- Email: [email protected]
Study on micro-topographical removals of diamond grain and metal bond in dry electro-contact discharge dressing of coarse diamond grinding wheel
International Journal of Machine Tools & Manufacture 88 (2015) 118–130
Contents lists available at ScienceDirect
International Journal of Machine Tools & Manufacture journal homepage: www.elsevier.com/locate/ijmactool
Study on micro-topographical removals of diamond grain and metal bond in dry electro-contact discharge dressing of coarse diamond grinding wheel Y.J. Lu, J. Xie n, X.H. Si School of Mechanical and Automotive Engineering, South China University of Technology, Guangzhou 510640, China
art ic l e i nf o
a b s t r a c t
Article history: Received 3 May 2014 Received in revised form 22 September 2014 Accepted 23 September 2014
A coarse diamond grinding wheel is able to perform smooth surface grinding with high and rigid grain protrusion, but it is very difficult to dress it. Hence, the dry electro-contact discharge (ECD) is proposed to dress #46 diamond grinding wheel for dry grinding of carbide alloy. The objective is to understand micro-topographical removals of diamond grain and metal bond for self-optimizing dressing. First, the pulse power and direct-current (DC) power were employed to perform dry ECD dressing in contrast to mechanical dressing; then the micro-topographies of diamond grains and metal bond were recognized and extracted from measured wheel surface, respectively; finally, the relationship between impulse discharge parameters and micro-topographical removals was investigated with regard to grain cutting parameters, dry grinding temperature and ground surface. It is shown that the dry ECD dressing along with spark discharge removal may enhance the dressing efficiency by about 10 times and dressing ratio by about 34 times against the mechanical dressing along with cutting removal. It averagely increases grain protrusion height by 12% and grain top angle by 23%, leading to a decrease 37% in grinding temperature and a decrease 46% in surface roughness. Compared with the DC-25V power along with arc discharges, the Pulse-25V power removes the metal bond at 0.241 mm3/min by utilizing discharge energy by 73% and diamond grain at 0.013 mm3/min through surface graphitization, respectively, leading to high and uniform grain protrusion. It is confirmed that the impulse discharge parameters are likely to control the microscopic grain protrusion topography for efficient dressing according to their relations to the micro-removal of metal bond. & Elsevier Ltd. All rights reserved.
Keywords: Grinding Diamond grinding wheel Electro-contact discharge dressing Grain protrusion topography
1. Introduction The hard and brittle materials such as ceramic, silicon, optical glass and carbide alloy have been widely applied in the components of automotive manufacturing, electronics, aerospace and precision optics, etc. It is known that a superabrasive diamond grinding wheel was an alternative to the machining of hard and brittle materials. Generally, superfine diamond grinding wheels were employed to perform a smooth surface grinding [1,2], but they led to poor grinding efficiency due to the limited height of grain protrusion and the rapid fall-off of superfine grains. It was reported that the critical cutting depth of optical glass was enhanced from 60 nm to 160 nm when the grain rake angle was changed from 35° to 90° using a single-point diamond tool of #46 grain [3]. In #60 coarse diamond grinding of Si wafer, the ductile-mode cutting was conducted by the cutting edge of n
Corresponding author. E-mail address: [email protected] (J. Xie).
http://dx.doi.org/10.1016/j.ijmachtools.2014.09.008 0890-6955/& Elsevier Ltd. All rights reserved.
single diamond grain when the cutting depth was less than 73 nm [4]. This means that the ductile-mode grinding may be performed by controlling coarse diamond grain topography along with suitable grinding conditions. Hence, a coarser diamond grinding wheel had been considered to improve both grinding efficiency and surface quality. For example, a #140 diamond grinding wheel was able to perform a super-smooth axial-feed grinding [5]. Furthermore, the 150-μm diamond grinding of BK7 optical glass may make the ground surface roughness be less than 50 nm [6]. A truncated #60 coarse diamond grinding wheel with large grain top angle and valid grain number may realize a mirror finish grinding of SiC ceramics in a macroscopic size [7]. However, it is very difficult to dress the metal-bonded coarse diamond grinding wheel. In the recent decades, many non-conventional dressing technologies, such as ELID (electrolytic inprocess dressing) [8], WEDD (wire electrical discharge dressing) [9] and laser dressing [10], have been used to dress metal-bonded diamond grinding wheel. However, they needed complex control system, expensive equipment and generated pollution emissions.
Y.J. Lu et al. / International Journal of Machine Tools & Manufacture 88 (2015) 118–130
Sm Sxz
Nomenclature a B D de dg E Ed hbr he hg hgr hpulse hwheel hy Id Ie me n N ng Npulse nspark Ra Sg
depth of cut (μm) the width of diamond wheel (mm) the diameter of diamond wheel (mm) impulse removal diameter (μm) nominal diameter of grain (μm) open-circuit voltage (V) discharge voltage (V) bond removal height (μm) impulse removal height (μm) grain protrusion height (μm) grain removal height (μm) the predicted removal height of metal bond (μm) wheel removal height (mm) removal height of grain micro-topography (μm) discharge current (A) impulse peak current (A) impulse removal volume (mm3) maximum discharge frequency (s 1) the grid point number of grain micro-topography grain number the number of impulse dischargecraters spark discharge frequency (s 1) surface roughness (μm) grain area (μm2)
Hence, the dry electro-contact discharge (ECD) dressing was developed along with ecological pollution-free and easy operation [11,12], but it was not applied to coarse diamond grinding wheel. It was reported that increasing valid grain number, negative grain rake angle and grain top angle can improve the ground surface quality [7,13], but there is no way to monitor these microscopic grain cutting parameters during dressing. Although the impulse discharge parameters were related to impulse removal crater sizes of metal bond for valid dressing [14], the grain cutting parameters were not investigated. Moreover, the grain protrusion topography can be measured and evaluated [7], but it has not yet been clear how the micro-topographies of diamond grain and metal bond are removed during dressing.
T td Th Ton Ts Tsύ VE vf Vgr Vpulse Vtbr Vtgr vw Vwheel Wd We
αg γ γg εg η τe ω
119
metal bond area (μm2) the projected area of grain topography on XOZ plane (μm2) dressing time (min) discharge time (μs) hole temperature (°C) actual contact discharge time (min) grinding temperature (°C) surface temperature (°C) dresser/electrode removal volume (mm3) feed speed of wheel (mm/min) grain removal volume (mm3) the predicted removal volume of metal bond (mm3) total bond removal volume (mm3) total grain removal volume (mm3) wheel speed (m/s) wheel removal volume (mm3) discharge energy rate (J/s) impulse discharge energy (J) grain top angle (°) dressing ratio grain rake angle (°) grain relief angle (°) dressing efficiency (mm3/min) impulse discharge duration (μs) effective spark discharge ratio
It was found that the grain surface graphitization (temperature 4700 °C) occurred under the concentrated electrical discharge [15]. Through Raman spectra testing, the diamond–graphite phase transformation was produced during the micro-EDM of PCD [16]. A graphite layer thickness of about 0.5 μm on fine diamond grain surface after wire electrical discharge dressing was measured by HRTEM analysis [17]. However, the actual micro-removal of diamond grain has not yet been quantitatively reported. In this paper, the dry ECD dressing of metal-bonded #46 coarse diamond grinding wheel is proposed using pulse and directcurrent power supplies, respectively. The dressed wheel surface topography was characterized using measured 3D information by white light interferometer (WLI). The objective is to understand
Fig. 1. Dry ECD dressing experimental setup: (a) dressing scene and (b) on-line discharge waveforms monitoring.
120
Y.J. Lu et al. / International Journal of Machine Tools & Manufacture 88 (2015) 118–130
how the micro-topographies of diamond grain and metal bond are removed at the same time in dry ECD dressing. First, the diamond grains and metal bond were extracted from measured microtopography of dressed wheel surface; then the relationship between impulse discharge parameters and micro-topographical removals was investigated; finally, the dressed coarse diamond grinding wheel was employed to perform the axial-feed dry grinding of carbide alloy with regard to the grinding temperature and ground surface roughness.
2. Dry ECD dressing and grinding experiments
Table 2 The dry ECD dressing conditions. CNC grinder
SMART B818
#46 (grain size:350 μm), metal-bonded Diameter D ¼150 mm, width B¼ 2.5 mm Electrode Copper, size: L B0 H ¼21.4 6 48.5 mm3 CNC tool path Curve interpolation path Dressing parameters vw ¼19.6 m/s, vf ¼150 mm/min, a¼ 30 μm Discharge conditions Pulse discharge E¼ 25 V (Pluseon ¼ 100 μs, Pluseoff ¼ 100 μs) DC discharge E ¼25 V and 32 V Coolant In air Diamond wheel
2.1. Dry ECD dressing Fig. 1 shows the dry electro-contact discharge (ECD) dressing experimental setup of #46 diamond grinding wheel. In experiments, a detachable diamond segment was installed on grinding wheel (see Fig. 1a). In order to ensure the diamond wheel positioning without any changes, the diamond segment was unloaded from machine after each dressing operation to measure dressed wheel surface topography by SEM (scanning electron microscope, FEI Quanta 200) and WLI (white light interferometer, BMT SMS Expert 3D). In dry ECD dressing, the pulse and directcurrent (DC) power supplies were employed along with diamond grinding wheel (anode) and copper electrode (cathode). The rotary diamond grinding wheel was driven to grind electrode along curve interpolation path in CNC grinder. When the grinding wheel cut copper electrode, the electric spark discharge occurred. The impulse discharge waveforms were on-line collected by voltage and current sensors along with a digital oscilloscope (RIGOL DS1052E) (see Fig. 1b). Before dressing experiments, the dry ECD truing along with DC open-circuit voltage E of 60 V was performed to ensure the concentricity of initial grinding wheel. In dry ECD dressing, the pulse open-circuit voltage E of 25 V, DC open-circuit voltage E of 25 V and DC open-circuit voltage E of 32 V was called Pusle-25V, DC-25V and DC-32V, respectively. The mechanical dressing, dry ECD dressing (Pusle-25V), dry ECD dressing (DC-25V) and dry ECD dressing (DC-32V) were performed in sequence. The mechanical dressing and dry ECD dressing conditions including wheel speed vw, feed speed of wheel vf, depth of cut a, open-circuit voltage E and so on are shown in Tables 1 and 2, respectively. 2.2. Dry grinding of carbide alloy Fig. 2 shows the dry grinding of carbide alloy and measurement of grinding temperature. Before grinding, the thermocouple 1 and thermocouple 2 were installed on workpiece surface and in the hole below the surface of 0.5 mm along with a simulated heat source of electric iron to measure surface temperature and hole temperature [18], respectively (see Fig. 2a). The temperature was acquired by USB-4718 temperature acquisition module. It is seen that the surface temperature Tsύ increased in polynomial form with hole temperature Th. Moreover, they had very good correlation due Table 1 The mechanical dressing conditions. CNC grinder
SMART B818
#46 (grain size:350 μm), metal-bonded Diameter D ¼150 mm, width B¼ 2.5 mm Dresser GC#120, size:100 6 60 mm3 CNC tool path Straight reciprocating motion Dressing parameters vw ¼23.5 m/s, vf ¼600 mm/min, a¼ 100 μm, Σa ¼30 mm Coolant Water Diamond wheel
Fig. 2. Temperature measurement and dry grinding of carbide alloy: (a) surface temperature Tsύ versus hole temperature Th and (b) dry grinding scene.
to the correlation coefficient R2 of 0.998. Hence, according to the fitted polynomial formula, the measured hole temperature Th may be used to predict the surface temperature Tsύ in grinding. The maximum surface temperature was defined as grinding temperature Ts in this paper. In grinding experiments, an axial-feed dry grinding of carbide alloy was employed along with the on-machine dressing of #46 diamond grinding wheel (see Fig. 2b). Moreover, an orthogonal experimental method was used to investigate the influences of grinding variables on micro-ground surface roughness. The surface
Y.J. Lu et al. / International Journal of Machine Tools & Manufacture 88 (2015) 118–130
Table 3 The dry grinding conditions of carbide alloy. CNC grinder
SMART B818
Diamond wheel
#46 (grain size:350 μm), metal-bonded Diameter D ¼150 mm, width B¼ 2.5 mm Carbide alloy Axial-feed interpolation vw ¼11.8, 15.7, 19.6 m/s vf ¼ 10, 30, 50 mm/min, a¼ 1, 5, 10, and 15 μm In air
Workpiece CNC tool path Grinding parameters Coolant
roughness was measured by TR200 roughness meter. The surface roughness Ra was the mean value of five measured data. The detailed dry grinding conditions including axial-feed speed vf, wheel speed vw, depth of cut a and so on are shown in Table 3.
3. Dry ECD dressing mechanism Fig. 3 shows dry ECD dressing mechanism. In contrast to mechanical dressing, the dry ECD dressing utilized an impulse electro-discharge to remove the metal bond of grinding wheel without any mechanical removal (see Fig. 3a). When the nonconductive diamond grain cut the copper electrode, the rolled-up copper chip approached metal bond to some extent so that the spark discharge occurred. Thus, the metal bond was melted and thrown into a discharge crater [14]. Gradually, the diamond grains were protruded from wheel working surface. Moreover, the impulse discharge heat flow also shocked diamond grain edge. It is seen that most Cu chips were formed by mechanical cutting according to the cut trace on chip surface (see Fig. 3b). Some melted Cu chips by electrical discharge were formed in dry ECD dressing (DC-60V). This is because arc discharges derived from short circuit occurred when the DC open-circuit voltage was larger than 20 V [14]. This would lead to large discharge heat rather than spark discharge crater on metal bond surface, leading to the formation of melted Cu chip. In order to observe the removal of diamond grain before and after dry ECD dressing, the Raman spectra testing was performed. It is found that there were two peaks at 1580 cm 1 (G band) and 1350 cm 1 (D band) on diamond grain surface (see Fig. 3c). It corresponded to the first and second order Raman peak of graphite, respectively [16,19]. Because the characteristic peak of pure diamond is a single sharp peak at 1332 cm 1, the diamond grain top may be carbonized into graphite due to dry electrical discharge. This means that the diamond grain top could be removed by impulse discharge heat and mechanical friction between grain and electrode. It contributes to uniform grain protrusion height in grinding.
121
4. Results and discussions 4.1. Grain protrusion topographies versus dressing conditions Fig. 4 shows the SEM photos of diamond grain protrusion on dressed wheel surface versus dressing conditions. It is shown that the diamond grains were completely protruded and sharpened from metal bond surface under previous three dressing conditions, but Grain 2 was fractured and Grains 1, 3 and 4 were pulled out in dry ECD dressing (DC-32V) (see Fig. 4d). This is because large discharge energy removed so much metal bond that diamond grains were pulled out. It is also found that the dry ECD dressing (Pulse-25V) was able to remove the melted clusters derived from arc discharges in contrast to mechanical dressing (see Fig. 4a and b). Moreover, dry ECD dressings (DC-25V and DC-32V) still produced micro melted clusters on the metal bond surface (see Fig. 4c and d). The chemical elements of Points A, B, C and D on dry ECD dressed wheel surface in Fig. 4 are shown in Table 4. They were measured by energy dispersive X-Ray spectroscopy (EDX, INCAP FET-X3). It is shown that the component of Point A at Grain 1 was mainly composed of diamond (91.70% C) and Oxygen (7.93% O) (see Fig. 4b). This means that diamond grain surface was oxidized in dry ECD dressing. However, the components of Cu and Sn appeared at Point B on grain bottom (see Fig. 4b), thus molten metal bond was adhered to the border of diamond grain. It is also found that the Cu component appeared at Points C and D at cracked Grain 2 (see Fig. 4d). It indicates that the cracked diamond grain surface was covered with melted Cu electrode materials.
4.2. Recognition and extraction of diamond grain and metal bond Fig. 5 shows 3D topographies of dressed diamond wheel surface corresponding to Fig. 4. It is shown that the measured grain topographies were identical to the SEM observations (see Fig. 4). This means that these reconstructed topographies may display 3D information of grain protrusions. In order to extract diamond grain from metal bond surface, the 3D topographies (see Fig. 5) were dispersed to 3D coordinate point cloud O(x, y, z) (see Fig. 6a); then ћ
its normal vector n (x n, yn , zn) was calculated (see Fig. 6b); finally, the diamond grain boundaries were recognized through setting the threshold s. The micro-topographical extraction of diamond grains from metal bond on wheel working surface is shown in Fig. 6. When the threshold value s was set as 0.8, the most integrated diamond grains boundaries may be extracted. They are described as follows:
Fig. 3. Dry ECD dressing mechanism: (a) micro-removal scheme, (b) SEM photo of Cu chip and (c) Raman spectra of diamond grain surface.
122
Y.J. Lu et al. / International Journal of Machine Tools & Manufacture 88 (2015) 118–130
Fig. 4. SEM photos of diamond grain protrusion versus dressing conditions: (a) mechanical dressing, (b) dry ECD dressing (Pulse-25V), (c) dry ECD dressing (DC-25V) and (d) dry ECD dressing (DC-32V). Table 4 The chemical elements on dry ECD dressed wheel surface in Fig. 4. Testing points
A B C D
Element (%) Cu
C
O
Fe
Sn
0.37 69.71 100.00 22.95
91.70 22.97 0.00 59.94
7.93 1.37 0.00 15.21
0.00 3.89 0.00 1.90
0.00 2.06 0.00 0.00
⎛№ ћ O № O № O⎞ n = (x n , yn , z n) = ⎜ , , ⎟ x № y № z⎠ ⎝№
(1)
0 < z n < ƾ = 0.8
(2)
According to the extracted grain boundary outline, the microtopographies of diamond grain and metal bond were recognized and extracted (see Fig. 6b and c).
topographies of measured wheel surfaces [20,21]. This is because the diamond grain shape was not changed even if their measuring coordinate frames were not identical. Fig. 7 shows the matched point clouds and micro-topographical removal of diamond grain before and after dry ECD dressing. It is shown that the extracted diamond grains before and after dressing may be well matched each other by ICP matching (see Fig. 7 left). This means that the matched grain topographies may be used to calculate the height difference between two measured grain topographies. It is also found that the Pulse-25V power produced more diamond grain removal than DC-25V power (see Fig. 7 right). Their maximum removal heights of diamond grain were 9.9 μm and 9.6 μm, respectively. The reason is that the grain top was carbonized into graphite and then removed by impulse discharge heat and shock (see Fig. 3). The mean grain removal height is called grain removal height hgr, which can be calculated as follows:
1 N
N
Å
4.3. Micro-topographical removals of diamond grain and metal bond
hgr =
In order to compare the micro-topographical heights of diamond grain and metal bond before and after dressing, the Iterative Closest Point (ICP) algorithm was employed to match 3D grain
where hy is the removal height of grain micro-topography for a grid on a XOZ plane. N is the grid point number of grain microtopography. After the matched grain topographies were extracted
i=1
hy
(3)
Y.J. Lu et al. / International Journal of Machine Tools & Manufacture 88 (2015) 118–130
123
Fig. 5. 3D topographies of dressed diamond wheel surface: (a) mechanical dressing, (b) dry ECD dressing (Pulse-25V), (c) dry ECD dressing (DC-25V) and dry ECD dressing (DC-32V).
Fig. 6. The micro-topographical extraction of diamond grains from metal bond surface: (a) measured wheel topography, (b) diamond grain topographies and (c) metal bond topography.
from a metal bond surface, the mean metal bond removal height, called bond removal height hbr, can be achieved. Hence, the grain removal volume Vgr can be calculated by
Vgr = hgr Sxz
(4)
where Sxz is the projected area of grain topography on the XOZ plane. Because the concentration of diamond abrasive was 100%, the volumetric proportions of diamond grain and metal bond were 25% and 75%, respectively. Hence, grain area Sg and metal bond area Sm on cylindrical wheel surface may be approximately calculated as follows:
Sg ↣
ưDB 4
(5)
3ưDB 4
(6)
Sm ↣
where D and B are the diameter and width of diamond wheel, respectively (see Table 2). On wheel surface, the grain was
regarded as a circle, thus grain number ng can be given by
ng =
4Sg ưdg 2
(7)
where dg ¼ 350 μm is the nominal diameter of #46 diamond grain. Hence, total grain removal volume Vtgr and total bond removal volume Vtbr can be calculated as follows:
Vtgr = ng Vgr
(8)
Vtbr = Smhbr
(9)
Fig. 8 shows the micro-topographical removal heights and volumes of diamond grain top and metal bond surface in dry ECD dressing. It is shown that the Pulse-25V power increased the grain removal height hgr by about 70% and bond removal height hbr by about 155 times against DC-25V power (see Fig. 8a). The former is because it easily removed the diamond grain top without any crack by surface graphitization derived from impulse discharge heat, shock and mechanical friction. The latter is because it produced little melted clusters on metal bond (see Fig. 4c).
124
Y.J. Lu et al. / International Journal of Machine Tools & Manufacture 88 (2015) 118–130
Fig. 7. The matched point clouds and micro-topographical removal of diamond grain before and after dressing: (a) dry ECD dressing (Pulse-25V) and (b) dry ECD dressing (DC-25V).
According to the discharge dressing time T ¼80 min, the Pulse25V power removed diamond grains at 0.013 mm3/min and metal bond at 0.241 mm3/min, which were larger than the ones for DC25V power. Moreover, the total bond removal volume Vtbr was about as much 18 times as total grain removal volume Vtgr for Pulse-25V power (see Fig. 8b). As a result, the proposed microtopographical removal mode of metal bond and diamond grain may be used to describe the grain protrusion mechanism during dressing. 4.4. Grain cutting parameters versus dressing conditions Fig. 9 shows the grain protrusion profile characterized with grain cutting parameters. The grain top was regarded as the highest point of 3D grain profile. Along the cutting direction at grain top, grain protrusion height hg, gain rake angle γg, grain relief angle εg and grain top angle αg were defined as the grain cutting parameters. The grain top angle αg was defined along the vertical cutting direction. It is shown that integrated diamond grains along with sharpened cutting edges were protruded from wheel working surface after dressing. Fig. 10 shows grain cutting parameters versus grain protrusion height hg. It is shown that the dry ECD dressing averagely increased the grain protrusion height hg by about 12% against mechanical dressing. This is because it removed the melted cluster on metal bond surface (see Fig. 4a and b). In dry ECD dressing, the Pulse-25V power increased the hg by about 6% against DC-25V power. It also decreased the height difference of grain protrusion from 6 μm to 3 μm for highest Grains 1 and 4, leading to more uniform grain protrusion. The reason is that it did not produce such micro-melted clusters on metal bond surface as DC-25V power (see Fig. 4c), and it increased the removal of diamond grain top (see Fig. 8). It is also found that an increase in grain protrusion height hg led to an increase trend in negative grain rake angle γg (see Fig. 10a) and a decrease trend in grain relief angle εg (see Fig. 10b). This is
because top grains tips were easily worn and obtuse during dressing. Moreover, the dry ECD dressing averagely increased grain top angle αg by about 23% compared with mechanical dressing (see Fig. 10c). Moreover, it greatly decreased negative γg and increased εg for Grain 3. The reason is that impulse discharge heat and shock force mainly influenced the grains with medium protrusion height (see Fig. 3a). After dry ECD dressing, the mean grain protrusion height reached to 178 μm, which was about 50% of #46 diamond grain size. However, it was only about 33% and 18% for #60 and #180 diamond grinding wheel in mechanical dressing, respectively [4,22]. It contributes to an increase in chip capacity, which may decrease cutting heat to protect from surface damage in grinding. Moreover, the negative grain rake angle ranged from 70° to 80°, which was larger than those of 65° and 50° for #60 and #270 diamond grinding wheels, respectively [4,13]. It also contributes to an increase in critical grain cutting depth transferred from brittlemode removal to ductile-mode removal for mirror grinding [3]. 4.5. The impulse discharge energy parameters in dry ECD dressing Fig. 11 shows the discharge current Id and discharge voltage Ed versus discharge time td in dry ECD dressing. The discharge current trace was characterized as the impulse discharge parameters including impulse peak current Ie and impulse discharge duration τe [14]. It is shown that the Pulse-25V power mainly produced micro-spark discharges, leading to large impulse discharge current Ie and short impulse discharge duration τe (see Fig. 11a). This may explain why the Pulse-25V power may gradually remove the metal bond (see Fig. 4b), leading to large grain protrusion height (see Fig. 8 and 10). However, the DC-25V power produced a majority of micro-arc discharges, thus leading to little impulse discharge current and long impulse discharge duration (see Fig. 11b). The reason is that many micro-melted clusters were formed on metal bond surface (see Fig. 4c). This may also explain that it achieved less grain
Y.J. Lu et al. / International Journal of Machine Tools & Manufacture 88 (2015) 118–130
125
Fig. 8. The micro-topographical removal heights and volumes of diamond grain top and metal bond surface in dry ECD dressing: (a) grain removal height hgr and bond removal height hbr and (b) total grain removal volume Vtgr and total bond removal volume Vtbr.
Fig. 9. The grain protrusion profile characterized with grain cutting parameters including grain protrusion height hg, gain rake angle γg, grain relief angle εg and grain top angle αg.
protrusion height (see Fig. 10). Moreover, the DC-32V power produced much larger arc discharges than DC-25V power (see Fig. 11c). According to the wheel speed vw of 19.6 m/s and impulse discharge duration τe of 178 μs, the moving length of its arc discharge along wheel surface reached up to 3.48 mm, which was about 10 times as much as grain size dg. This means that the arc discharges along with short circuit shocked the protruded diamond grains, leading to their fracture and pull-out (see Fig. 4d). The impulse discharge energy parameters included impulse discharge energy We, maximum discharge frequency n and discharge energy rate Wd. Their relations may be described as follows:
caused micro-spark discharges (see Fig. 11a) and micro-arc discharges (see Fig. 11b), respectively. Hence, the Pulse-25V power increased the discharge frequency n by 20% and discharge energy rate Wd by 22% against DC-25V power, respectively (see Fig. 12b and c). However, the DC-32V power produced too large impulse discharge energy We and discharge energy rate Wd, thus leading to large arc discharges that shocked diamond grains along with high speed rotary wheel. As a result, the Pulse-25V power along with spark discharge was suitable for ECD dressing. It is identical to the critical open-circuit voltage of 20 V transferred from arc discharge to spark discharge [14].
We =
ⅈ0
td
Wd = nWe
IdEdd(td)
(10) (11)
where n was achieved from collected discharge waveforms (see Fig. 11). Fig. 12 shows the impulse discharge energy parameters in dry ECD dressing. It is shown that Pulse-25V power and DC-25V power produced same impulse discharge energy (see Fig. 12a), but they
4.6. Impulse discharge parameters versus micro-removal of metal bond In dry ECD dressing, the impulse spark discharge produced a sphere-shaped discharge crater on the metal bond surface. The impulse discharge removal sizes such as impulse removal height he, impulse removal diameter de and impulse removal volume me were correlated to the impulse discharge parameters (Ie and τe) [14]. They were described as follows [14]:
126
Y.J. Lu et al. / International Journal of Machine Tools & Manufacture 88 (2015) 118–130
Fig. 10. Grain cutting parameters versus grain protrusion height hg: (a) Grain rake angle γg, (b) grain relief angle εg and (c) grain top angle αg.
he = 5.41 ¬ Ie 0.77 ¬ ǀ e 0.19
(12)
de = 29.00 ¬ Ie 0.78 ¬ ǀ e 0.17
(13)
me = ư((de2he )/8 + he 3/6)
(14)
Fig. 13 shows the impulse discharge removal sizes on metal bond surface in dry ECD dressing. The impulse discharge removal sizes were calculated according to Eqs. (12)–(14) along with the impulse discharge parameters (Ie and τe) shown in Fig. 11. It is shown that the Pulse-25V power may produce larger impulse removal height he, impulse removal diameter de and impulse removal volume me than DC-25V power. Moreover, the impulse removal height he was much less than the half of normal grain size dg/2. This means that the pulse power produced larger discharge removal of metal bond than DC power even if their discharge energies were same. Although the DC-32V power produced much larger impulse removal sizes of metal bond, many grains were pulled out from wheel working surface due to arc discharges. In order to predict the cumulative metal bond removal with impulse discharge crater mode, the spark discharge frequency nspark can be given by
nspark = Ǻ n
(15)
where ω is the effective spark discharge ratio. According to the observed discharge waveform results, the ω was regarded as 20%. Thus, the number of impulse discharge craters Npulse was defined by
Npulse = nspark Ton
(16)
where Ton is the actual contact discharge time. According to the wheel trajectories, dressing parameters (vw, vf) and electrode sizes (L, B0) shown in Table 2 in dressing process, Ton was achieved as about 2 min. The predicted removal volume of metal bond Vpulse was calculated by
Vpulse = Npulseme
(17)
According to Eq. (6), the predicted removal height of metal bond hpulse was given by
h pulse =
4Vpulse 3ưDB
(18)
According to Eqs. (17) and (18), Vpulse and hpulse were calculated as 26.4 mm3 and 29.8 μm in the case of ECD dressing (Pulse-25V). It is seen that the predicted values were a little larger than total bond removal volume Vtbr of 19.3 mm3 and bond removal height hbr of 21.8 μm (see Fig. 8). This is because the model of impulse discharge removal sizes (he, de, me) were achieved under static
spark discharges [14]. In actual dressing process, the high speed rotary wheel may take away a part of discharge energy and heat. Moreover, the electrical discharge also occurred between electrode/chip and diamond grain surface covered with metal bond (Cu and Sn), leading to a consumption of discharge energy. Hence, the predicted discharge removal Vpulse and hpulse were a little larger than the metal bond removal Vtbr and hbr, respectively. The discharge energy utilization ratio was about 73% according to the ratio of Vtbr to Vpulse. Hence, the impulse peak current, discharge duration and discharge frequency derived from discharge current trace may be used to monitor microscopic grain protrusion. This also means that the self-optimizing dressing may be employed according to the relationships between impulse discharge parameters and predicted discharge removal of metal bond. 4.7. Dressing ratio and dressing efficiency The dressing parameters included wheel removal height hwheel, wheel removal volume Vwheel, dressing ratio γ and dressing efficiency η. In the WLI measurement of actual dressing parameters, the wheel profile was on-machine replicated on the graphite plate surface positioned grinder after dressing. The difference between wheel profiles before and after dressing was regarded as the wheel removal height hwheel [20]. Hence, the wheel removal volume Vwheel may be calculated by
Vwheel = ưDBh wheel
(19)
In order to evaluate the dressing abilities versus dressing conditions, the dressing ratio γ and dressing efficiency η were defined as follows [12]:
Ɛ=
Vwheel VE
(20)
ƛ=
Vwheel T
(21)
where VE is the dresser/electrode removal volume. T is the dressing time, which was 250 min and 80 min in mechanical dressing and dry ECD dressing (Pulse-25V), respectively. Fig. 14 shows the dressing parameters of diamond wheel for mechanical dressing and dry ECD dressing (Pulse-25V). It is shown that the wheel removal height hwheel of 19.3 μm and wheel removal volume Vwheel of 22.7 mm3 (see Fig. 14a and b) were roughly identical to the bond removal height hbr of 21.8 μm and total bond removal volume Vtbr of 19.3 mm3 (see Fig. 8). This indicates that the proposed micro-topographical removal model of diamond grain and metal bond was valid to evaluate the grain protrusion in dry ECD dressing.
Y.J. Lu et al. / International Journal of Machine Tools & Manufacture 88 (2015) 118–130
127
Fig. 11. Discharge current Id and discharge voltage Ed versus discharge time td in dry ECD dressing conditions: (a) Pulse-25V, (b) DC-25V and (c) DC-32V.
Moreover, the dry ECD dressing increased the dressing ratio γ by about 34 times and the dressing efficiency η by about 10 times against mechanical dressing (see Fig. 14c and d). The reason is that dry ECD dressing may rapidly remove the metal bond by the impulse spark discharges between metal bond and electrode chip, but the mechanical dressing along with mechanical cutting was
limited to the space between metal bond and dresser, which was dominated by coarse diamond size. In dry ECD dressing, the dressing ratio of 0.0133 and the dressing efficiency of 0.284 mm3/min for coarse #46 diamond wheel dressing were much less than the truing ratio of 2.82 and the truing efficiency of 12.25 mm3/min for fine #600 fine diamond
128
Y.J. Lu et al. / International Journal of Machine Tools & Manufacture 88 (2015) 118–130
Fig. 12. The impulse discharge energy parameters in dry ECD dressing: (a) impulse discharge energy We, (b) maximum discharge frequency n and (c) discharge energy rate Wd.
Fig. 13. The impulse discharge removal sizes on metal bond surface in dry ECD dressing: (a) impulse removal height he, (b) impulse removal diameter de and (c) impulse removal volume me.
Fig. 14. The dressing parameters of diamond wheel for mechanical dressing and dry ECD dressing (Pulse-25V): (a) wheel removal height hwheel, (b) wheel removal volume Vwheel, (c) dressing ratio and (d) dressing efficiency η.
wheel [12]. This means it is quite difficult to dress coarser diamond grinding wheel. 4.8. Grinding temperature and ground surface roughness Fig. 15 shows grinding temperature Ts and surface roughness Ra versus dressing conditions and grinding parameters. It is shown that the grinding temperature Ts and the surface roughness Ra decreased with decreasing depth of cut a and feed speed vf and increasing wheel speed vw for three dressing conditions. The Ra changes with feed speed vf and wheel speed vw were identical to the theoretical model of axial-feed grinding [7]. This also means that a decrease in grinding temperature led to a decrease in ground surface roughness. However, it was an exception for the depth of cut larger than 10 μm and feed speed larger than 30 mm/ min. The reason is that large depth of cut and feed speed could cause large friction force and mechanical vibration. In the case of depth of cut a of 1 μm, the dry ECD dressing (Pulse-25V) decreased the grinding temperature Ts by about 37%
and about 10% against mechanical dressing and dry ECD dressing (DC-25V), respectively (see Fig. 15a). Moreover, it decreased the surface roughness Ra by about 46% and about 10%, respectively (see Fig. 15b). The Ra ranged from 73 nm to 89 nm. This is because large grain protrusion height could decrease grinding heat, thus leading to a decrease in grinding temperature and surface roughness. Fig. 16 shows the SEM photos of dry ground surfaces and edges of carbide alloy versus dressing conditions. The dry grinding parameters were a ¼1 μm, vf ¼10 mm/min and vw ¼11.8 m/s. It is shown that the dry ECD dressing produced less micro-scratches on ground surface and less micro-cracks on ground edge than mechanical dressing (see Fig. 16 up). It is identical to the results of grinding temperature and surface roughness. This is because it increased the grain protrusion height, leading to larger chip capacity space. In dry ECD dressing, the Pulse-25V power produced less micron-scale brittle-cracks on ground surface than the DC-25V power, leading to smoother ground surface. It also produced less end edge wear.
Y.J. Lu et al. / International Journal of Machine Tools & Manufacture 88 (2015) 118–130
129
Fig. 15. Dry grinding temperature Ts and surface roughness Ra: (a) grinding temperature Ts and (b) surface roughness Ra.
Fig. 16. SEM photos of dry ground surfaces and edges of carbide alloy: (a) mechanical dressing, (b) dry ECD dressing (Pulse-25V) and (c) dry ECD dressing (DC-25V).
130
Y.J. Lu et al. / International Journal of Machine Tools & Manufacture 88 (2015) 118–130
References
Fig. 17. The macroscopic dry ground surface of carbide alloy using dry ECD dressed #46 coarse diamond grinding wheel.
Fig. 17 shows the macroscopic dry ground surface of carbide alloy. It is shown that the mirror-like surface was achieved in a macroscopic size using the dry ECD dressed #46 coarse diamond wheel. This is because the dry ECD dressing may enhance the uniformity of grain protrusion heights, leading to an increase in valid grain number. The other reason is that the coarse diamond grain produced large rigidity without any pull-out. The surface roughness Ra reached to 0.089 μm, which was much less than the Ra of 0.82 μm that was achieved by #120-140 diamond wheel [23]. As a result, the dry ECD dressed coarse diamond grinding wheel may be used to grind smooth surface macroscopically due to large grain protrusion height and rigidity. 5. Conclusions (1) In dry ECD dressing, the coarse diamond grain is removed at 0.013 mm3/min through the surface graphitization; the metal bond is removed at 0.241 mm3/min through spark discharge removals. The impulse discharge parameters may be used to monitor and optimize the microscopic grain protrusion according to their relations to the micro-removal of metal bond. (2) The Pulse-25V power along with spark discharges produce larger impulse discharge frequency and discharge energy rate than the DC-25V power along with arc discharges even if their discharge energies are same. It increases metal bond removal and diamond grain removal, thus leading to high and uniform grain protrusion. (3) The dry ECD dressing increases dressing ratio by about 34 times and dressing efficiency by about 10 times against mechanical dressing along with cutting removal, respectively. It averagely increases grain protrusion height by 12% and grain top angle by 23%. The mean grain protrusion height of 178 μm is about 50% of actual diamond grain size. (4) In axial-feed dry grinding of carbide alloy with a #46 coarse diamond wheel, the dry ECD dressing may decrease the grinding temperature by about 37% against mechanical dressing, leading to a decrease 46% in surface roughness. The macroscopic smooth surface with the Ra of 73–89 nm may be produced due to large grain protrusion height and rigidity.
Acknowledgments The project was supported by the National Natural Science Foundation of China (Grant no. 61475046).
[1] J. Webster, M. Tricard, Innovations in abrasive products for precision grinding, CIRP Ann.—Manuf. Technol. 53 (2004) 597–617. [2] Z.Y. Zhang, F.W. Huo, Y.Q. Wu, H. Huang, Grinding of silicon wafers using an ultrafine diamond wheel of a hybrid bond material, Int. J. Mach. Tools Manuf. 51 (2011) 18–24. [3] J. Tamaki, G. Sato, K. Maekawa, J. Yan, A. Kubo, Effect of cutting edge shape of diamond abrasive grain on ductile-mode grinding of optical glass-measurement of critical depth of cut by means of single grit diamond cutting, J. Jpn. Soc. Abras. Technol. 49 (2005) 335–340. [4] J. Xie, H.F. Xie, X.R. Liu, T.W. Tan, Dry micro-grooving on Si wafer using a coarse diamond grinding, Int. J. Mach. Tools Manuf. 61 (2012) 1–8. [5] H. Yasui, T. Sawa, Ultra-smoothness grinding of fine ceramics by coarse grain size diamond wheel-effect of grinding fluid, J. Jpn. Soc. Precis. Eng. 70 (2004) 853–859. [6] Q.L. Zhao, J.Y. Chen, H.T. Huang, X.Y. Fang, Grinding damage of BK7 using copper-resin bond coarse-grained diamond wheel, Int. J. Precis. Eng. Manuf. 12 (2011) 5–13. [7] J. Xie, Y.X. Lu, Study on axial-feed mirror finish grinding of hard and brittle materials in relation to micron-scale grain protrusion parameters, Int. J. Mach. Tools Manuf. 51 (2011) 84–93. [8] H.S. Lim, K. Fathima, A.S. Kumar, M. Rahman, A fundamental study on the mechanism of electrolytic in-process dressing (ELID) grinding, Int. J. Mach. Tools Manuf. 42 (2002) 935–943. [9] E. Weingärtner, S. Jaumann, F. Kuster, M. Boccadoro, Special wire guide for onmachine wire electrical discharge dressing of metal bonded grinding wheels, CIRP Ann.—Manuf. Technol. 59 (2010) 227–230. [10] A. Hosokawa, T. Ueda, T. Yunoki, Laser dressing of metal bonded diamond wheel, CIRP Ann.—Manuf. Technol. 55 (2006) 329–332. [11] J. Xie, J. Tamaki, An experimental study on discharge mediums used for electro-contact discharge dressing of metal-bonded diamond grinding wheel, J. Mater. Process. Technol. 208 (2008) 239–244. [12] J. Xie, H.F. Xie, M.J. Luo, T.W. Tan, P. Li, Dry electro-contact discharge mutualwear truing of micro-diamond wheel V-tip for precision micro-grinding, Int. J. Mach. Tools Manuf. 60 (2012) 44–51. [13] J. Xie, F. Wei, J.H. Zheng, J. Tamaki, A. Kubo, 3D laser investigation on micronscale grain protrusion topography of truncated diamond grinding wheel for precision grinding performance, Int. J. Mach. Tools Manuf. 51 (2011) 411–419. [14] J. Xie, J. Tamaki, In-process evaluation of grit protrusion feature for fine diamond grinding wheel by means of electro-contact discharge dressing, J. Mater. Process. Technol. 180 (2006) 83–90. [15] E. Weingärtner, R. Roth, F. Kuster, M. Boccadoro, F. Fiebelkorn, Electrical discharge dressing and its influence on metal bonded diamond wheels, CIRP Ann.—Manuf. Technol. 61 (2012) 183–186. [16] Z.Y. Zhang, H.M. Peng, J.W. Yan, Micro-cutting characteristics of EDM fabricated high-precision polycrystalline diamond tools, Int. J. Mach. Tools Manuf. 65 (2013) 99–106. [17] A. Klink, Wire electro discharge trueing and dressing of fine grinding wheels, CIRP Ann.—Manuf. Technol. 59 (2010) 235–238. [18] J. Xie, M.J. Luo, K.K. Wu, L.F. Yang, D.H. Li, Experimental study on cutting temperature and cutting force in dry turning of titanium alloy using a noncoated micro-grooved tool, Int. J. Mach. Tools Manuf. 73 (2013) 25–36. [19] A.C. Ferrari, Determination of bonding in diamond-like carbon by Raman spectroscopy, Diam. Relat. Mater. 11 (2002) 1053–1061. [20] J. Xie, Y.W. Zhuo, T.W. Tan, Experimental study on fabrication and evaluation of micro-pyramid-structured silicon surface using a V-tip of diamond grinding wheel, Precis. Eng. 35 (2011) 173–182. [21] J. Xie, X.R. Liu, K.K. Wu, Y.J. Lu, P. Li, Evaluation on 3D micro-ground profile accuracy of micro-pyramid-structured Si surface using an adaptive-orientation WLI measurement, Precis. Eng. 37 (2013) 918–923. [22] J. Xie, J. Xu, Y. Tang, J. Tamaki, 3D graphical evaluation of micron-scale protrusion topography of diamond grinding wheel, Int. J. Mach. Tools Manuf. 48 (2008) 1254–1260. [23] Y. Wang, X.J. Zhou, D.J. Hu, An experimental investigation of dry-electrical discharge assisted truing and dressing of metal bonded diamond wheel, Int. J. Mach. Tools Manuf. 46 (2006) 333–342.
Contents lists available at ScienceDirect
International Journal of Machine Tools & Manufacture journal homepage: www.elsevier.com/locate/ijmactool
Study on micro-topographical removals of diamond grain and metal bond in dry electro-contact discharge dressing of coarse diamond grinding wheel Y.J. Lu, J. Xie n, X.H. Si School of Mechanical and Automotive Engineering, South China University of Technology, Guangzhou 510640, China
art ic l e i nf o
a b s t r a c t
Article history: Received 3 May 2014 Received in revised form 22 September 2014 Accepted 23 September 2014
A coarse diamond grinding wheel is able to perform smooth surface grinding with high and rigid grain protrusion, but it is very difficult to dress it. Hence, the dry electro-contact discharge (ECD) is proposed to dress #46 diamond grinding wheel for dry grinding of carbide alloy. The objective is to understand micro-topographical removals of diamond grain and metal bond for self-optimizing dressing. First, the pulse power and direct-current (DC) power were employed to perform dry ECD dressing in contrast to mechanical dressing; then the micro-topographies of diamond grains and metal bond were recognized and extracted from measured wheel surface, respectively; finally, the relationship between impulse discharge parameters and micro-topographical removals was investigated with regard to grain cutting parameters, dry grinding temperature and ground surface. It is shown that the dry ECD dressing along with spark discharge removal may enhance the dressing efficiency by about 10 times and dressing ratio by about 34 times against the mechanical dressing along with cutting removal. It averagely increases grain protrusion height by 12% and grain top angle by 23%, leading to a decrease 37% in grinding temperature and a decrease 46% in surface roughness. Compared with the DC-25V power along with arc discharges, the Pulse-25V power removes the metal bond at 0.241 mm3/min by utilizing discharge energy by 73% and diamond grain at 0.013 mm3/min through surface graphitization, respectively, leading to high and uniform grain protrusion. It is confirmed that the impulse discharge parameters are likely to control the microscopic grain protrusion topography for efficient dressing according to their relations to the micro-removal of metal bond. & Elsevier Ltd. All rights reserved.
Keywords: Grinding Diamond grinding wheel Electro-contact discharge dressing Grain protrusion topography
1. Introduction The hard and brittle materials such as ceramic, silicon, optical glass and carbide alloy have been widely applied in the components of automotive manufacturing, electronics, aerospace and precision optics, etc. It is known that a superabrasive diamond grinding wheel was an alternative to the machining of hard and brittle materials. Generally, superfine diamond grinding wheels were employed to perform a smooth surface grinding [1,2], but they led to poor grinding efficiency due to the limited height of grain protrusion and the rapid fall-off of superfine grains. It was reported that the critical cutting depth of optical glass was enhanced from 60 nm to 160 nm when the grain rake angle was changed from 35° to 90° using a single-point diamond tool of #46 grain [3]. In #60 coarse diamond grinding of Si wafer, the ductile-mode cutting was conducted by the cutting edge of n
Corresponding author. E-mail address: [email protected] (J. Xie).
http://dx.doi.org/10.1016/j.ijmachtools.2014.09.008 0890-6955/& Elsevier Ltd. All rights reserved.
single diamond grain when the cutting depth was less than 73 nm [4]. This means that the ductile-mode grinding may be performed by controlling coarse diamond grain topography along with suitable grinding conditions. Hence, a coarser diamond grinding wheel had been considered to improve both grinding efficiency and surface quality. For example, a #140 diamond grinding wheel was able to perform a super-smooth axial-feed grinding [5]. Furthermore, the 150-μm diamond grinding of BK7 optical glass may make the ground surface roughness be less than 50 nm [6]. A truncated #60 coarse diamond grinding wheel with large grain top angle and valid grain number may realize a mirror finish grinding of SiC ceramics in a macroscopic size [7]. However, it is very difficult to dress the metal-bonded coarse diamond grinding wheel. In the recent decades, many non-conventional dressing technologies, such as ELID (electrolytic inprocess dressing) [8], WEDD (wire electrical discharge dressing) [9] and laser dressing [10], have been used to dress metal-bonded diamond grinding wheel. However, they needed complex control system, expensive equipment and generated pollution emissions.
Y.J. Lu et al. / International Journal of Machine Tools & Manufacture 88 (2015) 118–130
Sm Sxz
Nomenclature a B D de dg E Ed hbr he hg hgr hpulse hwheel hy Id Ie me n N ng Npulse nspark Ra Sg
depth of cut (μm) the width of diamond wheel (mm) the diameter of diamond wheel (mm) impulse removal diameter (μm) nominal diameter of grain (μm) open-circuit voltage (V) discharge voltage (V) bond removal height (μm) impulse removal height (μm) grain protrusion height (μm) grain removal height (μm) the predicted removal height of metal bond (μm) wheel removal height (mm) removal height of grain micro-topography (μm) discharge current (A) impulse peak current (A) impulse removal volume (mm3) maximum discharge frequency (s 1) the grid point number of grain micro-topography grain number the number of impulse dischargecraters spark discharge frequency (s 1) surface roughness (μm) grain area (μm2)
Hence, the dry electro-contact discharge (ECD) dressing was developed along with ecological pollution-free and easy operation [11,12], but it was not applied to coarse diamond grinding wheel. It was reported that increasing valid grain number, negative grain rake angle and grain top angle can improve the ground surface quality [7,13], but there is no way to monitor these microscopic grain cutting parameters during dressing. Although the impulse discharge parameters were related to impulse removal crater sizes of metal bond for valid dressing [14], the grain cutting parameters were not investigated. Moreover, the grain protrusion topography can be measured and evaluated [7], but it has not yet been clear how the micro-topographies of diamond grain and metal bond are removed during dressing.
T td Th Ton Ts Tsύ VE vf Vgr Vpulse Vtbr Vtgr vw Vwheel Wd We
αg γ γg εg η τe ω
119
metal bond area (μm2) the projected area of grain topography on XOZ plane (μm2) dressing time (min) discharge time (μs) hole temperature (°C) actual contact discharge time (min) grinding temperature (°C) surface temperature (°C) dresser/electrode removal volume (mm3) feed speed of wheel (mm/min) grain removal volume (mm3) the predicted removal volume of metal bond (mm3) total bond removal volume (mm3) total grain removal volume (mm3) wheel speed (m/s) wheel removal volume (mm3) discharge energy rate (J/s) impulse discharge energy (J) grain top angle (°) dressing ratio grain rake angle (°) grain relief angle (°) dressing efficiency (mm3/min) impulse discharge duration (μs) effective spark discharge ratio
It was found that the grain surface graphitization (temperature 4700 °C) occurred under the concentrated electrical discharge [15]. Through Raman spectra testing, the diamond–graphite phase transformation was produced during the micro-EDM of PCD [16]. A graphite layer thickness of about 0.5 μm on fine diamond grain surface after wire electrical discharge dressing was measured by HRTEM analysis [17]. However, the actual micro-removal of diamond grain has not yet been quantitatively reported. In this paper, the dry ECD dressing of metal-bonded #46 coarse diamond grinding wheel is proposed using pulse and directcurrent power supplies, respectively. The dressed wheel surface topography was characterized using measured 3D information by white light interferometer (WLI). The objective is to understand
Fig. 1. Dry ECD dressing experimental setup: (a) dressing scene and (b) on-line discharge waveforms monitoring.
120
Y.J. Lu et al. / International Journal of Machine Tools & Manufacture 88 (2015) 118–130
how the micro-topographies of diamond grain and metal bond are removed at the same time in dry ECD dressing. First, the diamond grains and metal bond were extracted from measured microtopography of dressed wheel surface; then the relationship between impulse discharge parameters and micro-topographical removals was investigated; finally, the dressed coarse diamond grinding wheel was employed to perform the axial-feed dry grinding of carbide alloy with regard to the grinding temperature and ground surface roughness.
2. Dry ECD dressing and grinding experiments
Table 2 The dry ECD dressing conditions. CNC grinder
SMART B818
#46 (grain size:350 μm), metal-bonded Diameter D ¼150 mm, width B¼ 2.5 mm Electrode Copper, size: L B0 H ¼21.4 6 48.5 mm3 CNC tool path Curve interpolation path Dressing parameters vw ¼19.6 m/s, vf ¼150 mm/min, a¼ 30 μm Discharge conditions Pulse discharge E¼ 25 V (Pluseon ¼ 100 μs, Pluseoff ¼ 100 μs) DC discharge E ¼25 V and 32 V Coolant In air Diamond wheel
2.1. Dry ECD dressing Fig. 1 shows the dry electro-contact discharge (ECD) dressing experimental setup of #46 diamond grinding wheel. In experiments, a detachable diamond segment was installed on grinding wheel (see Fig. 1a). In order to ensure the diamond wheel positioning without any changes, the diamond segment was unloaded from machine after each dressing operation to measure dressed wheel surface topography by SEM (scanning electron microscope, FEI Quanta 200) and WLI (white light interferometer, BMT SMS Expert 3D). In dry ECD dressing, the pulse and directcurrent (DC) power supplies were employed along with diamond grinding wheel (anode) and copper electrode (cathode). The rotary diamond grinding wheel was driven to grind electrode along curve interpolation path in CNC grinder. When the grinding wheel cut copper electrode, the electric spark discharge occurred. The impulse discharge waveforms were on-line collected by voltage and current sensors along with a digital oscilloscope (RIGOL DS1052E) (see Fig. 1b). Before dressing experiments, the dry ECD truing along with DC open-circuit voltage E of 60 V was performed to ensure the concentricity of initial grinding wheel. In dry ECD dressing, the pulse open-circuit voltage E of 25 V, DC open-circuit voltage E of 25 V and DC open-circuit voltage E of 32 V was called Pusle-25V, DC-25V and DC-32V, respectively. The mechanical dressing, dry ECD dressing (Pusle-25V), dry ECD dressing (DC-25V) and dry ECD dressing (DC-32V) were performed in sequence. The mechanical dressing and dry ECD dressing conditions including wheel speed vw, feed speed of wheel vf, depth of cut a, open-circuit voltage E and so on are shown in Tables 1 and 2, respectively. 2.2. Dry grinding of carbide alloy Fig. 2 shows the dry grinding of carbide alloy and measurement of grinding temperature. Before grinding, the thermocouple 1 and thermocouple 2 were installed on workpiece surface and in the hole below the surface of 0.5 mm along with a simulated heat source of electric iron to measure surface temperature and hole temperature [18], respectively (see Fig. 2a). The temperature was acquired by USB-4718 temperature acquisition module. It is seen that the surface temperature Tsύ increased in polynomial form with hole temperature Th. Moreover, they had very good correlation due Table 1 The mechanical dressing conditions. CNC grinder
SMART B818
#46 (grain size:350 μm), metal-bonded Diameter D ¼150 mm, width B¼ 2.5 mm Dresser GC#120, size:100 6 60 mm3 CNC tool path Straight reciprocating motion Dressing parameters vw ¼23.5 m/s, vf ¼600 mm/min, a¼ 100 μm, Σa ¼30 mm Coolant Water Diamond wheel
Fig. 2. Temperature measurement and dry grinding of carbide alloy: (a) surface temperature Tsύ versus hole temperature Th and (b) dry grinding scene.
to the correlation coefficient R2 of 0.998. Hence, according to the fitted polynomial formula, the measured hole temperature Th may be used to predict the surface temperature Tsύ in grinding. The maximum surface temperature was defined as grinding temperature Ts in this paper. In grinding experiments, an axial-feed dry grinding of carbide alloy was employed along with the on-machine dressing of #46 diamond grinding wheel (see Fig. 2b). Moreover, an orthogonal experimental method was used to investigate the influences of grinding variables on micro-ground surface roughness. The surface
Y.J. Lu et al. / International Journal of Machine Tools & Manufacture 88 (2015) 118–130
Table 3 The dry grinding conditions of carbide alloy. CNC grinder
SMART B818
Diamond wheel
#46 (grain size:350 μm), metal-bonded Diameter D ¼150 mm, width B¼ 2.5 mm Carbide alloy Axial-feed interpolation vw ¼11.8, 15.7, 19.6 m/s vf ¼ 10, 30, 50 mm/min, a¼ 1, 5, 10, and 15 μm In air
Workpiece CNC tool path Grinding parameters Coolant
roughness was measured by TR200 roughness meter. The surface roughness Ra was the mean value of five measured data. The detailed dry grinding conditions including axial-feed speed vf, wheel speed vw, depth of cut a and so on are shown in Table 3.
3. Dry ECD dressing mechanism Fig. 3 shows dry ECD dressing mechanism. In contrast to mechanical dressing, the dry ECD dressing utilized an impulse electro-discharge to remove the metal bond of grinding wheel without any mechanical removal (see Fig. 3a). When the nonconductive diamond grain cut the copper electrode, the rolled-up copper chip approached metal bond to some extent so that the spark discharge occurred. Thus, the metal bond was melted and thrown into a discharge crater [14]. Gradually, the diamond grains were protruded from wheel working surface. Moreover, the impulse discharge heat flow also shocked diamond grain edge. It is seen that most Cu chips were formed by mechanical cutting according to the cut trace on chip surface (see Fig. 3b). Some melted Cu chips by electrical discharge were formed in dry ECD dressing (DC-60V). This is because arc discharges derived from short circuit occurred when the DC open-circuit voltage was larger than 20 V [14]. This would lead to large discharge heat rather than spark discharge crater on metal bond surface, leading to the formation of melted Cu chip. In order to observe the removal of diamond grain before and after dry ECD dressing, the Raman spectra testing was performed. It is found that there were two peaks at 1580 cm 1 (G band) and 1350 cm 1 (D band) on diamond grain surface (see Fig. 3c). It corresponded to the first and second order Raman peak of graphite, respectively [16,19]. Because the characteristic peak of pure diamond is a single sharp peak at 1332 cm 1, the diamond grain top may be carbonized into graphite due to dry electrical discharge. This means that the diamond grain top could be removed by impulse discharge heat and mechanical friction between grain and electrode. It contributes to uniform grain protrusion height in grinding.
121
4. Results and discussions 4.1. Grain protrusion topographies versus dressing conditions Fig. 4 shows the SEM photos of diamond grain protrusion on dressed wheel surface versus dressing conditions. It is shown that the diamond grains were completely protruded and sharpened from metal bond surface under previous three dressing conditions, but Grain 2 was fractured and Grains 1, 3 and 4 were pulled out in dry ECD dressing (DC-32V) (see Fig. 4d). This is because large discharge energy removed so much metal bond that diamond grains were pulled out. It is also found that the dry ECD dressing (Pulse-25V) was able to remove the melted clusters derived from arc discharges in contrast to mechanical dressing (see Fig. 4a and b). Moreover, dry ECD dressings (DC-25V and DC-32V) still produced micro melted clusters on the metal bond surface (see Fig. 4c and d). The chemical elements of Points A, B, C and D on dry ECD dressed wheel surface in Fig. 4 are shown in Table 4. They were measured by energy dispersive X-Ray spectroscopy (EDX, INCAP FET-X3). It is shown that the component of Point A at Grain 1 was mainly composed of diamond (91.70% C) and Oxygen (7.93% O) (see Fig. 4b). This means that diamond grain surface was oxidized in dry ECD dressing. However, the components of Cu and Sn appeared at Point B on grain bottom (see Fig. 4b), thus molten metal bond was adhered to the border of diamond grain. It is also found that the Cu component appeared at Points C and D at cracked Grain 2 (see Fig. 4d). It indicates that the cracked diamond grain surface was covered with melted Cu electrode materials.
4.2. Recognition and extraction of diamond grain and metal bond Fig. 5 shows 3D topographies of dressed diamond wheel surface corresponding to Fig. 4. It is shown that the measured grain topographies were identical to the SEM observations (see Fig. 4). This means that these reconstructed topographies may display 3D information of grain protrusions. In order to extract diamond grain from metal bond surface, the 3D topographies (see Fig. 5) were dispersed to 3D coordinate point cloud O(x, y, z) (see Fig. 6a); then ћ
its normal vector n (x n, yn , zn) was calculated (see Fig. 6b); finally, the diamond grain boundaries were recognized through setting the threshold s. The micro-topographical extraction of diamond grains from metal bond on wheel working surface is shown in Fig. 6. When the threshold value s was set as 0.8, the most integrated diamond grains boundaries may be extracted. They are described as follows:
Fig. 3. Dry ECD dressing mechanism: (a) micro-removal scheme, (b) SEM photo of Cu chip and (c) Raman spectra of diamond grain surface.
122
Y.J. Lu et al. / International Journal of Machine Tools & Manufacture 88 (2015) 118–130
Fig. 4. SEM photos of diamond grain protrusion versus dressing conditions: (a) mechanical dressing, (b) dry ECD dressing (Pulse-25V), (c) dry ECD dressing (DC-25V) and (d) dry ECD dressing (DC-32V). Table 4 The chemical elements on dry ECD dressed wheel surface in Fig. 4. Testing points
A B C D
Element (%) Cu
C
O
Fe
Sn
0.37 69.71 100.00 22.95
91.70 22.97 0.00 59.94
7.93 1.37 0.00 15.21
0.00 3.89 0.00 1.90
0.00 2.06 0.00 0.00
⎛№ ћ O № O № O⎞ n = (x n , yn , z n) = ⎜ , , ⎟ x № y № z⎠ ⎝№
(1)
0 < z n < ƾ = 0.8
(2)
According to the extracted grain boundary outline, the microtopographies of diamond grain and metal bond were recognized and extracted (see Fig. 6b and c).
topographies of measured wheel surfaces [20,21]. This is because the diamond grain shape was not changed even if their measuring coordinate frames were not identical. Fig. 7 shows the matched point clouds and micro-topographical removal of diamond grain before and after dry ECD dressing. It is shown that the extracted diamond grains before and after dressing may be well matched each other by ICP matching (see Fig. 7 left). This means that the matched grain topographies may be used to calculate the height difference between two measured grain topographies. It is also found that the Pulse-25V power produced more diamond grain removal than DC-25V power (see Fig. 7 right). Their maximum removal heights of diamond grain were 9.9 μm and 9.6 μm, respectively. The reason is that the grain top was carbonized into graphite and then removed by impulse discharge heat and shock (see Fig. 3). The mean grain removal height is called grain removal height hgr, which can be calculated as follows:
1 N
N
Å
4.3. Micro-topographical removals of diamond grain and metal bond
hgr =
In order to compare the micro-topographical heights of diamond grain and metal bond before and after dressing, the Iterative Closest Point (ICP) algorithm was employed to match 3D grain
where hy is the removal height of grain micro-topography for a grid on a XOZ plane. N is the grid point number of grain microtopography. After the matched grain topographies were extracted
i=1
hy
(3)
Y.J. Lu et al. / International Journal of Machine Tools & Manufacture 88 (2015) 118–130
123
Fig. 5. 3D topographies of dressed diamond wheel surface: (a) mechanical dressing, (b) dry ECD dressing (Pulse-25V), (c) dry ECD dressing (DC-25V) and dry ECD dressing (DC-32V).
Fig. 6. The micro-topographical extraction of diamond grains from metal bond surface: (a) measured wheel topography, (b) diamond grain topographies and (c) metal bond topography.
from a metal bond surface, the mean metal bond removal height, called bond removal height hbr, can be achieved. Hence, the grain removal volume Vgr can be calculated by
Vgr = hgr Sxz
(4)
where Sxz is the projected area of grain topography on the XOZ plane. Because the concentration of diamond abrasive was 100%, the volumetric proportions of diamond grain and metal bond were 25% and 75%, respectively. Hence, grain area Sg and metal bond area Sm on cylindrical wheel surface may be approximately calculated as follows:
Sg ↣
ưDB 4
(5)
3ưDB 4
(6)
Sm ↣
where D and B are the diameter and width of diamond wheel, respectively (see Table 2). On wheel surface, the grain was
regarded as a circle, thus grain number ng can be given by
ng =
4Sg ưdg 2
(7)
where dg ¼ 350 μm is the nominal diameter of #46 diamond grain. Hence, total grain removal volume Vtgr and total bond removal volume Vtbr can be calculated as follows:
Vtgr = ng Vgr
(8)
Vtbr = Smhbr
(9)
Fig. 8 shows the micro-topographical removal heights and volumes of diamond grain top and metal bond surface in dry ECD dressing. It is shown that the Pulse-25V power increased the grain removal height hgr by about 70% and bond removal height hbr by about 155 times against DC-25V power (see Fig. 8a). The former is because it easily removed the diamond grain top without any crack by surface graphitization derived from impulse discharge heat, shock and mechanical friction. The latter is because it produced little melted clusters on metal bond (see Fig. 4c).
124
Y.J. Lu et al. / International Journal of Machine Tools & Manufacture 88 (2015) 118–130
Fig. 7. The matched point clouds and micro-topographical removal of diamond grain before and after dressing: (a) dry ECD dressing (Pulse-25V) and (b) dry ECD dressing (DC-25V).
According to the discharge dressing time T ¼80 min, the Pulse25V power removed diamond grains at 0.013 mm3/min and metal bond at 0.241 mm3/min, which were larger than the ones for DC25V power. Moreover, the total bond removal volume Vtbr was about as much 18 times as total grain removal volume Vtgr for Pulse-25V power (see Fig. 8b). As a result, the proposed microtopographical removal mode of metal bond and diamond grain may be used to describe the grain protrusion mechanism during dressing. 4.4. Grain cutting parameters versus dressing conditions Fig. 9 shows the grain protrusion profile characterized with grain cutting parameters. The grain top was regarded as the highest point of 3D grain profile. Along the cutting direction at grain top, grain protrusion height hg, gain rake angle γg, grain relief angle εg and grain top angle αg were defined as the grain cutting parameters. The grain top angle αg was defined along the vertical cutting direction. It is shown that integrated diamond grains along with sharpened cutting edges were protruded from wheel working surface after dressing. Fig. 10 shows grain cutting parameters versus grain protrusion height hg. It is shown that the dry ECD dressing averagely increased the grain protrusion height hg by about 12% against mechanical dressing. This is because it removed the melted cluster on metal bond surface (see Fig. 4a and b). In dry ECD dressing, the Pulse-25V power increased the hg by about 6% against DC-25V power. It also decreased the height difference of grain protrusion from 6 μm to 3 μm for highest Grains 1 and 4, leading to more uniform grain protrusion. The reason is that it did not produce such micro-melted clusters on metal bond surface as DC-25V power (see Fig. 4c), and it increased the removal of diamond grain top (see Fig. 8). It is also found that an increase in grain protrusion height hg led to an increase trend in negative grain rake angle γg (see Fig. 10a) and a decrease trend in grain relief angle εg (see Fig. 10b). This is
because top grains tips were easily worn and obtuse during dressing. Moreover, the dry ECD dressing averagely increased grain top angle αg by about 23% compared with mechanical dressing (see Fig. 10c). Moreover, it greatly decreased negative γg and increased εg for Grain 3. The reason is that impulse discharge heat and shock force mainly influenced the grains with medium protrusion height (see Fig. 3a). After dry ECD dressing, the mean grain protrusion height reached to 178 μm, which was about 50% of #46 diamond grain size. However, it was only about 33% and 18% for #60 and #180 diamond grinding wheel in mechanical dressing, respectively [4,22]. It contributes to an increase in chip capacity, which may decrease cutting heat to protect from surface damage in grinding. Moreover, the negative grain rake angle ranged from 70° to 80°, which was larger than those of 65° and 50° for #60 and #270 diamond grinding wheels, respectively [4,13]. It also contributes to an increase in critical grain cutting depth transferred from brittlemode removal to ductile-mode removal for mirror grinding [3]. 4.5. The impulse discharge energy parameters in dry ECD dressing Fig. 11 shows the discharge current Id and discharge voltage Ed versus discharge time td in dry ECD dressing. The discharge current trace was characterized as the impulse discharge parameters including impulse peak current Ie and impulse discharge duration τe [14]. It is shown that the Pulse-25V power mainly produced micro-spark discharges, leading to large impulse discharge current Ie and short impulse discharge duration τe (see Fig. 11a). This may explain why the Pulse-25V power may gradually remove the metal bond (see Fig. 4b), leading to large grain protrusion height (see Fig. 8 and 10). However, the DC-25V power produced a majority of micro-arc discharges, thus leading to little impulse discharge current and long impulse discharge duration (see Fig. 11b). The reason is that many micro-melted clusters were formed on metal bond surface (see Fig. 4c). This may also explain that it achieved less grain
Y.J. Lu et al. / International Journal of Machine Tools & Manufacture 88 (2015) 118–130
125
Fig. 8. The micro-topographical removal heights and volumes of diamond grain top and metal bond surface in dry ECD dressing: (a) grain removal height hgr and bond removal height hbr and (b) total grain removal volume Vtgr and total bond removal volume Vtbr.
Fig. 9. The grain protrusion profile characterized with grain cutting parameters including grain protrusion height hg, gain rake angle γg, grain relief angle εg and grain top angle αg.
protrusion height (see Fig. 10). Moreover, the DC-32V power produced much larger arc discharges than DC-25V power (see Fig. 11c). According to the wheel speed vw of 19.6 m/s and impulse discharge duration τe of 178 μs, the moving length of its arc discharge along wheel surface reached up to 3.48 mm, which was about 10 times as much as grain size dg. This means that the arc discharges along with short circuit shocked the protruded diamond grains, leading to their fracture and pull-out (see Fig. 4d). The impulse discharge energy parameters included impulse discharge energy We, maximum discharge frequency n and discharge energy rate Wd. Their relations may be described as follows:
caused micro-spark discharges (see Fig. 11a) and micro-arc discharges (see Fig. 11b), respectively. Hence, the Pulse-25V power increased the discharge frequency n by 20% and discharge energy rate Wd by 22% against DC-25V power, respectively (see Fig. 12b and c). However, the DC-32V power produced too large impulse discharge energy We and discharge energy rate Wd, thus leading to large arc discharges that shocked diamond grains along with high speed rotary wheel. As a result, the Pulse-25V power along with spark discharge was suitable for ECD dressing. It is identical to the critical open-circuit voltage of 20 V transferred from arc discharge to spark discharge [14].
We =
ⅈ0
td
Wd = nWe
IdEdd(td)
(10) (11)
where n was achieved from collected discharge waveforms (see Fig. 11). Fig. 12 shows the impulse discharge energy parameters in dry ECD dressing. It is shown that Pulse-25V power and DC-25V power produced same impulse discharge energy (see Fig. 12a), but they
4.6. Impulse discharge parameters versus micro-removal of metal bond In dry ECD dressing, the impulse spark discharge produced a sphere-shaped discharge crater on the metal bond surface. The impulse discharge removal sizes such as impulse removal height he, impulse removal diameter de and impulse removal volume me were correlated to the impulse discharge parameters (Ie and τe) [14]. They were described as follows [14]:
126
Y.J. Lu et al. / International Journal of Machine Tools & Manufacture 88 (2015) 118–130
Fig. 10. Grain cutting parameters versus grain protrusion height hg: (a) Grain rake angle γg, (b) grain relief angle εg and (c) grain top angle αg.
he = 5.41 ¬ Ie 0.77 ¬ ǀ e 0.19
(12)
de = 29.00 ¬ Ie 0.78 ¬ ǀ e 0.17
(13)
me = ư((de2he )/8 + he 3/6)
(14)
Fig. 13 shows the impulse discharge removal sizes on metal bond surface in dry ECD dressing. The impulse discharge removal sizes were calculated according to Eqs. (12)–(14) along with the impulse discharge parameters (Ie and τe) shown in Fig. 11. It is shown that the Pulse-25V power may produce larger impulse removal height he, impulse removal diameter de and impulse removal volume me than DC-25V power. Moreover, the impulse removal height he was much less than the half of normal grain size dg/2. This means that the pulse power produced larger discharge removal of metal bond than DC power even if their discharge energies were same. Although the DC-32V power produced much larger impulse removal sizes of metal bond, many grains were pulled out from wheel working surface due to arc discharges. In order to predict the cumulative metal bond removal with impulse discharge crater mode, the spark discharge frequency nspark can be given by
nspark = Ǻ n
(15)
where ω is the effective spark discharge ratio. According to the observed discharge waveform results, the ω was regarded as 20%. Thus, the number of impulse discharge craters Npulse was defined by
Npulse = nspark Ton
(16)
where Ton is the actual contact discharge time. According to the wheel trajectories, dressing parameters (vw, vf) and electrode sizes (L, B0) shown in Table 2 in dressing process, Ton was achieved as about 2 min. The predicted removal volume of metal bond Vpulse was calculated by
Vpulse = Npulseme
(17)
According to Eq. (6), the predicted removal height of metal bond hpulse was given by
h pulse =
4Vpulse 3ưDB
(18)
According to Eqs. (17) and (18), Vpulse and hpulse were calculated as 26.4 mm3 and 29.8 μm in the case of ECD dressing (Pulse-25V). It is seen that the predicted values were a little larger than total bond removal volume Vtbr of 19.3 mm3 and bond removal height hbr of 21.8 μm (see Fig. 8). This is because the model of impulse discharge removal sizes (he, de, me) were achieved under static
spark discharges [14]. In actual dressing process, the high speed rotary wheel may take away a part of discharge energy and heat. Moreover, the electrical discharge also occurred between electrode/chip and diamond grain surface covered with metal bond (Cu and Sn), leading to a consumption of discharge energy. Hence, the predicted discharge removal Vpulse and hpulse were a little larger than the metal bond removal Vtbr and hbr, respectively. The discharge energy utilization ratio was about 73% according to the ratio of Vtbr to Vpulse. Hence, the impulse peak current, discharge duration and discharge frequency derived from discharge current trace may be used to monitor microscopic grain protrusion. This also means that the self-optimizing dressing may be employed according to the relationships between impulse discharge parameters and predicted discharge removal of metal bond. 4.7. Dressing ratio and dressing efficiency The dressing parameters included wheel removal height hwheel, wheel removal volume Vwheel, dressing ratio γ and dressing efficiency η. In the WLI measurement of actual dressing parameters, the wheel profile was on-machine replicated on the graphite plate surface positioned grinder after dressing. The difference between wheel profiles before and after dressing was regarded as the wheel removal height hwheel [20]. Hence, the wheel removal volume Vwheel may be calculated by
Vwheel = ưDBh wheel
(19)
In order to evaluate the dressing abilities versus dressing conditions, the dressing ratio γ and dressing efficiency η were defined as follows [12]:
Ɛ=
Vwheel VE
(20)
ƛ=
Vwheel T
(21)
where VE is the dresser/electrode removal volume. T is the dressing time, which was 250 min and 80 min in mechanical dressing and dry ECD dressing (Pulse-25V), respectively. Fig. 14 shows the dressing parameters of diamond wheel for mechanical dressing and dry ECD dressing (Pulse-25V). It is shown that the wheel removal height hwheel of 19.3 μm and wheel removal volume Vwheel of 22.7 mm3 (see Fig. 14a and b) were roughly identical to the bond removal height hbr of 21.8 μm and total bond removal volume Vtbr of 19.3 mm3 (see Fig. 8). This indicates that the proposed micro-topographical removal model of diamond grain and metal bond was valid to evaluate the grain protrusion in dry ECD dressing.
Y.J. Lu et al. / International Journal of Machine Tools & Manufacture 88 (2015) 118–130
127
Fig. 11. Discharge current Id and discharge voltage Ed versus discharge time td in dry ECD dressing conditions: (a) Pulse-25V, (b) DC-25V and (c) DC-32V.
Moreover, the dry ECD dressing increased the dressing ratio γ by about 34 times and the dressing efficiency η by about 10 times against mechanical dressing (see Fig. 14c and d). The reason is that dry ECD dressing may rapidly remove the metal bond by the impulse spark discharges between metal bond and electrode chip, but the mechanical dressing along with mechanical cutting was
limited to the space between metal bond and dresser, which was dominated by coarse diamond size. In dry ECD dressing, the dressing ratio of 0.0133 and the dressing efficiency of 0.284 mm3/min for coarse #46 diamond wheel dressing were much less than the truing ratio of 2.82 and the truing efficiency of 12.25 mm3/min for fine #600 fine diamond
128
Y.J. Lu et al. / International Journal of Machine Tools & Manufacture 88 (2015) 118–130
Fig. 12. The impulse discharge energy parameters in dry ECD dressing: (a) impulse discharge energy We, (b) maximum discharge frequency n and (c) discharge energy rate Wd.
Fig. 13. The impulse discharge removal sizes on metal bond surface in dry ECD dressing: (a) impulse removal height he, (b) impulse removal diameter de and (c) impulse removal volume me.
Fig. 14. The dressing parameters of diamond wheel for mechanical dressing and dry ECD dressing (Pulse-25V): (a) wheel removal height hwheel, (b) wheel removal volume Vwheel, (c) dressing ratio and (d) dressing efficiency η.
wheel [12]. This means it is quite difficult to dress coarser diamond grinding wheel. 4.8. Grinding temperature and ground surface roughness Fig. 15 shows grinding temperature Ts and surface roughness Ra versus dressing conditions and grinding parameters. It is shown that the grinding temperature Ts and the surface roughness Ra decreased with decreasing depth of cut a and feed speed vf and increasing wheel speed vw for three dressing conditions. The Ra changes with feed speed vf and wheel speed vw were identical to the theoretical model of axial-feed grinding [7]. This also means that a decrease in grinding temperature led to a decrease in ground surface roughness. However, it was an exception for the depth of cut larger than 10 μm and feed speed larger than 30 mm/ min. The reason is that large depth of cut and feed speed could cause large friction force and mechanical vibration. In the case of depth of cut a of 1 μm, the dry ECD dressing (Pulse-25V) decreased the grinding temperature Ts by about 37%
and about 10% against mechanical dressing and dry ECD dressing (DC-25V), respectively (see Fig. 15a). Moreover, it decreased the surface roughness Ra by about 46% and about 10%, respectively (see Fig. 15b). The Ra ranged from 73 nm to 89 nm. This is because large grain protrusion height could decrease grinding heat, thus leading to a decrease in grinding temperature and surface roughness. Fig. 16 shows the SEM photos of dry ground surfaces and edges of carbide alloy versus dressing conditions. The dry grinding parameters were a ¼1 μm, vf ¼10 mm/min and vw ¼11.8 m/s. It is shown that the dry ECD dressing produced less micro-scratches on ground surface and less micro-cracks on ground edge than mechanical dressing (see Fig. 16 up). It is identical to the results of grinding temperature and surface roughness. This is because it increased the grain protrusion height, leading to larger chip capacity space. In dry ECD dressing, the Pulse-25V power produced less micron-scale brittle-cracks on ground surface than the DC-25V power, leading to smoother ground surface. It also produced less end edge wear.
Y.J. Lu et al. / International Journal of Machine Tools & Manufacture 88 (2015) 118–130
129
Fig. 15. Dry grinding temperature Ts and surface roughness Ra: (a) grinding temperature Ts and (b) surface roughness Ra.
Fig. 16. SEM photos of dry ground surfaces and edges of carbide alloy: (a) mechanical dressing, (b) dry ECD dressing (Pulse-25V) and (c) dry ECD dressing (DC-25V).
130
Y.J. Lu et al. / International Journal of Machine Tools & Manufacture 88 (2015) 118–130
References
Fig. 17. The macroscopic dry ground surface of carbide alloy using dry ECD dressed #46 coarse diamond grinding wheel.
Fig. 17 shows the macroscopic dry ground surface of carbide alloy. It is shown that the mirror-like surface was achieved in a macroscopic size using the dry ECD dressed #46 coarse diamond wheel. This is because the dry ECD dressing may enhance the uniformity of grain protrusion heights, leading to an increase in valid grain number. The other reason is that the coarse diamond grain produced large rigidity without any pull-out. The surface roughness Ra reached to 0.089 μm, which was much less than the Ra of 0.82 μm that was achieved by #120-140 diamond wheel [23]. As a result, the dry ECD dressed coarse diamond grinding wheel may be used to grind smooth surface macroscopically due to large grain protrusion height and rigidity. 5. Conclusions (1) In dry ECD dressing, the coarse diamond grain is removed at 0.013 mm3/min through the surface graphitization; the metal bond is removed at 0.241 mm3/min through spark discharge removals. The impulse discharge parameters may be used to monitor and optimize the microscopic grain protrusion according to their relations to the micro-removal of metal bond. (2) The Pulse-25V power along with spark discharges produce larger impulse discharge frequency and discharge energy rate than the DC-25V power along with arc discharges even if their discharge energies are same. It increases metal bond removal and diamond grain removal, thus leading to high and uniform grain protrusion. (3) The dry ECD dressing increases dressing ratio by about 34 times and dressing efficiency by about 10 times against mechanical dressing along with cutting removal, respectively. It averagely increases grain protrusion height by 12% and grain top angle by 23%. The mean grain protrusion height of 178 μm is about 50% of actual diamond grain size. (4) In axial-feed dry grinding of carbide alloy with a #46 coarse diamond wheel, the dry ECD dressing may decrease the grinding temperature by about 37% against mechanical dressing, leading to a decrease 46% in surface roughness. The macroscopic smooth surface with the Ra of 73–89 nm may be produced due to large grain protrusion height and rigidity.
Acknowledgments The project was supported by the National Natural Science Foundation of China (Grant no. 61475046).
[1] J. Webster, M. Tricard, Innovations in abrasive products for precision grinding, CIRP Ann.—Manuf. Technol. 53 (2004) 597–617. [2] Z.Y. Zhang, F.W. Huo, Y.Q. Wu, H. Huang, Grinding of silicon wafers using an ultrafine diamond wheel of a hybrid bond material, Int. J. Mach. Tools Manuf. 51 (2011) 18–24. [3] J. Tamaki, G. Sato, K. Maekawa, J. Yan, A. Kubo, Effect of cutting edge shape of diamond abrasive grain on ductile-mode grinding of optical glass-measurement of critical depth of cut by means of single grit diamond cutting, J. Jpn. Soc. Abras. Technol. 49 (2005) 335–340. [4] J. Xie, H.F. Xie, X.R. Liu, T.W. Tan, Dry micro-grooving on Si wafer using a coarse diamond grinding, Int. J. Mach. Tools Manuf. 61 (2012) 1–8. [5] H. Yasui, T. Sawa, Ultra-smoothness grinding of fine ceramics by coarse grain size diamond wheel-effect of grinding fluid, J. Jpn. Soc. Precis. Eng. 70 (2004) 853–859. [6] Q.L. Zhao, J.Y. Chen, H.T. Huang, X.Y. Fang, Grinding damage of BK7 using copper-resin bond coarse-grained diamond wheel, Int. J. Precis. Eng. Manuf. 12 (2011) 5–13. [7] J. Xie, Y.X. Lu, Study on axial-feed mirror finish grinding of hard and brittle materials in relation to micron-scale grain protrusion parameters, Int. J. Mach. Tools Manuf. 51 (2011) 84–93. [8] H.S. Lim, K. Fathima, A.S. Kumar, M. Rahman, A fundamental study on the mechanism of electrolytic in-process dressing (ELID) grinding, Int. J. Mach. Tools Manuf. 42 (2002) 935–943. [9] E. Weingärtner, S. Jaumann, F. Kuster, M. Boccadoro, Special wire guide for onmachine wire electrical discharge dressing of metal bonded grinding wheels, CIRP Ann.—Manuf. Technol. 59 (2010) 227–230. [10] A. Hosokawa, T. Ueda, T. Yunoki, Laser dressing of metal bonded diamond wheel, CIRP Ann.—Manuf. Technol. 55 (2006) 329–332. [11] J. Xie, J. Tamaki, An experimental study on discharge mediums used for electro-contact discharge dressing of metal-bonded diamond grinding wheel, J. Mater. Process. Technol. 208 (2008) 239–244. [12] J. Xie, H.F. Xie, M.J. Luo, T.W. Tan, P. Li, Dry electro-contact discharge mutualwear truing of micro-diamond wheel V-tip for precision micro-grinding, Int. J. Mach. Tools Manuf. 60 (2012) 44–51. [13] J. Xie, F. Wei, J.H. Zheng, J. Tamaki, A. Kubo, 3D laser investigation on micronscale grain protrusion topography of truncated diamond grinding wheel for precision grinding performance, Int. J. Mach. Tools Manuf. 51 (2011) 411–419. [14] J. Xie, J. Tamaki, In-process evaluation of grit protrusion feature for fine diamond grinding wheel by means of electro-contact discharge dressing, J. Mater. Process. Technol. 180 (2006) 83–90. [15] E. Weingärtner, R. Roth, F. Kuster, M. Boccadoro, F. Fiebelkorn, Electrical discharge dressing and its influence on metal bonded diamond wheels, CIRP Ann.—Manuf. Technol. 61 (2012) 183–186. [16] Z.Y. Zhang, H.M. Peng, J.W. Yan, Micro-cutting characteristics of EDM fabricated high-precision polycrystalline diamond tools, Int. J. Mach. Tools Manuf. 65 (2013) 99–106. [17] A. Klink, Wire electro discharge trueing and dressing of fine grinding wheels, CIRP Ann.—Manuf. Technol. 59 (2010) 235–238. [18] J. Xie, M.J. Luo, K.K. Wu, L.F. Yang, D.H. Li, Experimental study on cutting temperature and cutting force in dry turning of titanium alloy using a noncoated micro-grooved tool, Int. J. Mach. Tools Manuf. 73 (2013) 25–36. [19] A.C. Ferrari, Determination of bonding in diamond-like carbon by Raman spectroscopy, Diam. Relat. Mater. 11 (2002) 1053–1061. [20] J. Xie, Y.W. Zhuo, T.W. Tan, Experimental study on fabrication and evaluation of micro-pyramid-structured silicon surface using a V-tip of diamond grinding wheel, Precis. Eng. 35 (2011) 173–182. [21] J. Xie, X.R. Liu, K.K. Wu, Y.J. Lu, P. Li, Evaluation on 3D micro-ground profile accuracy of micro-pyramid-structured Si surface using an adaptive-orientation WLI measurement, Precis. Eng. 37 (2013) 918–923. [22] J. Xie, J. Xu, Y. Tang, J. Tamaki, 3D graphical evaluation of micron-scale protrusion topography of diamond grinding wheel, Int. J. Mach. Tools Manuf. 48 (2008) 1254–1260. [23] Y. Wang, X.J. Zhou, D.J. Hu, An experimental investigation of dry-electrical discharge assisted truing and dressing of metal bonded diamond wheel, Int. J. Mach. Tools Manuf. 46 (2006) 333–342.