Machinability of plantation-grown Eucalyptus globoidea timber

ABSTRACT

Eucalyptus globoidea is an emerging plantation species. Its naturally durable heartwood has the potential to be used for solid wood outdoor products. Good machinability of a timber is essential for solid wood processing but was unknown for E. globoidea.

The ASTM D1666 standard was used to assess the machinability of E. globoidea and compared to that of Pinus radiata, the dominating resource for the local wood processing industry and well-known for its good machinability. This study showed that E. globoidea machined equally well or better than P. radiata in planing, boring, mortising, grooving, edging, and turning. Sanding E. globoidea (723 kg/m³) was not possible at the prescribed settings, whereas a smooth finish was obtained for the lower density P. radiata (461 kg/m³). Reducing the sanding depth resulted in a satisfactory sanding finish. Defects that determined the machine scores were identified. When boring, mortising and grooving E. globoidea timber, most defects were chipping caused by the tool exiting the piece. When edging, most care needs to be taken at the corner.

The comparable machinability of E. globoidea and P. radiata according to ASTM D1666 indicated that the well-established P. radiata wood processing industry should be able to process E. globoidea.

KEYWORDS:

• ASTM D1666
• Pinus radiata
• white stringybark
• wood processing

1. Introduction

High-value markets for naturally durable timbers are well-established globally, but these timber species are rare and typically found in tropical forests (Scheffer and Morrell 1998; Bootle 2005). Supplies of durable timbers are dwindling with strong consumer demand, and supply is often sourced from unsustainably managed natural stands (Nellemann, 2012; Bösch, 2021). To combat the decline of many hardwood species, legislation to prohibit the trade of illegally logged timber (Polo Villanueva et al., 2023) and to protect overharvested resources are now more common. For example, Western Australia's jarrah (Eucalyptus marginata) forests will be closed to commercial harvesting by 2024 (Conservation and Parks Commission, 2022). Preservative-treated, sustainably harvested timber can substitute such naturally durable timbers, but at the cost of converting the biodegradable material wood into toxic waste (Altaner, 2022).

The New Zealand Drylands Forest Initiative (NZDFI) identified the opportunity to sustainably grow ground durable timbers for agricultural sectors (Millen et al., 2020). E. globoidea has been identified as a species featuring ground-durable heartwood (AS5604 2005), which can be grown sustainably in New Zealand's climate. After over a decade of development, the first improved E. globoidea became available for sale in 2021 from a breeding programme. E. globoidea timber is not only ground-durable, but the higher density timber also features excellent mechanical properties (Poynton, 1979; Bootle, 2005), opening outdoor joinery and structural engineered wood products markets (Jones et al., 2010; Guo & Altaner, 2018). The machinability of a timber, largely the ease of obtaining a smooth surface, is a critical characteristic for solid wood processing. The American Society for Testing and Materials (ASTM) has developed a standard to quantify the machinability of timber (ASTM D1666-17). Tests are designed to reflect the variability of the surface finishes for specific machining processes, which are visually graded. Literature on the machinability of timbers has been reviewed recently (Sofuoglu & Kurtoglu, 2014; Gupta et al., 2019). There is little correlation of machinability characteristics between species (Merhar et al., 2020); however, there are indications that some machinability parameters are influenced by within-species density variation (Hernandez et al., 2001). The machinability of some higher-density eucalypts has been reported (Belleville et al., 2016) but shown to be species-dependent. The machinability of E. globoidea is unknown.

This study aimed to give New Zealand's radiata pine-focused wood processing industry confidence to utilise E. globoidea timber. To account for the inherently qualitative data of machinability characteristics (as summarised by Merhar et al., 2020), the results are best compared against timbers with known machinability. P. radiata was chosen as it dominates New Zealand's timber industry. It machines well and is widely used for structural and non-structural applications (Young, 1988). It is highly permeable and typically treated with copper chrome arsenic (CCA) for outdoor uses (Altaner, 2022). Further, the locations of machining defects were noted, allowing to recommend improvements of the machining processes.

2. Materials and methods

2.1. Material

Eucalyptus globoidea boards originated from 31 trees felled at age 28 years old on Banks Peninsula (43°47’32.0“S; 172°50’20.8“E), New Zealand as part of a taper and volume study (Boczniewicz et al., 2022). Logs were then live sawn into 30 mm thick boards using a mobile horizontal band saw and ripped into 110 mm wide heartwood boards without pith. The boards were air-dried indoors for three years. The least defective and most approximately flat-sawn boards were selected from the available material. Deviating from ASTM D1666-17, specimens with live knots not larger than 15 mm in width were included due to material constraints. As a control, thirteen flat-sawn, kiln-dried, rough-sawn, clears-grade Pinus radiata boards (5,200 mm × 150 mm × 25 mm) were purchased locally.

From the rough-sawn boards; samples for planing and sanding; grooving, edging, boring, mortising; and turning; as well as density and moisture content determination, were prepared according to ASTM D1666-17 and equilibrated to ∼12% MC in a climate-controlled room at 65% relative humidity and 20°C.

2.2. Grading

The ASTM D1666-17 standard distinguishes five visually assigned grades for each machined surface. To ensure consistency, only one person graded all samples. The grades were interpreted as follows:

Grade 1: defect-free pieces, where no remediation (sanding) would be required to finish the product to a high standard.

Grade 2: defects that could be remediated by light sanding.

Grade 3: remediation would require significant sanding or chiselling, resulting in slightly rounded edges or altered dimensions.

Grade 4: as Grade 3, except remediation will result in deep-rounded edges or visibly altered dimensions.

Grade 5: No defect remediation is possible to achieve a high-quality surface finish.

Machining defects associated with knots were excluded from the surface quality analysis, particularly for sanding and planing, as wood near knots is at higher risk of torn and fuzzy grain (Öhman et al., 2016). The worst defect determined the grade of a piece. The machine score for a tool was defined as the percentage of the combined Grades 1 and 2 or Grades 1–3 as defined in (ASTM D1666-17).

2.3. Planing

Equilibrised boards were planed in random order regardless of species to remove the effect of tool wear with an SCM S630 thicknesser operated at 4500 RPM (revolutions per minute) with two new high-speed steel knives installed at a cutting angle of 30°. The feeding speed was 11.25 m/min, resulting in 0.8 cuts/mm. Samples were planed in two passes, with a first pass taking off 1 mm to remove artefacts of previous machining and a second pass taking off a further 1.6 mm of the same face. The surface quality was measured after the second pass.

2.4. Sanding

After planing, the samples were sanded using a Hafco Woodmaster DS-25 Dual Drum Sander fitted with 80-grit and 120-grit aluminium oxide sandpaper. The feed speed was 6.1 m/min, and 1.2 mm was sanded off one face for assessment. Each species was machined independently on fresh sandpaper.

2.5. Boring

A Dyco 12 mm bench drill press was used with a 25 mm Bosch Forstner drill bit featuring a high-speed steel cutting edge to assess boring. The targeted feed speed was 0.5 m/min. The drill was run at 1200 RPM, deviating from the standard-specified 3600 RPM, as the samples burned at a higher speed. All samples were randomised. Two holes were bored into each sample, and the order was reversed in the second run to remove the effect of tool wear.

For all boreholes, three data points were graded independently: entry, the edge where the tool entered the wood; hole, the plane inside the borehole; and exit, the edge where the tool exited the wood. The grade of a data point on a board was the worst grade of the two holes per board. Typical grades are depicted in Figure 1.

Figure 1. Boring grades: a) Grade 1, b) Grade 2, c) Grade 3, and d) Grade 4.

2.6. Shaping

An Ascent Pro computer numerical control (CNC) machine was used with three router bits featuring high-speed steel cutting edges: a 6.4 mm straight bit, a 10 mm rebating bit, and an 8 mm rounding bit. The CNC machine was run with a 4240 mm/min feed speed at 10,000 RPM. The samples were machined in random order to reduce to effect of tool wear. First, a J-shaped groove running parallel and perpendicular to the grain was cut with the 6.4 mm straight bit, followed by a rebate along the edge of the specimen, first parallel, then perpendicular to the grain, using the 10 mm rebating bit (Supplementary Figure 1). Finally, the lower step of the rebate was rounded with the 8 mm rounding bit.

Four data points were graded for each groove: Side-grain inside the groove; the corner inside the groove; the end grain inside the groove; and the edge where the tool exited the sample. Three data points were graded for each edge: the side-grain surface, the corner, and the end-grain surface.

2.7. Mortising

A Luxcut TM-3VSL turret mill was used with a 13 mm custom-made high-speed cutting steel mortising chisel and a single spiral auger centre. The targeted feed speed was 0.23 m/min. Holes were mortised at 1200 RPM as the machine could not perform this task safely at the (ASTM D1666-17) specified 3600 RPM. The samples were mortised in a randomised order, and the two mortised holes per sample were machined in reversed order to minimise the effect of tool wear.

Four data points were graded for each mortise hole independently: entry, the edge where the tool entered the wood; side-grain, the side-grain plane inside the borehole; end-grain, the end-grain plane inside the borehole; and exit: the edge where the tool exited the wood. The grade of a data point on a board was the worst grade of the two holes per board.

2.8. Turning

A Luxcut Z330X1000 lathe was used for turning with a custom-made knife profiled as specified in (ASTM D1666-17) featuring a high-speed steel cutting edge. The targeted feed speed was 0.02 m/min. The lathe was run at 2000 RPM. Samples were randomised to reduce the effect of tool wear.

Four data points were graded independently per sample at the angled, flat, concave, and convex part.

2.9. Data analysis

Regression modelling and t-tests were performed with R (R Core Team, 2022).

3. Results and discussion

Air-dry densities were significantly higher in E. globoidea than in P. radiata (Table 1). The mean air-dry density for E. globoidea heartwood boards cut from 28-year-old trees was 723.3 kg/m³ with a coefficient of variation (CV) of 9%. This was lower than the reported air-dry density for Australian old-growth trees (820-900 kg/m³) (Bootle, 2005) but consistent with the reported basic density of ∼540 kg/m³ for 25-year-old trees grown in Rotoehu Forest, New Zealand (Jones et al., 2010). The lower growing temperature might contribute to the lower density of New Zealand-grown timbers, as warmer climates were reported to increase wood density of E. grandis (Thomas et al. 2007).

Table 1. Summary statistics of P. radiata and E. globoidea machinability data. CV: coefficient of variation. Significant differences between species on the 0.05 level are highlighted in bold.

		Eucalyptus globoidea n = 37				Pinus radiata n = 49
		Mean	CV	Min	Max	Mean	CV	Min	Max	P-value
Density (kg/m^3)		723.3	9%	553.7	872.8	460.9	9%	372.4	548.3	< 2.2e-16
Moisture content (%)		12.8	3%	12.1	13.5	12.0	7%	10.2	13.6	1.834e-08
Planing	0.8 cuts/mm	1.9	46%	1	4	1.9	39%	1	4	0.8515
Sanding	1.2 mm	5.0	0%	5	5	1.0	0%	1	1	< 2.2e-16
Boring	Entry	1.0	0%	1	1	1.5	38%	1	3	6.283e-08
	Hole	1.2	32%	1	2	3.1	23%	2	5	< 2.2e-16
	Exit	3.3	32%	2	5	4.1	15%	3	5	3.305e-05
	Overall	3.3	32%	2	5	4.1	15%	3	5	2.401e-05
Mortising	Entry	2.5	32%	1	4	3.4	27%	2	5	4.535e-06
	Side grain	2.5	22%	2	4	3.0	16%	2	4	0.0002647
	End grain	2.4	21%	2	3	4.4	11%	3	5	< 2.2e-16
	Exit	3.5	16%	3	5	4.8	9%	3	5	< 2.2e-16
	Overall	3.6	16%	3	5	4.8	8%	4	5	< 2.2e-16
Grooving	Side grain	1.4	40%	1	3	2.6	20%	1	3	< 2.2e-16
	Corner	1.5	35%	1	2	2.6	19%	2	3	< 2.2e-16
	End grain	2.0	8%	2	3	3.0	18%	2	4	< 2.2e-16
	Exit	2.9	24%	1	4	3.0	23%	2	5	0.5387
	Overall	3.0	22%	2	4	3.3	16%	3	5	0.005251
Edging	Side grain	1.2	59%	1	5	2.1	25%	1	3	8.592e-09
	Corner	2.9	42%	1	5	2.6	34%	1	5	0.1267
	End grain	2.1	27%	2	5	2.9	8%	2	3	5.208e-10
	Overall	3.1	38%	2	5	3.1	17%	2	5	0.8414
Turning	Wedge	2.2	83%	1	5	2.5	78%	1	5	0.5069
	Convex	2.8	51%	1	5	2.9	58%	1	5	0.8597
	Flat	2.7	58%	1	5	3.1	49%	1	5	0.1953
	Concave	2.6	60%	1	5	2.8	63%	1	5	0.7284
	Overall	2.9	47%	2	5	3.2	48%	1	5	0.4370

The mean equilibrium moisture contents at 65% relative humidity and 20°C were with 12.8% for E. globoidea and 12% for P. radiata boards (12%) in the expected range Table 1. The variation was low (E. globoidea CV = 3% and P. radiata CV = 7%), indicating sufficient equilibration.

3.1. Planing

No statistically significant difference was found between the surface finishes of E. globoidea and P. radiata after planing to 1.6 mm depth at 0.8 knife cuts/mm (Table 1). Both wood species planed well (mean visual grade of 1.9 (Table 1) and a machine score of ∼80% (Table 2)). This was consistent with P. radiata's previously reported good planing characteristics (Young, 1988). Optimisation of the machine settings as described in ASTM D1666-17 will not necessarily achieve significantly improved planing quality as characteristics were already good (Merhar et al., 2020).

Table 2. Overall machine scores for E. globoidea and P. radiata according to (ASTM, 2017).

Eucalyptus globoidea
n = 37	Grade 1	Grade 2	Grade 3	Machine Score
Planing	15	13	0	76%
Sanding^a	0	0	0	0%
Boring	0	10	0	27%
Mortising	0	0	17	46%
Grooving	0	8	0	22%
Edging	0	16	0	43%
Turning	0	25	0	68%
Pinus radiata
n = 49	Grade 1	Grade 2	Grade 3	Machine Score
Planing	16	25	0	84%
Sanding	46	0	0	100%
Boring	0	0	0	0%
Mortising	0	0	0	0%
Grooving	0	0	0	0%
Edging	0	1	0	2%
Turning	5	18	0	47%

^a Excellent sanding finish can be achieved at 0.5 mm instead of 1.2 mm depth.

3.2. Sanding

Consistent with previous reports, all P. radiata boards were sanded to a Grade 1 finish (Young, 1988). E. globoidea boards were almost impossible to sand with the machine settings prescribed in the standard (ASTM, 2017), as the sandpaper was clogged up after sanding less than a metre, causing surface charring. Similar observations were made for other dense and durable eucalypts (Belleville et al., 2016).

The E. globoidea boards were significantly denser than the P. radiata boards (Table 1), implying more material to be removed by the sandpaper at a given time resulting in more friction-induced heat. Reducing the sanding depth from 1.2 mm to 0.5 mm resulted in Grade 1 finishes for subsequent E. globoidea boards, consistent with what was reported for other high-density eucalypts (Belleville et al., 2016). Consequently, E. globoidea can be sanded to the same excellent surface quality as P. radiata if surface defects are not deeper than 0.5 mm or by repeating the procedure. Apart from significantly larger density, E. globoidea heartwood contains extractives (Iyiola et al., 2022), which could have contributed to the clogging-up of the sandpaper.

3.3. Boring

Boring finishes were significantly better for E. globoidea than P. radiata (Tables 1 and 2). The predominate defect occurred during the exit of the drill (Tables 1 and 3). In contrast to E. globoidea, where the inside of the hole was machined well, frequent defects were found for P. radiata at this location (Tables 1 and 3). The difficulty with boring P. radiata was unexpected considering the published rating of 4 out of 5 (Young, 1988) and potentially due to more stringent grading or deviation from the standard regarding RPM and borer type. An effect of borer type on the drilling grades of eucalypts was reported (Belleville et al., 2016).

Table 3. Machinability scores for selected processing features of E. globoidea and P. radiata. Problematic processing features are highlighted bold.

		Eucalyptus globoidea n = 37	Pinus radiata n = 49
Boring	Entry	100%	96%
	Hole	100%	18%
	Exit	27%	0%
Mortising	Entry	84%	61%
	Side	97%	90%
	End	100%	0%
	Exit	54%	2%
Grooving	Side	97%	37%
	Corner	100%	37%
	End	97%	12%
	Exit	22%	20%
Edging	Side	97%	78%
	Corner	47%	55%
	End	95%	6%
Turning	Wedge	70%	63%
	Convex	68%	61%
	Flat	70%	47%
	Concave	70%	63%

3.4. Mortising

Similar to boring, E. globoidea produced significantly better-quality mortising finishes than P. radiata (Tables 1 and 2). The known difficulty in mortising P. radiata (Young, 1988; Kotlarewski et al., 2019) was reflected in the low mean grade (4.8 out of 5) and machine machining score (0%). As for boring, the worst defect occurred for both species when the tool exited the piece (Tables 1 and 3). For P. radiata, additionally, the end-grain inside of the hole was problematic.

3.5. Grooving

Grooving finishes were significantly better in E. globoidea (mean grade 3.0, machine score 22%) than in P. radiata (mean grade 3.3, machine score 0%) (Tables 1 and 2). While the interior side- and end-grain faces machined well for E. globoidea, defects mainly occurred when the tool exited the piece (Tables 1 and 3). Tool exit was also problematic for P. radiata, which, in line with mortising, showed additional difficulties with the interior end-grain face of the groove. Therefore, E. globoidea ‘exit’ defects should be avoidable by employing backing pieces (Belleville et al., 2016) or starting the grooving process from both ends and meeting in the interior.

3.6. Edging

While the mean edging grade for the two species was the same (Table 1), the machine score was significantly better for E. globoidea (Table 2). The main reason for the worse performance of P. radiata was the end-grain surface, which had a machine score of 6% compared to the 95% for E. globoidea (Table 3). The sharp corner had a machine score of ∼50% for both species, the limiting factor for E. globoidea (Figure 2). Improved corner finishes were reported for high-density eucalypts by cutting against the feed direction (Belleville et al., 2016).

Figure 2. Example of typical Grade 5 defect on corner caused by edging E. globoidea.

3.7. Turning

No statistically significant difference in the turning grades was found between the two species (Table 1). The higher machine scores of 70% for E. globoidea compared to the 50-60% of P. radiata were encouraging (Tables 2 and 3), particularly as P. radiata is typically known to turn well (Young, 1988). The different observations for average turning grade and machine score of the two species indicated that they differed in the proportion of ‘extreme’ (i.e. acceptable and completely failed) samples. Turning in this assessment resulted in frequent complete torsion failure of the samples in the lathe, likely caused by a suboptimal clamping method.

4. Conclusion

This study has confirmed that, when tested according to ASTM D1666, New Zealand grown E. globoidea heartwood machines equally well or better than P. radiata, except for sanding, which can be remedied by reducing the sanding depth. Consequently, solid wood processors working with P. radiata will not need to invest in specialty machinery when working with E. globoidea, if no unforeseen challenges arise during industrial scale processing. Forest growers will have the confidence to invest in planting this emerging species, knowing it can be machined with standard equipment. The generally good machinability of E. globoidea was consistent with reports for other plantation-grown higher-density eucalypts (Belleville et al. 2016). In contrast, plantation-grown non-durable, medium-density eucalypts (E. nitens and E. globulus) were more challenging to machine than P. radiata, as reported by Kotlarewski et al. (2019).

Optimisation of the machining parameters can be expected to improve the machine scores. For example, grooving having two entry points and no exit should avoid most defects.

Author contributions

Conceptualisation: Clemens Altaner; Methodology: Hamish Scown, Clemens Altaner, Hyungsuk Lim; Formal analysis and investigation: Hamish Scown, Clemens Altaner; Writing – original draft preparation: Hamish Scown; Writing – review and editing: Clemens Altaner, Hamish Scown, Hyungsuk Lim; Funding acquisition: Clemens Altaner; Resources: Clemens Altaner, Hyungsuk Lim; Supervision: Clemens Altaner, Hyungsuk Lim.

Acknowledgments

The authors would like to appreciate the technical assistance of Gert Hendriks and Monika Sharma (School of Forestry, University of Canterbury).

Disclosure statement

The authors have no relevant financial interests to disclose. Clemens Altaner is Science Team Leader of the New Zealand Dryland Forests Initiative (NZDFI).

Additional information

Funding

This work has been financially supported by the New Zealand Ministry of Business, Innovation and Employment (MBIE) Specialty Wood Products Partnership (FFRX1501).