friction and wear behavior resulting from interface will affect the transmitted signal waveform, thus establishing the direct correlation between the physical tribological behavior and the transmitted signal quality.

To in-situ analyze the morphology and roughness of the worn surface, the sliding process was suspended every 10000 cycles (the total number of sliding cycles is 80000). Then the worn surface at the degree of 0, 90, 180, and 270 were measured by using an electron microscope and a surface roughness instrument, respectively. Finally, based on the roughness data as well as the friction coefficient (FC), friction temperature (FT), contact resistance (CR), wear particles, and their composition, the effects of the friction and wear behavior on the signal waveform distortion during the whole life time were analyzed.

2.2 Signal generation principle

It is well known that the strain ε of the resistance strain sensor has a linear relationship with the resistance change rate ∆R/R [15], thus conversely under an alternating load, the sensor will generate alternating resistance. Based on this principle, as shown in Fig. 1, we first stuck a self-temperature compensating resistance sensor (BE120-3AA) on the main shaft with glue (KAFUTER 502), then on the end of the shaft a vertical downward 50 N pressure was applied to form a cantilever beam structure. Thus, under the load 50 N, as the shaft rotates at a constant speed, the sensor strain will be changed alternately in tension and pressure, hence resulting in an alternating sinusoidal resistance.

Usually, the resistance change is small, so the direct-current Wheatstone bridge [16] is used for measurement. As shown in Fig. 2, the bridge consists of four strain gauges R₁, R₂, R₃, and R₄ (i.e. four bridge arms), where points C and D are connected to the input voltage, while A and B are the voltage output points, and G is a galvanometer for checking the voltage in the branches.

When no current flows through the galvanometer G, the bridge is in balance, thus the resistance of the four bridge arms satisfies the following relationship [17]:

R ₁×R₄ = R₂×R₃ (1)

As the main shaft rotates, the resistance R₃ will change with the rotation. For instance, when the strain gauge turned to the top, R₃ increased due to the tensile force. On the contrary, R₃ decreases as turned to the bottom. Additionally, as the strain gauge turned to the midline of each side, R₃ is zero due to the balance of tension and pressure. So, under a uniform load, the resistance R₃ changes sinusoidally. Based on Eq. (1), the output voltage E can be calculated by the following formula:

$$\Delta E=\frac{U}{4}(\frac{{\Delta {R_1}}}{{{R_1}}}-\frac{{\Delta {R_2}}}{{{R_2}}}+\frac{{\Delta {R_3}}}{{{R_3}}}-\frac{{\Delta {R_4}}}{{{R_4}}})$$

Where U is the supply voltage, ∆R₁, ∆R₂, ∆R₃, and ∆R₄ are the change values. During the test, the resistances of R₁, R₂, and R₃ are fixed 120 Ω, thus ∆R₁=∆R₂=∆R₄ = 0. So, Eq. (2) can be described as:

$$\Delta E=\frac{U}{4}\frac{{\Delta {R_3}}}{{{R_3}}}$$

Here, the voltage ∆E is linearly related to ∆R₃, which changes in a sinusoidal waveform during rotation, so ∆E can be used as a standard input sine voltage signal. In the circuit, the brush and ring are connected directly in series with R₃, thus the friction and wear at the interface will directly affect the signal quality.

2.3 Signal waveform distortion

Usually, the SW distortion refers to the variation degree of the actual SW relative to an ideal SW [18]. So, the SW distortion is a ratio with the definition of both frequency domain and time domain. The result of frequency domain analysis is the average effect of the distortion in the sinusoidal waveform period, so cannot give the distribution of the waveform distortion in a sinusoidal waveform period. However, the distortion defined in the time domain is mainly obtained by waveform measurement, so the distortion value of a sinusoidal waveform can be given quantitatively at each phase angle, which is beneficial to accurate location analysis of the relationship between surface wear and waveform distortion. In this case, we assume the sinusoidal voltage outputted by the resistive sensor is expressed as

$$E(t)=A\sin (2\pi ft+\varphi )$$

Where E(t) is the instantaneous voltage value of the signal, A is the peak amplitude of the signal, f is the frequency of the sinusoidal signal, and φ is the initial phase of the sinusoidal signal.

As signal passes through the brush-ring interface (Fig. 2), due to the influence of friction and wear, the output voltage is

$${E_0}(t)=A\sin (2\pi ft+{\varphi _0}){\text{+C}}$$

Where E₀(t) is the instantaneous voltage value, φ₀ is the initial phase of the signal, and C is the value of the direct current component of the signal.

Therefore, there is a deviation function δ(t) between the input and output voltages, namely δ(t) = E(t) - E₀(t), which can be expressed in the form of phase coordinates as δ(φ/2πf). For any phase φ and its neighborhood interval [φ-θ/2, φ + θ/2], there exists

$$\overline {\delta } \left( {\varphi ,\theta } \right)=\frac{1}{\theta }\int_{{\varphi - \theta /2}}^{{\varphi +\theta /2}} {\delta \left( {\frac{\varphi }{{2\pi f}}} \right)} d\varphi$$

$${D_c}\left( {\varphi ,\theta } \right)=\frac{{\overline {\delta } \left( {\varphi ,\theta } \right)}}{{A/\sqrt 2 }}$$

Where, D_c(φ, θ) is the SW distortion in the interval [φ-θ/2, φ + θ/2]. We can adjust the resolution ratio of SW distortion by changing the size θ. So D_c(φ, θ) can quantitatively describe the distortion in any phase interval and can compare the distortion differences of different waveform cycles, more can physically locate the waveform distortion.

2.4 Testing materials

In this case, the copper-graphite brush is a commercial product (J204, 8×10×32 mm³) with 25 vol% copper and 75 vol% graphite (Fig. 3). So, the main component is graphite. As described by the manufacturer, the brush with a resistivity of 0.96 Ω⸱m was fabricated by the powder metallurgy technology using raw materials copper, graphite particles, and resin additives. The microstructure is shown in Fig. 3a, where the white particle is copper and the dark gray base is graphite. It can be seen that copper particles are uniformly distributed in the graphite matrix. The hardness of copper particle, graphite base, and their boundaries were tested and listed in Fig. 3c. It shows that the hardness of copper particle (35 HV₁₀) is high than that of graphite base (21 HV₁₀), thereby under the same friction and wear conditions, the graphite base will be easily worn. It was found that some cracks appear around the boundary indentation (37 HV₁₀), indicating that copper particles are not firmly combined with the graphite. The hardness of the ring with an outer diameter of 14 mm and an inner diameter of 10 mm is relatively uniform, such as 129, 134, and 137 HV₁₀, with an average of 133 HV₁₀ (Fig. 3d). Obviously, the ring hardness is much higher than that of the brush.

2.5 Testing analysis method

During the sliding motion, the surface morphology was in-situ analyzed using a SHOCREX DP1920 electron microscope (EM) with 16 million pixels, and the surface roughness was in-situ online measured using a TR240 roughmeter. The friction coefficient µ was calculated using

µ = F/F_n (8)

where the frictional force F and the normal force F_n were measured using S-type force sensors (0 ~ 5kg). In this case, the F_n is 20N, and the ring speed is 5 r/s (corresponding to a linear sliding velocity v = 220 mm/s). The µ was continuously and automatically recorded as a function of sliding time. The FT was measured online using the dynamic thermocouple method. The contact resistance was measured by using the four-point electrical circuit (also called Kelvin sensor [19]). The optical microstructure (OM) characteristics were examined using a NEOTHOT-32 Zeiss optical microscope. The volume fraction of copper and graphite in Fig. 3 was measured using ImageJ software. The worn surfaces were examined using JEOL JSM 6510A scanning electron microscope (SEM), INCA Energy 300 energy dispersive spectroscopy (EDS), and X-ray photoelectron spectroscope (XPS). Micro-hardness of the brush and ring were tested using an MH-5 Vickers hardness tester with testing load 10N and dwell time 15s. The main testing conditions are listed in Table 1.

Table 1

Test operating parameters and conditions.
Parameter	Value
Brush material	Copper-graphite (25 vol% Cu, 75 vol% C)
Brush size	8×10×32 mm³
Brush normal force	20 N
Brush resistivity	0.96 Ω⸱m
Slip-ring material	Brass alloy (65 wt% Cu, 35 wt% Zn)
Slip-ring resistivity	0.0172 Ω⸱m
Slip-ring nominal diameter	14 mm
Slip-ring rotational speed	300 RPM
Sliding speed	220 mm/s
Total sliding cycle	80000
Electrical voltage	5 V
Resistance strain sensor size	3.5×6.4 mm²
Strain sensor resistance	120 Ω
Main shaft load	50 N
Data acquisition frequency	500 Hz
Atmosphere	Ambient air
Ambient temperature	23 ^oC
Ambient humidity	35–65% rH

3.1 Evolution of signal waveform distortion

During the whole 80000 cycles, the evolution of SW of brush-ring system are shown in Fig. 4, where the first five SWs of every 10000 cycles were selected for comparison. It shows that the initial signal is nearly an ideal sine waveform without noise or distortion (Fig. 4a). As the sliding cycle increases to 10000, the signal is still a sine wave with an amplitude of -1.28 ~ 1.52mV, however, appeared a small amount of ringing, showing the SW was slightly distorted (Fig. 4b). When the cycle increases from 20000, 30000 to 40000, the SW still presents a sine wave, the minimum value of the amplitude increases from − 1.23, -1.16 to -1.11mV, while the maximum value increases from 1.59, 1.78 to 1.98mV. During the sliding periods, the noise on the SW does not increase significantly (Fig. 4c-e). However, as the cycle increases to 50000, the SW evolves from a sine wave to an approximate triangular wave. Moreover, a large number of glitches appear on the SW (Fig. 4f). As the cycle continues to increase to 60000, the SW is very uneven and appears abrupt noise, even some sub-peaks (dotted circle) appeared (Fig. 4g). When the cycle increases from 70000 to 80000, the SW become more and more distorted. It is hard to identify what characteristics the SW is (Fig. 4h, i). Obviously, in the whole life time of brush-ring system, with the increase of the sliding cycle, the local noise or distortion continue to increase, by which the signal waveform was destroyed.

In order to further analyze the signal distortion, as listed in Fig. 5, we select a cycle waveform from each 10000 cycles in the same time domain. Since the rotation period of the ring is synchronized with the sinusoidal waveform, the friction and wear behavior on the signal distortion can be continuously reflected on the waveform with same phase. In other words, the waveform distortion at each phase angle is the result of the accumulated friction and wear on the ring surface. Results show that not only the signal amplitude but also the phase angle were all increased (dotted area in Fig. 5).

In addition, the distortion at each phase angle (θ = 3.6^o) was calculated in Fig. 6. It shows that the signal distortion increases with the cycle number, and varies with the phase angle (Fig. 6a), such as the distortion in ranges 60^o ~ 350^o increase rapidly, while the value of the other angles are relatively slower, indicating that over the entire surface the friction and wear have a non-uniform effect on the SW distortion. The relationship between the average distortion and the sliding cycle was fitted in Fig. 6b. It shows that the relationship satisfies the exponential equation: y = exp(-2.83 + 5.3e^− 5x-9.2e^− 11x²), where x and y are the sliding cycle and average distortion, respectively. The R-square of the equation is 0.95, hence the relationship between x and y is strong, indicating that the increase of sliding cycle will directly aggravate the signal distortion.

3.2 Evolution of friction and wear behavior

3.2.1 Friction coefficient

The FC of each 10000 cycle (duration 2000 s) under normal load 20 N and sliding speed 220 mm/s was shown in Fig. 7a. It indicates that, during 10000 cycles, the FC continues to increase from 0.288 to 0.383, while during 20000 cycles, the FC gradually decreases from 0.405 to 0.385, and similar to which, the FC during 30000, 40000 and 50000 cycles decreases from 0.397 to 0.367, 0.376 to 0.353, and 0.362 to 0.342, respectively. The change rate of FC all exceeds 0.02. However, during 60000, 70000 to 80000 cycles, the FC decreases from 0.349 to 0.339, 0.344 to 0.335, and 0.341 to 0.337, respectively. The change rate of FC is all below 0.01. In the whole sliding cycle, the FC first increases rapidly and then decreases slowly, indicating that the friction of brush-ring system is gradually stabilized.

The average FC of each 10000 cycles was calculated in Fig. 7b. It shows that from 10000 to 80000 cycles the FC first increases from 0.357 to 0.398 then decreases from 0.398 to 0.338, and a maximum FC appears at 20000 cycles. The relationship between the average FC and sliding cycle can be fitted with extreme model: y = 0.338 + 0.059*exp(-exp(-z)-z + 1), where z = (x-21467)/9703, y is the friction coefficient and x is the sliding cycle. The R-square of the model is 0.99, indicating that the relationship between y and x is strong.

3.2.2 Friction temperature

The FT of each 10000 cycles under normal load 20 N and sliding speed 220 mm/s was shown in Fig. 8a. It can be seen that, in the 10000 cycles, the FT gradually increases from room temperature 23^oC to the maximum temperature 49.3^oC during 2000 s. However, in the 20000 cycles, the FT increased to 49.3^oC within only 830 s, then continued to rise to the maximum 50.1^oC at 1250 s, finally decreased to 48.8^oC at 2000 s. In the 30000 cycles, the FT increased to the maximum 47.9 ^oC at 900 s, then the FT remains relatively stable during the rest time. In the 40000 cycles, the FT increased to the maximum 47.1 ^oC at 805 s, then similarly to the 30000 cycles the temperature remains relatively stable during the rest time. But in the subsequent 50000, 60000, 70000, and 80000 cycles, the FT curves almost overlapped with little change, showing a stable friction period.

Further, the average FT was calculated in Fig. 8b. It shows that during the 80000 sliding cycles the average FT increased rapidly from 43.65 to 47.29 ^oC then slowly decreased to 44.28 ^oC. The relationship between the average FT and sliding cycle is fitted with extreme equation: y = 44.2 + 3.29*exp(-exp(-z)-z + 1), where z = (x-22758)/6647, y and x is the average FT and sliding cycle respectively. The R-square of the equation is 0.93, indicating that the FT and sliding cycle have a certain correlation. Comparing Fig. 7b and Fig. 8b, it is easy to find that the FC and FT are positively correlated. As reported in Ref.[20], the friction heat (q) is decided by q = µF_nv, where µ is the FC, and v is the sliding velocity. In this case, the normal force F_n and the sliding velocity v are two fixed 20 N and 220 mm/s respectively. So, the friction heat q is proportional to the friction coefficient µ.

3.2.3 Morphology of worn surface

During the 80000 sliding cycles, the morphology of ring surface near phase angles 0^o, 90^o, 180^o, and 270^o were in-situ observed by EM (see Fig. 9). It can be seen that the surface morphology near phase angles 0^o, 90^o, 180^o, and 270^o are very different. For instance, after 10000 cycles, the worn surface near 0^o is relatively smooth with slight wear tracks and a few micro-bulges (Fig. 9b₁), while the surface near 90^o is smooth with some unworn areas (Fig. 9b₂), and some wear tracks appeared near 180^o (Fig. 9b₃), also these tracks appeared near 270^o (Fig. 9b₄). As the sliding cycle increases to 20000, the wear tracks near 0^o, 180^o, and 270^o gradually disappeared (Fig. 9c_1,3,4), and the unworn areas reduced near 90^o and 270^o (Fig. 9c_2,4). After 30000 sliding cycles, a few micro-bulges appear on all surfaces (Fig. 9d_1 ~ 4). And after 40000 sliding cycles, the size and area of micro-bulges increased significantly. As the sliding cycle increases to 50000, the unworn area near 90^o disappeared (Fig. 9f₂), and a large number of micro-bulges accumulated to form a black mud-like body near 0^o, 90^o, and 180^o, indicating that the adhesive wear occurred. However, the surface near 270^o is relatively smooth (Fig. 9f₄). After 60000 sliding cycles, some scrape tracks appear in nearly the same position on all surfaces, where many mud-like bodies adhered after 50000 cycles, indicating that the morphology and structure of the worn surface are dynamically changing during the sliding process. As the sliding cycle increases from 70000 to 80000, the mud-like film gradually increases, as well as the scratched area. Therefore, during the sliding friction, the friction films are continually wiped off and re-formed on the worn surface. Due to the friction and wear behavior varying with the surface profile, structure, and film, the electrical signal is distorted at different angles.

3.2.4 Roughness of worn surface

The roughness of ring surface was in-situ measured every 10000 cycles and shown in Fig. 10. It shows that, during the sliding process, the height, length, width, density, shape, and texture of the wear track continues to change. For example, after 10000 cycles, the max depth and height of the wear track are − 0.73µm and 0.91 µm, respectively. As increasing to 20000 cycles, the max depth and height of wear tracks were reduced to -0.63µm and 0.82 µm, respectively (Fig. 10b). Meanwhile, the width of the wear track is also reduced. So, the worn surface is smother compared to 10000 cycles. As the sliding cycle increases to 30000, the max depth and height of wear tracks were increased to -0.91 and 1.19 µm respectively. The width and density of wear tracks were also increased significantly (Fig. 10c). As increasing to 40000 cycles, the max depth and height of wear tracks were increased to -0.97 and 1.22 µm respectively. However, the density and shape of wear tracks were not changed obviously (Fig. 10d). After sliding 50000 cycles, the max depth and height of wear tracks were increased to -1.25 and 1.35 µm respectively, and most tracks were close to 0.91 µm in height and 0.06 mm in width (Fig. 10e). After 60000 sliding cycles, the inheritance of wear track appears on some local areas (dotted area in Fig. 10f), where the wear tracks also appeared after 50000 cycles, even more, appeared after 70000 and 80000 cycles. In addition, the average depth and height of wear tracks were calculated separately in Fig. 11. It shows that before 20000 cycles the average height first decreases from 0.334 to 0.289 µm, then after 20000 cycles it increases from 0.351, 0.357 to 0.411µm. Overall, the surface roughness tends to increase with increasing the sliding cycle.

3.2.5 Friction film

The SEM images of ring surface were shown in Fig. 12. The top view shows that the surface was covered with plenty of films and non-uniformly distributed with wear tracks (Fig. 12a). The side view shows that the micro valleys filled with films present a multi-layered structure with an uneven thickness (Fig. 12b). In addition, the higher magnified SEM images show that some cracks formed in the film (Fig. 12c). Therefore, during the sliding motion, these films are easily slipped or peeled off from the worn surface. To understand the film effects on the FC, FT, roughness, and signal distortion, the compositions were analyzed by EDS. It shows that the film is mainly composed of C (24.22 at%), Cu (42.82 at%), and O (26.47 at%) (Fig. 12d ~ f). The C rich film indicates that plenty of graphite was transferred from the brush to the ring.

In addition, Raman spectroscopy was used to analyze the chemical states of Cu and O in the film. As shown in Fig. 13a, the primary peaks were identified as C1s, O1s, Cu 2p, and Zn 2p, indicating the existence of C, O, Cu, and Zn elements. Further, Fig. 13b shows that the fine spectrum of C 1s could be fitted into three peaks at 282.42, 284.40, and 288.54 eV, hence associating with graphitic sp² carbon (C = C/C-C) [21, 22]. The fine spectrum of O 1s consists of three oxygen contributions at 530.34, 531.73, and 533.52 eV (Fig. 13c), thus corresponding to oxygen bonded to Cu [23, 24]. Figure 13d shows that the two peaks around 932.30 and 952.14 eV correspond to Cu 2p_3/2 and Cu 2p_1/2 peaks of Cu⁺, respectively, hence confirming the existence of Cu₂O in the friction film [25–27].

3.2.6 Contact resistance

During the sliding process, the CR between brush and ring was measured online every 10000 cycles (Fig. 14). It shows that the trend of CR is different in each cycle stage (Fig. 14a). For instance, during the 10000 cycles, the CR drops rapidly from 0.27 to 0.15 mΩ in the first 300 s (arrow #1), then stabilizes at 0.15 mΩ from 300 to 1200 s (arrow #2), finally increases slowly from 0.15 to 0.22 mΩ (arrow #3). While in the 20000 cycles the CR decreases slowly from 0.29 to 0.23 mΩ in the first 600 s and then stabilizes around 0.22 mΩ. Yet, during the 30000 cycles, the CR is close to 0.23 mΩ without significant change. However, during the 40000 cycles, the CR increases gradually from 0.28 to 0.39 mΩ. Then after 40000 cycles, the CR continues to increase, such as from 0.42 to 0.44 mΩ (50000 cycles), from 0.43 to 0.55 mΩ (60000 cycles), from 0.56 to 0.65 mΩ (70000 cycles) and from 0.65 to 0.70 mΩ (80000 cycles). In addition, the average CR of each 10000 cycle was shown in Fig. 14b. It can be seen that as the sliding cycle increases, the CR increases continuously from 0.17, 0.23, and 0.27 to 0.68. The relationship between the contact resistance (y) and sliding cycle (x) fits the exponential model: y = exp (-1.96 + 2.53e^− 5x-7.2e^− 11x²). The R-square of the model is 0.998, indicating the relationship between y and x is strong.

4.1 Friction and wear behavior of brush-ring system

From a frictional point of view, the life time of a brush-ring system can be divided into three stages (see Fig. 7): ascending stage, descending stage, and steady stage. In the ascending stage, the FC gradually increased by 0.095 over 2000 seconds with an average increase of about 4.75×10^− 5 per second. The increase of FC can be inferred from the surface morphology and structure (Fig. 9b). It shows that the real surface of the ring is not flat but comprises many asperities. For their hardness is much higher than that of the brush (Fig. 3), so as sliding contact is made between brush and ring, these harder asperities are just like many plowshares cutting the brush. So the main friction mechanism of the brush-ring system is ploughing friction. During the friction, a part of the particles ploughed out from the brush would be extruded and dropped off the brush-ring interface. While under the normal force F_n, the other parts would be coated on the ring surface. As a result, the ring surface appears to be polished and becomes smoother and smoother (Fig. 9), hence changing from ploughing friction to adhesion friction.

According to the classical theory proposed by Bowden et al. [28] in their study of the ploughing and adhesion of sliding metals, the friction coefficient (µ) consists of two parts, ploughing (µ_p) and adhesion (µ_a) respectively, i.e., µ = µ_p + µ_a. The theory also applies to biology [29], composite [30], and non-metals, such as MoS₂ [31], carbon and glass fabric [32]. As reported in Ref.[28, 33, 34], the FC induced by ploughing (µ_p) could be calculated from the relation

$${\mu _p}=\frac{{{F_p}}}{{{F_n}}}=\frac{{{A_s}{\sigma _{s2}}}}{{{A_r}{\sigma _{s1}}}}$$

where F_p is the ploughing force, A_r and A_s are the real contact area and the cross-section area of the track respectively, σ_s1 and σ_s2 are the yield strengths in normal and sliding directions respectively.

In addition, the FC induced by adhesion (µ_a) could be calculated from the equation

$${\mu _a}=\frac{{{F_a}}}{{{F_n}}}=\frac{{{A_r} \cdot \tau }}{{{F_n}}}$$

where F_a is the shear force, and τ is the shear strength.

During the initial stage, due to the transformation of graphite from brush to ring, the ring surface gradually becomes smoother, thus the real contact area A_r becomes larger, yet the scratch area A_s becomes smaller. As a result, the µ_p decreases, while the µ_a increases, then taken them together, the friction coefficient µ gradually increases.

In the descending stage, due to the adhesive wear, graphite particles accumulate to form micro-bulges, by which the abrasive wear occurs. The compositional analysis show that the film is rich in copper oxide (Fig. 13), indicating that in addition to abrasive wear, oxidation wear also occurs. According to Ref.[35], the oxidation film has excellent friction-reducing properties, hence decreasing the FC. Finally, in the steady stage, the formation and wear of the friction film reach a dynamic balance, thus entering a stable friction.

4.2 Friction and wear behavior on signal waveform distortion

The friction and wear behavior on signal transmission were analyzed comprehensively from tribological and electrical aspects to find out the key factors, including FC, FT, CR, surface morphology, roughness, wear particles, and composition. Results show that the FC changes greatly in the initial stage then later tends to be stable (Fig. 7). Contrarily, the initial signal waveform is clean, yet later the signal distortion gradually increases (Fig. 4). So, the opposite trend indicates that the FC is not the cause of signal distortion. When contact is made between ring and brush, friction film formed on the ring surface increase with the transferred particles. At the same time, the surface roughness changes with the friction film. So, from a macro viewpoint, the surface roughness appears as a morphological inheritance at some angles (Fig. 10). In addition, at these angles, the signal distortion increases with the sliding cycle (Fig. 5). Therefore, we can infer that the increase of signal distortion is related to the surface roughness, or more specifically, the increase of friction film.

For brush-ring system the mechanism of friction film on signal distortion is illustrated in Fig. 15. Firstly, as the brush contacts with the slip ring, current can only transfer where the surfaces are in actual contact [36]. The CR in the brush-ring system is mainly related to the surface structure [37]. When brush and ring contacts at some asperities (Fig. 15a), the electric current lines bundle together to pass through the individual spots (known as the a-spots [38]). The resistance of the individual spot is usually called shrinkage resistance [14]. As a rule, the shrinkage resistance (R_s) can be calculated from Holm’s famous equation [38]:

R _s = (ρ₁ + ρ₂)/4a (11)

where ρ₁ and ρ₂ is the resistivity of brush and ring, respectively, and a is the average radius of a-spots, indicating that the shrinkage resistance is inversely proportional to the a-spot radius.

Secondly, under the normal load F_n and shear force F_a, the brush spots will be plastically deformed by the ring asperities, hence increasing the contact area, yet decreasing the shrinkage resistance (Fig. 15b). As micro valleys are filled with graphite film, a valley-film resistance (R_vf) is added to the contact resistance (R_c), which can be expressed as [20]:

R _c = R_s + R_vf (12)

$${R_{vf}}={\rho _{vf}}\frac{{{d_{vf}}}}{{\pi {a^2}}}$$

Where ρ_vf and d_vf are the valley-film resistivity and thickness, respectively [34].

Since the resistivity of valley film (0.96 Ω⸱m) is higher than that of the ring (0.0172 Ω⸱m), as shown in Fig. 15c, the formation of valley film will increase the contact resistance as well as the signal distortion (Eq. (3)). In addition, micro valleys are lower than the real contact surface thus their geometry (d_vf and a) will not be affected when filled. Therefore, during each sliding cycle, the signal waveform will be repeatedly distorted by the added film resistance R_vf (Fig. 4).

Finally, as illustrated in Fig. 15d, when valleys are filled with graphite film, the following transferred particles will cover the ring surface. Thereby, in addition to the resistances R_s and R_vf, a new cover-film resistance R_cf is added to the R_c

R _c = R_s + R_vf + R_cf (14)

$${R_{cf}}={\rho _{cf}}\frac{{{d_{{\text{c}}f}}}}{{\pi {a^2}}}$$

Where ρ_vf and d_vf are the cover-film resistivity and thickness, respectively. Results show that the cover film has a great influence on the contact resistance R_c (Fig. 14a). For example, the average R_c increases from 0.23 to 0.68 mΩ (Fig. 14b), meanwhile, the average signal distortion increases from 0.11 to 1.94 (Fig. 6b).

In the later stage, the cover film accelerates the change of CR as well as the signal distortion. As the cover film increases from 10000 to 40000 cycles, according to Eq. 15, the increase of d_cf will result in a linear increase in film resistance R_cf. During 40000–80000 cycles, due to abrasive wear and oxidation wear, not only the film thickness d_cf but also the content of oxide Cu₂O increases. The oxide layer can deteriorate the electrical contact and impede current flow [39], thus increasing the resistivity ρ_vf. Under the dual action of film thickening and oxide increasing, the signal waveform is rapidly distorted. Moreover, as illustrated in Fig. 15d, due to the frictional force, the frictional film moves along the frictional direction (Fig. 10). As a result, the distorted signal angle also increases with the sliding cycle (Fig. 5).

This paper presents a new tribotester for the evaluation of friction and wear effects on the signal transmission of brush-ring system. Owing to the input signal is synchronized with the ring rotation, the relationship between tribological behavior and signal waveform distortion can be analyzed in the same time domain. The main findings can be summarized as follows:

1) From a frictional point of view, the life time of a brush-ring system can be divided into three stages, such as the ascending stage, descending stage, and steady stage. The main wear mechanisms are ploughing wear, adhesive wear, and abrasive wear mixed with oxidative wear.

2) For the brush-ring system, the initially signal waveform is clean, yet later the distortion gradually increases with the sliding cycle. During the sliding friction, the fluctuation of friction coefficient is different from the signal waveform. The former is random fluctuation, yet the latter is inherited fluctuation due to cumulative effect. So, there is no direct relationship between them.

3) During the friction process, with the increase of friction film, the signal waveform is gradually degraded, which causing the signal distorts from sine waveform to triangular waveform.

4) From an electrical point of view, the total contact resistance of brush-ring system is mainly composed of shrinkage resistance, valley-film resistance, and cover-film resistance. During the sliding process, the contact resistance is firstly reduced by the shrinkage resistance, and then increased with the valley-film resistance, finally is affected by the cover-film resistance.

5) For the brush-ring system, the signal transmission is mainly affected by the friction film, such as the signal distortion increases with the film and the signal angle shifts as the film moving along the sliding direction.

PT Power transmission

ST Signal transmission

DC Direct current

CT Computed tomography

PC Personal computer

FC Friction coefficient

FT Friction temperature

CR Contact resistance

EM Electron microscope

OM Optical microstructure

SEM Scanning electron microscope

EDS Energy dispersive spectroscopy

XPS X-ray photoelectron spectroscope

Acknowledgments

This work was supported by the Open Fund of Hubei Engineering Research Center for Graphite Additive Manufacturing Technology and Equipment (HRCGAM202105), the Research Foundation of Hubei Key Laboratory of Hydroelectric Machinery Design & Maintenance (2019KJX03), and the Open Fund of Shanghai Key Laboratory of Materials Laser Processing and Modification (MPLM2018-2).

Liu, Z., Song, C., Li, J., Hou, X., Wang, L., Zhang, Y. Effect of Rotation Speed on Electric Damage of Copper Tribo/Electric Rolling Pairs under Lubricated Single-Point Contact. MATERIALS TRANSACTIONS 61:111-118 (2020).
Kalin, M., Poljanec, D. Influence of the contact parameters and several graphite materials on the tribological behaviour of graphite/copper two-disc electrical contacts. Tribology International 126:192-205 (2018).
Cho, K.H., Hong, U.S., Kang, S.L., Jang, H. Tribological Properties and Electrical Signal Transmission of Copper–Graphite Composites. Tribology Letters 27:301-306 (2007).
Ho, Jang, Hyung, Kyu, Park. Effect of the Friction Characteristics of Sliding Contacts on Electrical Signal Transmission. KSTLE International Journal 2:22-28 (2001).
Feng, D., Liang, B., Zhong, F., Zhou, Y. Development of sliding contact resistance testing equipment for carbon brush confluence ring. IOP Conference Series: Materials Science and Engineering 1207:012015 (2021).
Chevallier, E., Bourny, V., Bouzerar, R., Fortin, J., Durand-Drouhin, O., Da Ros, V. Voltage noise across a metal/metal sliding contact as a probe of the surface state. Journal of Applied Physics 115:154903 (2014).
Fu, Y., Qin, H., Xu, X., Zhang, X., Guo, Z. What are the Progresses and Challenges, from the Electrical Properties of Current-Carrying Friction System to Tribological Performance, for a Stable Current-Carrying Interface? Journal of Bio- and Tribo-Corrosion 8:4 (2021).
Wu, T., Buesink, F., Canavero, F. Overview of Signal Integrity and EMC Design Technologies on PCB: Fundamentals and Latest Progress. IEEE Transactions on Electromagnetic Compatibility 55:624-638 (2013).
Santoro, C., Hayes, R., Herman, J.: Brushless Slip Ring for High Power Transmission. AIAA SPACE 2009 Conference & Exposition. Place
Holzapfel, C., Heinbuch, P., Holl, S.: Sliding Electrical Contacts: Wear and Electrical Performance of Noble Metal Contacts. 2010 Proceedings of the 56th IEEE Holm Conference on Electrical Contacts, pp. 1-8. Place (2010)
Holzapfel, C.: Wear and electrical properties of slip rings. 26th International Conference on Electrical Contacts (ICEC 2012), pp. 86-90. Place (2012)
Holzapfel, C.: Selected Aspects of the Electrical Behavior in Sliding Electrical Contacts. 2011 IEEE 57th Holm Conference on Electrical Contacts (Holm), pp. 1-9. Place (2011)
Holzapfel, C.: Time Series Analysis in the Study of Sliding Electrical Contacts. 2012 IEEE 58th Holm Conference on Electrical Contacts (Holm), pp. 1-7. Place (2012)
Dorsey, G.F., Coleman, D.S., Witherspoon, B.K.: High Speed Data across Sliding Electrical Contacts. 2012 IEEE 58th Holm Conference on Electrical Contacts (Holm), pp. 1-12. Place (2012)
WILLIAMS, W.: Strain sensing technology for high temperature applications. AlAA 4th International Aerospace Planes Conference. Place
BAZERGUI, A. On the initial balance of Wheatstone bridges. Strain 7:21-24 (1971).
Chen, H.-c., Cao, Z.-q., Chen, G.-y., Wang, B., Gong, C.-y., Zhang, Z.-x., et al. Design and construction of a novel instrument for high-frequency micro-force electrical sliding friction testing. Review of Scientific Instruments 90:095103 (2019).
Bartman, J., Kwiatkowski, B. The influence of measurement methodology on the accuracy of electrical waveform distortion analysis. Measurement science review 18:72-78 (2018).
Liskiewicz, T.W., Kubiak, K.J., Mann, D.L., Mathia, T.G. Analysis of surface roughness morphology with TRIZ methodology in automotive electrical contacts: Design against third body fretting-corrosion. Tribology International 143:106019 (2020).
Braunovic, M., Konchits, V.V., Myshkin, N.K.: Electrical contacts: fundamentals, applications and technology. CRC press, (2017)
Zhang, Q., Huang, L., Kang, S., Yin, C., Ma, Z., Cui, L., et al. CuO/Cu2O nanowire arrays grafted by reduced graphene oxide: synthesis, characterization, and application in photocatalytic reduction of CO2. RSC Advances 7:43642-43647 (2017).
Blyth, R.I.R., Buqa, H., Netzer, F.P., Ramsey, M.G., Besenhard, J.O., Golob, P., et al. XPS studies of graphite electrode materials for lithium ion batteries. Applied Surface Science 167:99-106 (2000).
Vasquez, R.P. Cu2O by XPS. Surface Science Spectra 5:257-261 (1998).
Zhu, C., Osherov, A., Panzer, M.J. Surface chemistry of electrodeposited Cu2O films studied by XPS. Electrochimica Acta 111:771-778 (2013).
Li, H., Lei, Y., Huang, Y., Fang, Y., Xu, Y., Zhu, L., et al. Photocatalytic reduction of carbon dioxide to methanol by Cu2O/SiC nanocrystallite under visible light irradiation. Journal of Natural Gas Chemistry 20:145-150 (2011).
Lin, F., Shao, Z., Li, P., Chen, Z., Liu, X., Li, M., et al. Low-cost dual cocatalysts BiVO4 for highly efficient visible photocatalytic oxidation. RSC Advances 7:15053-15059 (2017).
Panzner, G., Egert, B., Schmidt, H.P. The stability of CuO and Cu2O surfaces during argon sputtering studied by XPS and AES. Surface Science 151:400-408 (1985).
Bowden, F., Moore, A., Tabor, D. The ploughing and adhesion of sliding metals. Journal of Applied Physics 14:80-91 (1943).
van Kuilenburg, J., Masen, M.A., van der Heide, E. The role of the skin microrelief in the contact behaviour of human skin: Contact between the human finger and regular surface textures. Tribology International 65:81-90 (2013).
Chang, L., Zhang, Z., Ye, L., Friedrich, K. Tribological properties of high temperature resistant polymer composites with fine particles. Tribology International 40:1170-1178 (2007).
Zhang, H.-J., Zhang, Z.-Z., Guo, F. Studies of the Influence of Graphite and MoS2 on the Tribological Behaviors of Hybrid PTFE/Nomex Fabric Composite. Tribology Transactions 54:417-423 (2011).
Suresha, B., Shiva Kumar, K., Seetharamu, S., Sampath Kumaran, P. Friction and dry sliding wear behavior of carbon and glass fabric reinforced vinyl ester composites. Tribology International 43:602-609 (2010).
Hou, B., Guo, H., Zhang, N., Zhi, Q., Wang, B., Yang, J. Anisotropic friction behavior of aligned and oriented graphite flakes/copper composite. Carbon 186:64-74 (2022).
Slade, P.G.: Electrical contacts: principles and applications. CRC press, (2017)
Ma, G.-z., Xu, B.-s., Wang, H.-d., Si, H.-j., Yang, D.-x. Effect of surface nanocrystallization on the tribological properties of 1Cr18Ni9Ti stainless steel. Materials Letters 65:1268-1271 (2011).
Grandin, M., Wiklund, U. Wear and electrical performance of a slip-ring system with silver–graphite in continuous sliding against PVD coated wires. Wear 348-349:138-147 (2016).
Zhai, C., Hanaor, D., Proust, G., Brassart, L., Gan, Y. Interfacial electro-mechanical behaviour at rough surfaces. Extreme Mechanics Letters 9:422-429 (2016).
Holm, R.: Electric contacts: theory and application. Springer Science & Business Media, New York (1967)
Kogut, L., Komvopoulos, K. Electrical contact resistance theory for conductive rough surfaces separated by a thin insulating film. Journal of Applied Physics 95:576-585 (2004).

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

The effect of friction and wear behavior on the signal transmission of brush-ring system during electrical contact sliding

Status:

Version 1

Abstract

Figures

1. Introduction

2. Experimental Details

2.1 Testing devise

2.2 Signal generation principle

2.3 Signal waveform distortion

2.4 Testing materials

2.5 Testing analysis method

3. Results

3.1 Evolution of signal waveform distortion

3.2 Evolution of friction and wear behavior

3.2.1 Friction coefficient

3.2.2 Friction temperature

3.2.3 Morphology of worn surface

3.2.4 Roughness of worn surface

3.2.5 Friction film

3.2.6 Contact resistance

4. Discussion

4.1 Friction and wear behavior of brush-ring system

4.2 Friction and wear behavior on signal waveform distortion

5. Conclusions

Abbreviations

Declarations

Acknowledgments

References

Additional Declarations

Status:

Version 1