**3.2 Compression testing**

A summary of the compression test results, as displayed in Table 5, shows strengths that are higher than those obtained in tension testing. Higher compressive strengths are often observed in polymers, and specifically for bulk ABS materials (Ahn et al., 2002). In this study, the average tensile yield strength for FDM specimens was 56% of the average compressive yield strength for FDM specimens. The mean compressive ultimate and yield strengths (0.2% offset) were found to be the largest for the 90° raster orientation, 34.69 and 29.48 MPa respectively. The 45° raster specimens displayed the smallest mean yield strength, 24.46 MPa, representing 82.97% of the yield strength of the 90° raster specimens. The mean yield strength for the injection molded specimens was 35.50 MPa, a value higher than any of the FDM specimens tested. The 90° raster specimens performed the most closely to the injection molded parts, achieving a mean yield strength that was 83.0% of that of the molded specimens.


Table 5. Compression test results

Although mean ultimate strengths are provided in Table 5, it is typically difficult to pinpoint the instance of rupture in compression loading. Most plastics do not exhibit rapid fracture in compression and the focus is therefore on measuring the compressive yield stress at the point of permanent yield on the stress-strain curve (Riley et al., 2006). Although many of the FDM specimens failed by separation between layers, resulting in two or three distinct pieces, yield stresses were analyzed for consistency and to allow for comparisons with that of the injection molded specimens. As a result, a one-way ANOVA was conducted to compare the effect of raster orientation on mean compressive yield strengths in 0º, 45º, 90º, and +45º/-45º conditions. The test indicated that raster orientation had a significant effect on mean compressive yield strength at the *p* < 0.05 level for the four conditions, F(3, 16) = 31.25, p = 0.0001. Post hoc comparisons using the Tukey HSD test indicated that this significance was limited to and specifically in regard to comparisons with the 45º diagonal condition. The mean yield strength for the 45º raster orientation (24.46 MPa) was significantly different (lower) than that of the other three raster orientations (Table 6), while all other paired comparisons indicated statistically insignificant differences in the mean yield strengths.

Inspection of the failed compression specimens provided additional evidence that the 45º raster specimens were significantly weaker in compression than the other raster orientations. The specimens ultimately separated into two or three pieces, following the displacement of the cylinder's top relative to its bottom, as seen in Figure 10. This distortion occurred as a result of the shearing or sliding along the 45º rasters as the specimens was subjected to an axial compressive load. The other three raster orientations displayed less distortion prior to failure and had mean compressive yield strengths that were significantly larger than that of the 45º raster specimens.


\* Mean difference is significant at the 0.05 level

168 Mechanical Engineering

injection molded specimens displayed tensile strengths greater than that of any of the FDM parts, achieving a mean yield and ultimate tensile strength of 26.95 and 27.12 MPa respectively. The mean UTS achieved by the 0° raster specimens was closest to that of the

A summary of the compression test results, as displayed in Table 5, shows strengths that are higher than those obtained in tension testing. Higher compressive strengths are often observed in polymers, and specifically for bulk ABS materials (Ahn et al., 2002). In this study, the average tensile yield strength for FDM specimens was 56% of the average compressive yield strength for FDM specimens. The mean compressive ultimate and yield strengths (0.2% offset) were found to be the largest for the 90° raster orientation, 34.69 and 29.48 MPa respectively. The 45° raster specimens displayed the smallest mean yield strength, 24.46 MPa, representing 82.97% of the yield strength of the 90° raster specimens. The mean yield strength for the injection molded specimens was 35.50 MPa, a value higher than any of the FDM specimens tested. The 90° raster specimens performed the most closely to the injection molded parts, achieving a mean yield strength that was 83.0% of that of the

Longitudinal (0°) 28.83, 1.16 32.32, 0.58 402.64, 3.64 Diagonal (45°) 24.46, 0.30 33.43, 0.20 417.20 10.06 Transverse (90°) 29.48, 0.75 34.69, 0.99 382.21, 10.31 Default (+45°/-45°) 28.14, 0.64 34.57, 0.86 410.44, 11.23

Although mean ultimate strengths are provided in Table 5, it is typically difficult to pinpoint the instance of rupture in compression loading. Most plastics do not exhibit rapid fracture in compression and the focus is therefore on measuring the compressive yield stress at the point of permanent yield on the stress-strain curve (Riley et al., 2006). Although many of the FDM specimens failed by separation between layers, resulting in two or three distinct

*Mean Ultimate Strength (MPa), Std Dev* 

*Mean Effective Modulus* 

injection molded specimens, representing 94.8% of its value.

Fig. 9. SEM image of air voids seen on fractured 0º raster specimen

**3.2 Compression testing** 

molded specimens.

*Raster Orientation Mean Yield Strength* 

Table 5. Compression test results

*(MPa), Std Dev* 

Table 6. Post hoc Tukey HSD multiple comparisons of mean yield compressive strengths

Fig. 10. Photos of failed 45º raster specimens under compression loading

Anisotropic Mechanical Properties of ABS Parts Fabricated by Fused Deposition Modelling 171

Source DF SS MS F P

Post hoc analysis further indicated a significant difference between all paired mean comparisons other than that of the 45º raster condition (21.3 MPa) in comparison to the 90º raster condition (20.8 MPa). These flexural strength results further confirm that the raster orientation of the FDM specimens contributes to directionally dependent performance. The specimen fracture patterns for the 45º and the 90° specimens were similar to those described for the tensile testing. In contract, the 0º and the +45°/-45° specimens never fractured during

Examination of the fracture surfaces of those specimens that broke into two pieces, i.e. the 45° raster specimens and the 90° specimens, revealed that failure initiated on the side of the part that was under tension loading. As fracture began, the specimen initially remained together by unbroken fibers on portion of it that was in compression. Crack propagation along load direction was erratic and not uniform. This is apparent in Figure 12, which displays clusters of fibers that have bent and then ruptured individually in a

Fig. 12. SEM image of the fracture surface of a 45° raster specimen after flexural loading

Specimens with 0° raster orientation will have fibers that are able to offer more resistance to bending because they are parallel to the bending plane. There is more fiber length over which the load can be distributed. As the raster angle increases to 45° or 90°, the fiber inclination relative to the plane of bending produces rasters with smaller lengths. This results in a net decrease in the ability of the specimen to resist the load. This effect is observed in Figure 13 where the 90° raster specimen shows little evidence of bending. The bottom of the specimen shows a large flat area initially affected by the failure of several

Error 16 32.22 2.01

three-point bend testing, but retained some degree of permanent deformation.

Total 19 614.77

Table 8. One-way ANOVA results for flexural testing

Angle 3 582.55 194.18 96.44 0.0001

Raster

catastrophically brittle manner.
