Beech sawing patterns: cant sawing wins on volume, quarter sawing on radial share

Introduction

Beech (Fagus sylvatica L.) is one of the most abundant broadleaved species in Central Europe, and in Slovakia it is a mainstay of the hardwood sawmilling sector. It is also a demanding raw material. Beech logs frequently carry red false heartwood, reaction wood and dote, and their cross-sections are often far from round. Because a sawmill buys logs by volume but sells boards by volume and quality, the way a log is broken down on the headrig has a direct effect on revenue.

A recent paper from the Department of Wood Technology at the Technical University in Zvolen (Slovakia) looks at exactly this decision. The authors compared three log-sawing patterns on beech - through-and-through (live) sawing, cant sawing and quarter sawing - and asked two questions. How much lumber does each pattern recover from a log (the quantitative yield)? And how much of that lumber is radial, that is, cut with growth rings roughly perpendicular to the board face (the qualitative yield)? Radial boards are valued for dimensional stability, so the share of radial material is a sensible proxy for quality in a species that warps and checks readily.

This article summarises the study for sawmill managers and production planners. It is worth stating at the outset that the experiment is very small - six logs in total - so the numbers are best read as an illustration of trade-offs rather than as benchmarks to plug into a business case.

Research methodology

Material. The study used six beech logs, labelled V1 to V6, taken from the Hronska Dubrava forest area (457 m above sea level), managed by the Technical University Forestry Enterprise in Zvolen. The logs were chosen on the day of felling. Mid-diameters ranged from 43 to 60.5 cm, with log volumes between 0.56 and 1.10 m3. The tables in the paper give log lengths of roughly 4.0 to 4.2 m, although the running text mentions "approximately six metres"; the tables appear to be the reliable figure.

Sawing patterns. Two logs were allocated to each pattern:

Equipment and target. All logs were processed on a horizontal band saw (MEBOR 1000 HTZ). The target lumber thickness was 33 mm. After sawing, each board was cleaned and marked.

Measurement. Thickness, width and length were recorded for every board. Dimensions and volumes followed the European standard STN EN 1309-1 for measuring round timber and sawn timber. Width and thickness were taken with a digital caliper (accuracy stated as 0.03 mm) and length with a tape. Because boards can be edged, semi-edged or unedged, the width was determined by the method appropriate to each case, and rounded edges were handled with the permissible allowances in the standard.

Yield definitions. Lumber was sorted into radial and tangential groups. Quantitative yield is the total lumber volume (radial plus tangential) divided by log volume. Qualitative yield is the share of radial lumber. The paper does not describe a separate defect-based grading scheme beyond this radial/tangential classification, and it does not give per-log defect counts.

Statistics. The authors used one-way analysis of variance (ANOVA) at a 0.05 significance level, supported by box plots, to test whether the sawing pattern influenced quantitative yield, qualitative yield and the volumes of radial and tangential lumber.

Key findings

Per-log results. The table below reproduces the main per-log values reported in the paper (rounded as published).

Pattern averages. Averaging the two logs per pattern gives roughly the following picture.

Quantity. Cant sawing delivered the highest recovery. The best single result in the whole experiment was log V4, a large cant-sawn log, at about 84% - a figure the abstract highlights. The cant average across both logs was 78.5%. Quarter sawing recovered the least lumber, at about 61% on average. Through-and-through sawing fell in between. The authors link the weaker through-and-through result partly to log taper, which causes extra waste when a parallel face has to be produced on a tapered stem, and they note that cant sawing produces boards of larger usable dimensions.

Quality. The pattern was reversed for radial content. Quarter sawing produced by far the highest share of radial lumber, averaging about 55% (62.69% and 47.86% for the two logs). Through-and-through sawing gave around 20%, and cant sawing only about 7%. In other words, the pattern that maximised volume was the one that least favoured the dimensionally stable, radially oriented boards.

Statistical tests. The ANOVA indicated a significant effect of the pattern on qualitative yield (p = 0.0105) and on radial lumber volume (p = 0.0065). The effect on quantitative yield was not significant (p = 0.1721). The authors therefore report that their hypothesis about quality was confirmed while the hypothesis about quantity was not, even though the descriptive differences in quantity look sizeable.

Other observations. The authors discuss several influences on the outcome: log curvature (noted for V3 and V6), ovality (V5 and V6), the characteristics of the horizontal band saw, and the presence of juvenile wood and pith in the log centre. They recommend removing the juvenile core because boards from near the pith tend to crack. They also point out that red false heartwood, reaction wood and dote are typical beech problems, and cite literature associating yield with red heartwood, but this experiment does not quantify the effect of those defects on any individual log.

Comparison with the literature. The paper compares its numbers with other studies. It cites an oak live-sawing result of 57.29% quantitative yield with 87.89% radial share, a beech study with 56.8% (through-and-through) and 60.5% (cant and quarter) yields, and a Masson pine study with a 82.7% yield for a hexagon pattern against 53.3% for live sawing. These are different species, equipment and definitions, so the comparison is indicative only.

Practical applications for sawmills

From the paper.

• The sawing pattern is a real lever on both yield and the radial share of the product. A mill cannot maximise volume and radial content with the same pattern.
• In this experiment cant sawing gave the best volume recovery, and quarter sawing gave the best radial share.
• Taper, curvature and ovality visibly affect what a given pattern can deliver, so log sorting by form matters.
• Removing the juvenile core and pith area is advised for beech because of cracking.

Portal interpretation (not claims of the paper).

• Which pattern is "best" depends on the price list. If radial or quarter-sawn beech earns a premium that exceeds the lost volume, quarter sawing can be justified; if the mill sells mainly commodity boards by volume, cant sawing looks more attractive. A simple margin calculation with the mill's own prices would settle this.
• A mixed strategy may be sensible: allocating larger, straighter logs to cant sawing and keeping selected logs for quarter sawing when an order for stable radial boards exists. The paper's allocation of the largest log to cant sawing is consistent with that idea, though it does not test it as a strategy.
• Anyone replicating the comparison in-house should record bark, defects (red heartwood, reaction wood) and log form for every log, because the paper does not provide these data and they will confound a small trial.

Limitations

• Very small sample. Only six logs were sawn, two per pattern. Statistical tests on two observations per group have little power and should be treated cautiously; the significant p-values are best seen as suggestive.
• Confounding with log size. The logs differed in diameter (43 to 60.5 cm), and the text indicates that allocation was linked to size. Pattern and log size are therefore partly mixed up. The largest log, V4, delivered the headline 84% and was cant sawn, so part of its advantage may come from its size rather than the pattern.
• Inconsistencies in the text. The paper gives log length as about six metres in the methods but about 4 m in the tables, and the described log selection rule does not fully match the table of allocations. The "17.3% higher" yield statement in the abstract does not reproduce simply from the tables as a relative difference; it appears to correspond to a gap in percentage points against the other logs. Some reported yields differ slightly from a direct ratio of lumber to log volume. Readers should rely on the per-log tables and verify before reusing any figure.
• Defects and grading. There is no defect-based grading and no per-log defect data, even though red false heartwood and reaction wood are central to the motivation. Qualitative yield only measures the radial share.
• Allowances. Edge allowances under the standard were used to compute volumes, which can raise reported yields compared with measuring only fully formed boards.
• No economic analysis. The study does not translate yield differences into value.
• No dedicated limitations section. The authors suggest only that the work could be extended to other hardwoods with similar defects, such as oak and hornbeam.

Conclusions

The study shows, on a small beech sample, that the choice between through-and-through, cant and quarter sawing shifts the balance between lumber volume and radial lumber share. Cant sawing gave the highest recovery (best log about 84%, pattern mean 78.5%), while quarter sawing gave the highest radial share (mean about 55%). Only the effect on quality reached statistical significance in the authors' ANOVA. Given the sample of six logs, the work is useful mainly as a pointer to the trade-off and as a template for a larger trial with controlled log sizes, recorded defects and a price-based evaluation. Mills processing beech should test the trade-off against their own price lists rather than adopt the figures directly.

Original research

• Authors: Peter Vilkovsky, Ivan Klement, Tatiana Vilkovska
• Title: The Impact of the Log-Sawing Patterns on the Quantitative and Qualitative Yield of Beech Timber
• Journal: Applied Sciences (MDPI), 2023, volume 13, issue 14, article 8262
• Year: 2023
• DOI: 10.3390/app13148262
• Link: https://www.mdpi.com/2076-3417/13/14/8262
• Licence: Creative Commons Attribution (CC BY); the MDPI page states CC BY 4.0 in its footer

This article is an independent summary prepared by the portal; the original paper should be consulted for full details and exact figures.