Printer Friendly

Comparison of orbital and linear vibration welding of thermoplastics.

INTRODUCTION

The use of frictional heat to produce plastic welds has been an industrial process since the mid 1970s (1). The method is fast, reliable and works with a wide range of materials and applications. The process works by clamping two parts together under relatively high pressure (0.5-5 MPa at the faying surface) and rubbing the two parts together generating frictional heating (2). The relative cyclic motion between the parts is typically at frequencies between 100 and 250 Hz. with peak amplitude between 0.5 and 2.0 mm of displacement. This results in frictional heating until the surface melts and viscous heating of the melt occurs until a predetermined weld time or amount of melt is achieved and the vibrations are stopped. Once the melt solidities, the parts are welded. Stokes (3), (4) studied linear vibration welding experimentally and developed a process model showing that the process can be divided into four discreet phases: Phase I-solid/solid friction heating, Phase II-transient flow. Phase Ill-steady state flow, and Phase IV-solidilication of the melt. Each of these phases can be relatively easily determined from the melt down rate or penetration-time curve.

Using frictional heat to join plastics is fast, reliable, and works with a wide range of materials and applications. In the past, nearly all the frictional welding processes were linear (except for spin welding, which is only applicable for relatively small and round parts); however over a decade ago, orbital welding (5), (6) was introduced to the market. In this case, the relative motion between the two parts is not linear, but has an orbital path where each point on the moving part forms an orbit around a stationary point on the nonmoving part. The goal of this work was to compare linear and orbital vibration welding and determine their differences for a variety of materials, acrylonitrile-buta-diene-styrene (ABS), polycarbonate (PC), Nylon (PA) and polypropylene/polyethylene copolymer (PP/PE). The primary effects that were studied included: (1) weld strength, (2) impact strength (toughness), (3) amount of weld flash, (4) cycle time, (5) residual stress level, (6) welding of unsupported walls, (7) weld consistency when vibration motion is parallel and perpendicular to the walls, (8) minimum amplitude required to generate heating and (9) power dissipation. Although not detailed in this article, it is important to note that with orbital welding, part geometry and the joint design must accommodate relative motion between the parts in all directions in the welding plane, while linear vibration welding requires relative motion between the parts in only one axis in the welding plane.

The instantaneous displacement and velocity for linear vibration welding are sinusoidal as shown by the following relations:

x(t) = [A.sub.0] sin ([omega]t) (1)

V(t) = dx(t)/dt = [A.sub.0][omega] cos ([omega]t) (2)

where x(t) is the displacement, V(t) is the velocity, [A.sub.0] is the amplitude of vibration, [omega] is the radial frequency of the vibration and t is time. As observed from Eqs. 1 and 2, the magnitude of both the displacement and velocity vary with respect to time, oscillating from zero to a positive maximum back to zero and a negative maximum and then back to zero again repeatedly. However, for orbital welding, if the amplitude of vibration for the x and v direction are equal (circular motion), then the magnitude of the velocity is constant or time independent. The displacement for orbital welding can be divided into an x and a v component with their respective amplitudes shown in Eq. 3.

x(t) = [A.sub.x] sin([omega]t)

y(t) = [A.sub.y] cos([omega]t] (3)

Where [A.sub.x] = Peak amplitude in x direction

[A.sub.y] = Peak amplitude in y direction (3)

Similarly, the velocity for orbital welding can be expressed as shown in Eq. 4. When the amplitude of vibration is the same in the x and y directions than the motion becomes circular and the velocity is constant as shown below.

[v.sub.x] = dx/dt = [A.sub.x][omega] cos ([omega]t)

[v.sub.y] = dy/dt = [A.sub.y][omega] sin ([omega]t) (4)

velocity = v = [square root of ([v.sub.x.sup.2] + [v.sub.y.sup.2])]

= [square root of ([A.sub.x.sup.2] [[omega].sup.2][cos.sup.2] ([omega]t) + [A.sub.y.sup.2] [[omega].sup.2] [sin.sup.2]([omega]t)]

If [A.sub.x] = [A.sub.y] = [A.sub.0], then

v = [square root of ([A.sub.0.sup.2] [[omega].sup.2][[cos.sup.2] ([omega]t) + [sin.sup.2] ([omega]t) = [A.sub.0][omega])] (5)

With the velocities defined, the frictional heating equations can be derived. During a frictional welding process, there are two main forces of concern: (1) Clamp or Normal force; this force affects the working force and it also influences squeeze flow, molecular alignment and out-gassing [7] and (2) Working force; this is the opposing force produced by the part-to-part interfacial friction and it resists the vibrational displacement.

Assuming Coulomb friction with a constant friction coefficient ([mu]), then the working force ([F.sub.w]) is related to the clamp force (F) as follows:

[F.sub.w] = [mu]*F (6)

For equal amplitudes and clamp forces, orbital motion will dissipate significantly more power (8) than linear motion during Phase I as shown Table 1. Thus, the ratio of linear to orbital motion power dissipated during Phase I is 64% as shown in Eq. 7.
TABLE 1. Power dissipated during Phase 1 and Phase 3 for linear and
orbital vibration welding.

 Linear vibration Orbital vibration

Phase 1--Solid/solid [mu]F[A.sub.0]| [mu]F|
friction heating [omega] [A.sub.0][omega]|
Instantaneous power P = cos([omega]t)|
[F.sub.w] . v =

Average power 2[mu]F[A.sub.0] [mu]F|
[P.sub.avg] = [omega]/[pi] [A.sub.0[omega]|
[F.sub.W].[v.sub.avg] =

Phase 3--Steady state [eta][A.sub.0.sup2] [eta]([[A.sub.0]
flow Instantaneous heat ([[omega]cos [omega]).sup.2]/2h
flux q = [eta]. ([omega]t).sup.2]/2h)
[v.sup.2]/2h =

Average heat flux [eta][A.sub.0.sup.2] [eta][A.sub.0.sup.2]
[q.sub.avg] [ = [[omega].sup.2]/4h [[omega].sup.2]/2h
[eta].[v.sub.avg.sup.2]
=


[[P.sub.linear]/[P.sub.orbital]] = [[2[micro][FA.sub.0][omega]]/[pi]]/[[mu][FA.sub.0][omega]]] = [2/[pi]] = 0.64 = 64% (7)

It is known that, once melting starts, the coefficient offriction is no longer applicable and the working force is defined by shearing of the melt [4], which will be discussed in the following. Therefore, during Phase II of the cycle, a transition occurs from solid-to-solid frictional heating to melt viscous shear heating [4] and the heating ratio in Eq. 7 is no longer applicable. As described in detail by Stokes [4] for a Newtonian fluid, the behavior during Phase II is complex and entails melting as well as squeeze flow. Nevertheless, during this phase, orbital motion would result in more heating than linear motion, and therefore it will also result in more squeeze flow than linear motion.

As modeled by Stokes for a Newtonian fluid [4], during Phase III steady state is reached where the amount of polymer that is melted due to viscous shear heating is equal to the amount of melt that is squeezed out due to the clamping force. Therefore, during Phase III the melt layer thickness remains constant. As shown by Stokes [4], heating during Phase III depends on the viscosity of the melt, the steady state melt layer thickness and the amplitude of vibration. For a melt layer of thickness 2h. the instantaneous power dissipated per unit area or the heat flux generated due to shear heating is as follows:

q = [eta][v.sub.inst.sup.2]] (8)

where, q is the instantaneous heat flux, [eta] is the viscosity, and [[upsilon].sub.inst] is the instantaneous velocity. Table 1 shows the instantaneous heat flux for Phase III for linear and orbital vibration welding. By including the velocities for linear and orbital motions and averaging over a cycle the average heat flux is found as shown in Table 1. Thus, for a given melt layer thickness that is the same for both linear and orbital motion, the ratio of the heat flux for linear and orbital motion in Phase III is then:

[[q.sub.linear]/[q.sub.orbital]] = [[[eta][A.sup.2][[omega].sub.2]/4h]/[[eta][A.sup.2][[omega].sup.2]]/2h = 0.50 = 50% (9)

Thus, during Phase I of the weld cycle, the ratio of the power dissipation between linear and orbital welding should be ~0.64 and during Phase III this ratio should be ~0.50 assuming a constant melt thickness. It should be noted that because orbital welding produces more heat, the steady state melt thickness should be lower for orbital welding than linear welding, further increasing the power dissipation. Although this effect is noted, it is not modeled in this article.

Thus, if adiabatic heating is assumed (no heat losses), then it is possible to estimate the expected relative increase in power dissipation for orbital welding, and the relative decrease in cycle time, as shown in Table 2. Although the assumption of adiabatic heating may not be fully accurate for these estimates it is believed that it is valid for first order approximations. The higher heating rate for orbital welding has a number of advantages, including short cycle times and lower minimum required amplitude. However, the additional heating can also reduce the melt viscosity as a result of the higher bond-line temperature (9). This higher viscosity can result in shear thinning, adverse molecular alignment in post-weld morphology and relatively high residual stresses. It has also been noted that higher bondline temperatures can lead to out-gassing of volatiles, such as moisture [7]. That is to say, if the vapor pressure exceeds the bondline pressure, the bubbles nucleate and grow if sufficient time is allowed.
TABLE 2. Details on the duration of each phase of the weld (PC.
Amplitude = 1.77 mm. 2 MPa)

 Phase Phase Decrease in
 number duration/time (%)

I Solid/Solid friction 36% (1.62-1.03 S)

II Sold/Liquid/Solid (solid and shear 26% (0.77-0.57 S)
 friction)

III Steady state flow-shear friction 30% (1.92-1.35 S)

I + II + III Total cycle 32% (4.31-2.95 S)


EXPERIMENTAL

Table 3 details the materials that were studied. Dissimilar material combinations were not considered. The only filled material was PA-6, which had 25% by weight short glass fibers added.
TABLE 3. Material properties.

Number Material Manufacturer Material-ID Tensile Flow rate
 strength (g/10 min)
 (Mpa)

 1 PP/PE (cp) Huntsman AP 6120-HS 23.44 20

 2 PC GE Plastics Lexan 141 62.05 10.5

 3 ABS GE Plastics Cycolac 46.54 2.0
 GDT6400

 4 25 wt% Glass AlliedSignal Capron 8233 179.2 Not
 reinforced determined
 PA-6(Nylon)


All materials were molded and machined into three sample configurations: (1) 6.35 mm plaques (6.35 X 101.6 x 71.1 [mm.sup.3]) weld area = 645 [mm.sup.2] (2) Modified 6.35 mm plaques (see Fig. 1). and (3) Ribbed samples (detailed in Fig. 2) weld area ~2581 [mm.sup.2]. The 6.35 mm plaques were welded in a butt joint configuration and in all cases, unless otherwise noted, oriented so that the 101.6-mm length was parallel to the linear vibration motion.

[FIGURE 1 OMITTED]

A modified Branson VW-4 vibration welder with an orbital head that had isolated magnets was used for both linear and orbital welding studies. Each motion was independently achieved by modifying the controller's parameters. In those studies involving the orbital motion, the controller was programmed to produce a circle with a diameter equal to the desired peak-to-peak linear amplitude. The equipment had a maximum clamp force of 15,000 N. In the studies requiring relatively low clamp force (less than 890 N), the fixture was fitted with a pneumatic cylinder. In these studies, the welder's hydraulic clamping system was brought against mechanical stops and the pneumatic cylinder in the fixture was used to apply the welding pressure.

The welding system was equipped with a data acquisition system that monitored and recorded the: (1) power, (2) melt distance. (3) amplitude (x and y direction), (4) clamp force, and (5) frequency at a minimum of 100 samples per second for each.

Overall, there were three designs of experiments that were used. Each of these is detailed in the following sections and each set of experiments was conducted with both linear and orbital welding and all four materials.

Full-Factorial-DOE

These experiments evaluated the effects of: (1) amplitude (0.77, 1.27, 1.52, and 1.78 [mm.sub.pp]), (2) clamp pressure at weld (1. 2. 4.8. and 6.9 MPa) and (3) collapse/melt down (0.77. 1.27. 1.52. and 1.78 mm).

The 101.6 X 6.35 [mm.sup.2] samples were welded in a butt joint configuration for this study. The welds were made with the machine set to weld until the preset collapse value was achieved. The machine measured the collapse with an internal LVDT (Linear Variable Differential Transformer) based on the displacement of the welding heads. Because this displacement does not account for deflections within the fixture, samples and machine, the actual collapse was determined by comparing the before and after welded heights of the samples.

Welds were made at all variable combinations for a total of 64 welds (4 x 4 x 4) for each material and each process. One weld sample was then cut into multiple test samples. Additional experiments were conducted to evaluate the effect of welding with a relatively low amplitude (0.5 [mm.sub.pp]) with four clamp pressures at weld (1, 2. 4.8, and 6.9 MPa), and one collapse/melt down (0.5 mm). Each weld was visually examined and the resulting weld flash was rated on a scale from 1 to 10; with 1 indicating no visible flash and 10 indicating excessive flash.

Wall Height Test

These experiments were conducted to determine if orbital welding could weld unsupported walls better than linear welding. These experiments evaluated the effects of: (1) amplitude (1.27 and 1.78 [mm.sub.pp]), (2) clamp pressure at weld (1. 2. and 4.8 MPa) and (3) wall height (6.35, 12.7. and 19.1 mm; measured from edge of sample to fixture).

The 101.6 X 6.35 [mm.sup.2] samples were welded in a butt joint configuration with the linear vibration motion being perpendicular to the wall thickness for this study. The machine was set to weld to a collapse of 1.27 mm. The actual collapse was determined by comparing the before and after welded heights of the samples. Each weld was visually examined for the resulting weld flash, which was rated on a scale from 1 to 10. with 1 indicating no visible flash and 10 indicating excessive flash

Power Test

These tests were conducted to determine the difference between linear and orbital welding in terms of power dissipation/draw. The ribbed samples (see Fig. 2) were used in this study. These experiments evaluated the effects of: (1) amplitude (0.76. 1.27, and 1.78 [mm.sub.pp]). (2) clamp pressure at weld (0.5, 1.7, and 3.4 MPa) and (3) collapse (1.27 mm). The power was estimated by measuring the power into the drive unit on all three electrical phases and summing of the power of the separate phases. The power dissipated for orbital welding was compared to the power dissipated power in linear welding for a given set of conditions and the relative increase was calculated as detailed in Table 2.

[FIGURE 2 OMITTED]

Weld Characterization

Two primary tests were conducted on the resulting welds, namely tensile test and impact test (toughness test). For tensile and impact testing, the samples were machined from the welded plaques as shown in Fig. 3. In most cases, tensile tests were performed on the welds. A Tinius Olsen 5000 testing machine was used with a cross head speed of 2 mm/min. The maximum load was recorded as the failure load. In all cases, the cross-sectional weld area was measured and the tensile strength is reported in terms of stress (MPa). It should be noted that in a few cases the samples yielded near the grips. In these cases, the weld strengths were assumed to be similar to parent material strength.

[FIGURE 3 OMITTED]

The impact strengths were measured using a standard Charpy swing-mass Riehle model P1-2 impact tester. It is important to note that no notch was machined at the weld.

The residual stresses were measured only for the PC welds because this is the only material studied that had a published procedure for solvent-residual stress measurements (10). The protocol details the application of different concentrations of a solvent to the weld for 3 min and then examining the weld for cracks. Once a particular concentration is identified that promotes cracking, a calibration curve is then used to estimate the level of residual stress. The "micrograph" samples were used for this evaluation, see Fig. 3.

RESULTS AND DISCUSSION

Weld Strength Results

Figure 4a shows the relationship between collapse and weld strength for PC welded with orbital and linear motion. The clamp pressure and amplitude were held constant at 0.35 MPa and 1.4 [mm.sub.pp], respectively. Each data point is the average of two weld samples. The lines are added to show the trend and are not statistically significant. Generally, the weld strength increases with collapse and then remains constant once the steady state phase (Phase III) is reached. It is important to note that there is little difference between the two processes. However, the line for the orbital process appears to be shifted slightly to the left (lower collapse values) relative to the linear process. This suggests slightly less collapse is required with orbital welding to result in similar weld strength compared to linear welding. Another point of interest is that the maximum achievable strength for both processes is similar (57 MPa).

Figure 4b shows the relationship between collapse and weld strength for PP/PE welded with orbital and linear motion. The welding parameters were the same as detailed for PC. Again, the data for the orbital process is shifted slightly to the left (lower collapse values) relative to the linear process. Thus, again it is seen that the orbital welding process may require slightly less collapse. Again the maximum achievable strength for both processes is similar (15 MPa).

[FIGURE 4 OMITTED]

Based on the balance of the experimental data. Fig. 5 compares the maximum recorded weld strength for all the data collected from the full-factorial DOE for each material and each process, which may not correspond to the conditions for which the maximum weld strength was achieved in Fig. 4. As can be seen, except for ABS, the difference in strength between linear and orbital welding is minimal. In addition, it can be seen that all these maximum weld strength values are very close to the bulk material strength (see Table 3) with the exception of the glass reinforced PA, where the strength is close to that of the PA matrix.

[FIGURE 5 OMITTED]

Impact Weld Strength Results

Figure 6a shows impact strength as a function of collapse for orbital and linear welding with PC at various collapse values. The clamp pressure and amplitude were held constant at 0.35 MPa and 1.4 [mm.sub.pp], respectively. Because each data point represents only a single test sample and because impact testing typically produces a significant amount of scatter, it is difficult to determine an absolute trend for the data. It is seen that there is little difference in the general trend for each process. Again trend lines are added to aid with visualization of the results.

[FIGURE 6 OMITTED]

Figure 6b shows impact strength as a function of collapse for orbital and linear welding with PP/PE copolymer using the same welding parameters as with PC. The results are very similar to those seen with PC, except that the absolute values are lower, which is expected because of the relative difference in base material impact strength. That is to say, the impact strength of PC is between 20 and 30 (Nm/[m.sup.2]) while for PP/PE it is only 3 to 7 (Nm/[m.sup.2]) (11). Thus, it is seen that for both PC and PP/PE, the impact strength factor (impact strength of weld/base material impact strength) is only 0.72 for PC and 0.30 for PP/PE copolymer.

Cycle Time

The time required to achieve a preset collapse value for each weld with orbital welding was compared to linear welding and the percent change in weld time was calculated using Eq. 10.

Weld time decrease % = [[([t.sub.L] - [t.sub.R])]/[[t.sub.L]]] x 100% (10)

Where [t.sub.L] is the linear vibration weld time and [t.sub.R] is the orbital vibration weld time. It is important to note that for each calculation, the process parameters for each process were identical, that is collapse, pressure and amplitude. As seen in Fig. 7, in all cases the cycle time for orbital welding was significantly shorter. The average decrease in weld time for all four materials was 32.8%, which is within reasonable agreement of the theoretical value between 36% and 50%.

[FIGURE 7 OMITTED]

To gain insight into the reduction of cycle time for orbital welding, the melt distance during the two welding cycles, as seen in Fig. 8, was characterized. It is seen that for all phases (I, II. and III as defined in V.J. Stokes' [3] work) of the weld cycle, the duration of each phase was reduced for orbital welding. Table 4 shows the average time reduction for orbital welding compared to linear vibration welding for all four phases for PC with the standard 6.3-mm butt joint. It is seen that the largest reduction in cycle time occurs during Phase I, which is in agreement with the models which predict that greatest reduction in cycle time during this phase.

[FIGURE 8 OMITTED]
TABLE 4. Summary of findings.

 Potential Benefit Results Impact of
 orbital
 welding

Decrease in weld time Seen in all materials Positive

Better at welding unsupported Seen in all materials Positive
walls

More consistent w/unsupported Seen in all materials Positive
walls

Higher power dissipation Seen in all materials Positive

Lower minimal amplitude Significant in only 1 (ABS) Limited
 material-limited benefit in 2
 (PP/PE, PA) materials

Increased weld strength Only seen in l(ABS) of the Limited
 four materials

Increased impact strength No visible difference None

Decrease in residual stress Orbital produces higher Negative
 residual stresses

Effects of disengagement Orbital welding produced Negative
 welds with a weld strength
 that was a compromise between
 0 and 90 degree linear
 welding with thin walls

Appearance of weld flash In one application the Limited
 appearance changed from hair
 like flash to particulates

Decrease in weld flash No visible difference None


Power Dissipation

The results of the power studies fell within the theoretical values as detailed in Table 2. To increase the power levels, in the majority of the tests the so-called ribbed plaques were used to compare the power dissipation between orbital and linear welding. As shown in Fig. 9 for most materials, orbital welding dissipated from 50% to 150% more power compared to linear welding. In addition, the orbital welding achieved a set collapse value in 50% of the time required for linear welding. Thus, for most materials, the experimental values and theoretical values are in agreement.

[FIGURE 9 OMITTED]

However, the average increase in power draw for PP/PE copolymer is over 150% higher for orbital welding compared to linear welding. This increase is significantly higher than the expected value of between 56% and 100%. It is believed that this increase is due to shear thinning of the molten layer in orbital welding, which increases the work force. That is to say. the molten layer in the orbital welds is thinner compared to linear welds. To verify this effect, the resulting welds were microtomed and the final melt thickness was measured for orbital and linear welding. It was found that for PC the melt thickness is ~25% less with orbital welding than linear welding; however, with PE/PP the thickness is nearly 50% less. Although there is no conclusive explanation for this observation, it is believed to be related to the rheological properties of the various materials.

Weld Flash Comparison

Figure 10 shows photographs of typical welds created by linear and orbital motions for the various materials. These samples are 25.4-mm wide and are the samples remaining from the sectioned welds. The left edge of each photograph is the outer edge of the weld and the right side of the photograph is the cut edge of the sample. The weld Hash is seen through the center of the photographs. The photographs show that there is no significant difference in the appearance of the weld flash between the two processes for ABS and PP/PE. However, for PC and PA, the linear weld appears to have slightly more "hair-like" flash, although the difference in minimal.

[FIGURE 10 OMITTED]

Minimum Required Amplitude

On the basis of the welds made with the full-factorial DOE matrix (102 mm x 6.4 mm plaques), it was found that for most of the materials evaluated, for a given clamp-force setting, it was possible to weld with a lower amplitude and achieve 80% of the base material strength with orbital motion compared to a linear motion. As seen in Fig. 11). the decrease in minimum required amplitude varied from 0% with PC to 30% for ABS (5% for PA and 10% for PP/PE). The advantage of lower required minimal amplitude is the ability to weld higher unsupported walls and design parts with less clearance for motion. It is interesting to note that ABS required relatively high amplitudes compared to the other materials. Although this is not fully understood, it is believed that this is related to ABS's high coefficient friction and low stiffness which may have resulted in attenuation of the amplitude.

[FIGURE 11 OMITTED]

Effect of Disengagement

Figure 12a shows weld strength as a function of amplitude for PC samples using the U-joint design. In this graph, the direction of vibration was parallel to the wall length. It is interesting to note that both processes are relatively amplitude-independent. The main item to note is that the orbital process produced welds that were only ~50% as strong as those produced by the linear process. This is probably caused by either exposure of the melt to atmosphere, causing premature solidification, or side wall deflection. It is also important to note that the minimal required amplitude with linear welding to achieve 80% of the base material strength is slightly lower than previously reported with the standard butt joint.

[FIGURE 12 OMITTED]

When the same tests were completed with the tee-joint samples, which are more representative of an application, the two processes produced welds with more comparable strength. In these studies, an additional experiment was conducted, in which 90[degrees] linear (cross-thickness) welding was also evaluated. As shown in Fig. 12b, orbital welding produced weld strengths that had values between welds made with 0[degrees] linear and 90[degrees] linear (cross-thickness) welding.

To allow relative comparison for PC and disengagement, weld strength for linear (0[degree]) and orbital welding were plotted as a function of the ratio of amplitude and wall thickness for the different joint designs (see Fig. 13). It is seen that with relatively thick walls, orbital and linear welding produce welds with similar strengths; however, as the wall thickness decreases, the orbital welds experience a significant loss in weld strength [from 62 to 25 MPa (56% reduction)], while linear welds experience only a modest loss in strength [55 MPa to 47 psi (6% reduction)]. Thus, it is again seen that with unsupported thin walls, orbital welds tend to be weaker than welds made with linear (0[degrees]) and in most cases have weld strengths between welds made with 0[degrees] and 90[degrees] linear welding.

[FIGURE 13 OMITTED]

With PP/PE samples, only the tee-joint design was studied because it was more representative of designs used in real applications. As shown in Fig. 14, orbital welding performed similarly to welding in a 0[degrees] linear and 90[degrees] linear mode. It is believed that the issues of premature solidification caused by disengagement of the parts at the interface are not as serious with PP/PE compared to the other materials when using the tee-joint. This is probably due to PP/PE's relatively low melt temperature, high viscosity and low coefficient of friction.

[FIGURE 14 OMITTED]

Unsupported Wall Heights

As seen in Fig. 15a, there was little difference between linear-0[degrees] and orbital welding strengths as a function of wall height with ABS samples. It should be noted that while data from the full-factorial experiments showed an increase in weld strength with orbital welding over linear welding, this data did not show this trend. It is believed that this discrepancy is due to ABS's low modulus (low stiffness), which reduced the relative motion perpendicular to the wall at the faying surface, thus reducing the benefits of orbital welding that promoted higher weld strength. The graph does show a difference between linear-90[degrees] and orbital welding, which shows that for wall heights of 12.5 mm the linear-90 [degrees]process did not weld the ABS plaques, while the orbital process was able to weld the plaques nearly independently of the wall height.

[FIGURE 15 OMITTED]

In the case of PC samples (Fig. 15b). it is seen that with linear-0 [degrees] welding, there is a slight decrease in weld strength with increased wall height, while with orbital welding there was relatively little difference in weld strength relative to wall height. It is also important to note that linear-90 [degrees] welding did not weld the plaques regardless of the wall height.

As seen in Fig. 15c, there was little difference between linear-0 [degrees] and orbital welding strengths as a function of wall heights with PP/PE. Again, when considering linear-90[degrees]welding, there is a significant loss in weld strength with increasing wall height (from ~ 14-0 MPa when the wall height was increased to 12.5 mm).

In the case of PA (Fig. 15d), it is seen that with a linear-0[degrees]welding, there is a slight decrease in weld strength with increased wall height, while orbital welding was relatively unaffected. It is also important to note that linear-90 welding did not weld the plaques regardless of the wall height.

Residual Stresses

These tests were only performed on PC because this is the only material studied that had published procedures for measuring residual stresses. Selected samples were tested using the GE Plastics solvent tests [10].

Two sets of welding conditions were selected to compare the residual stresses between linear and orbital welding: a low and high clamp pressure, 1 and 5 MPa. These were selected because it was theorized that the majority of the stresses are produced by: (1) different heating and cooling rates caused by changes in the clamping force, (2) different shearing of the melt caused by changes in the heating rates, and (3) different shearing of the melt caused by different clamp pressures.

It was seen that at both clamp pressures, orbital welding produces higher residual stresses. This is consistent with the findings that showed PC orbital welds had thinner bondlines, which is indicative of higher shear rates and faster heating and cooling rates. It is interesting to note that at the higher clamp pressures, the relative increase in the amount of residual stresses from linear to orbital is less. For example, with 1 MPa of clamp pressure, there is a 44% increase in the amount of residual stresses when comparing linear and orbital welding (4.25-6.10 MPa), while there is only 27% increase at the high clamp pressure (7.75-9.85 MPa). This is expected because at the higher clamp pressure, independent of the process. the higher shearing of the melt and faster heating and cooling promotes more residual stresses.

Summary

It was theorized that there were 10 potential benefits of orbital welding. Table 4 lists these potential benefits, the experimental results and the impact of the results on orbital welding compared to linear welding. As seen in the Table 4, 4 of the 10 potential benefits were actually observed, and another 2 were marginally seen. In addition, there were two findings that indicated that orbital welding had shortcomings compared to linear welding, higher residual stresses and problems when welding thin walls.

CONCLUSIONS

For the same welding conditions, the cycle time for orbital welding was 30% to 50% faster than linear vibration welding. Power dissipation levels for orbital welding were between 80% and 120% higher than for linear vibration welding. Orbital welding appears to produce thinner melt layers and thinner heat affected zones. In addition, orbital welding was able to weld unsupported walls better compared to linear welding. Orbital welding also produced slightly less flash compared to linear welding. The main shortcomings of orbital welding are that it produces slightly higher residual stresses at the weld region and has greater problems with material disengagement for thin walls. Overall, orbital welding is slightly better in some application compared to linear welding, especially in terms of shorter cycle times and welding of unsupported walls.

ACKNOWLEDGMENTS

The authors thank Branson Ultrasonics for their support and for allowing them to publish these data. They also thank Dr. V.J. Stokes for his help and insight in this work.

REFERENCES

(1.) J. Mengason, Plast, Eng., 36, 20 (1980).

(2.) H. Potente and M. Ruthmann, Kunststoffe, 77, 711 (1987).

(3.) V. Stokes, Polym, Eng. Sci., 28. 718 (1988).

(4.) V. Stokes, Palym. Eng. Sci., 28. 728 (1988).

(5.) D. Grewell, A. Benatar, and J. Park, Plasties and Composites Welding Handbook, Hanser, Munich. Germany, I (2003).

(6.) I. Froment, SPE ANTEC Proc., 1, 1285 (1995).

(7.) H. Potente, SPE ANTEC Pro., 1, 1320 (1994).

(8.) F. Sear. M Zemansky, and H. Young, University Physics, Addison-Wesley, Menlo Park. CA. 6. 25 (1982).

(9.) W. Myers. Textbook of Polymer Science. Wiley-Interscience, New York. NY. 315 (1984).

(10.) GE Lexan Resin Technifacts T-77. Solvent Stress Analysis, GE Lexan Products Division, Pittsfield, MA (1982).

(11.) R. Throne. Polymer Engineering Principles, Hanser, Munich, Germany, 792 (1993).

David A. Grewell, (1) Avraham Benatar (2)

(1) Department of Agricultural and Biosystems Engineering, Iowa State University, 100 Davidson Hall, Ames, IA 50011

(2) Plastics and Composites Joining Laboratory, Department of Industrial, Welding, and Systems Engineering, The Ohio State University, 1248 Arthur E. Adams Drive, Columbus, OH 43221

Correspondence to: David Grewell; e-mail: dgrewell@iastate.edu DOI 10.1002/pen.21254

Published online in Wiley InterScience (www.interscience.wiley.com).

[C] 2009 Society of Plastics Engineers
COPYRIGHT 2009 Society of Plastics Engineers, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2009 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Grewell, David A.; Benatar, Avraham
Publication:Polymer Engineering and Science
Article Type:Technical report
Date:Jul 1, 2009
Words:5863
Previous Article:Dimensional stability of single-site ethylene copolymers in rotational molding.
Next Article:The use of crack layer theory to predict the lifetime of the fatigue crack growth of high density polyethylene.
Topics:

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters