Stiffness Grading Of

STIFFNESS GRADING OF
GREEN ROUGH-SAWN TIMBER

by
Doug Gaunt
Grant Emms
Barry Penellum
John Roper
October 2003

Correspondence to:
Laboratory Manager:
Timber Engineering Laboratory
New Zealand Forest Research Institute Limited Portfolio Manager:
Private Bag 3020
ROTORUA
Ph: (07) 343 5899
Fx: (07) 343 9380

©NEW ZEALAND FOREST RESEARCH INSTITUTE LIMITED 1998
All rights reserved. Unless permitted by contract or law, no part of this work may be reproduced, stored or copied in any form or by any means without the express permission of the NEW ZEALAND FOREST RESEARCH INSTITUTE LIMITED.
IMPORTANT DISCLAIMER. The contents of this publication are not intended to be a substitute for specific specialist advice on any matter and should not be relied on for that purpose. NEW ZEALAND FOREST RESEARCH INSTITUTE LIMITED and its employees shall not be liable on any ground for any loss, damage or liability incurred as a direct or indirect result of any reliance by any person upon information contained, or opinions expressed in this work.

OBJECTIVE

To evaluate in the effectiveness of using an acoustic stiffness grading approach on green rough-sawn timber to predict the dry gauged plank and joist stiffnesses.

MATERIAL

53 pieces of green 100x50 - 4.8m minimum length was selected from Pukepine’s Te Puke sawmill off the green table. The timber selected aimed to cover the visual grades No 1Framing, No 2Framing and box grade. At Forest Research the timber was then cut into 2.35m lengths (giving 106 sticks of approximately ‘stud’ length) to suit the length constraints of the Forest Research drying kiln. All the subsequent testing was thus done of these 2.35m lengths.

TEST METHODS

Each stick of timber was assessed for stiffness at three stages namely: green rough-sawn, kiln dried rough-sawn and finally kiln dried gauged. At each stage the timber was tested for stiffness as follows:
1. Each piece of timber was assigned a unique laboratory number ranging from 211976 to 212081. The numbers were marked on one end so that they could be used a reference position for stiffness and dimension measurements.
2. Measured for bending stiffness in both plank and joist orientations over a span of 2.1m as Figure 1.

Figure 1. Bending test configuration

3. Measured for weight.
4. Measure for width and depth. The rough-sawn timber measurements were done at 400, 800, 1200, 1600 & 2000mm, whereas the gauged timber was only measured at the midpoint (1200mm).
5. Acoustic MoE measurements were made:–

• One end of the timber was struck with a hammer and the resulting sound produced by the timber was recorded with a microphone placed near one end of the timber. This recording was processed using a Fourier transform to show the longitudinal vibration resonant frequencies of the timber.
• The first two of these resonant frequencies were combined with the length of the timber to calculate the average speed of sound along the length of the timber. Timber density was then calculated from weight and dimension measurements. The speed of sound was combined with the timber density to calculate the acoustic MoE, using the formula:
  MoE = (speed of sound)² x density.

6. The timber was kiln dried at 120/70°C for 22 hours followed by steam reconditioning (98/98°C) for four hours in the Forest Research laboratory kiln.

Results

Figures 2 and 3 show the cumulative frequency distributions for the bending stiffness in both the joist and plank directions for the green rough-dawn and kiln dried gauged states. With reference to the joist orientation the study timber can be considered as typical of the nationwide forest resource in that it has a similar average stiffness as the nationwide data set but with slightly fewer low and high stiffness pieces.

Figure 2: Bending longspan stiffness as a joist (LMoEj)

Figure 3: Bending longspan stiffness as a plank (LMoEp)

The stiffnesses in the joist and plank directions as expected are very similar for both the green rough-sawn and kiln dried gauged states.
The following Figures 4 & 5 show the green rough-sawn acoustic MoE vs the kiln dried gauged MoE in both joist and plank directions. This data is based on the average of five width and depth measurements taken for each stick.

Figure 4: Green RS vs KD gauged MoEj using actual dimensions.

Figure 5: Green RS vs KD gauged MoEp using actual dimensions.

Figures 6 & 7 show the green rough-sawn acoustic MoE vs the kiln dried gauged MoE in both joist and plank directions based on the nominal 103x53 green dimensions.

Figure 6: Green RS vs KD gauged MoEj using nominal 103x53 dimensions.

Figure 7: Green RS vs KD gauged MoEp using nominal 103x53 dimensions.

The potential effect on grading accuracy of using the nominal dimensions over actual dimensions is minimal in this case. This observation will potentially simply the grading machine. Figure 8 shows the range in the differences of the acoustic MoE when using average actual dimensions versus nominal dimensions with Figures 9 and 10 showing the range in actual dimensions. The sawing variation of ± 2mm covering around two standard deviations is about the mean was typical of New Zealand sawmills in this market.

Figure 8: Difference in Acoustic MoE with actual and nominal dimensions.

Figure 9: Range in Green rough-sawn depth dimensions.

Figure 10: Range in Green rough-sawn width dimensions.

The following:
Figure 11, shows the range in timber density in terms of nominal density (Oven dry weight/green volume), Green piece density and Kiln dried piece density.
Figure 12, shows the range in green moisture contents.

Figure 11: Range in Timber Densities Oven-dry, Dry & Green moisture conditions.

Figure 12: Range in Green Moisture Contents.

Figure 13: Dry G4S vs KD G4S MoEj using nominal dimensions.

Figure 14: Dry G4S vs KD G4S MoEp using nominal dimensions.

Figure 15: Dry G4S vs KD gauged MoEj using nominal dimensions & 474kg/m3 nominal density.

Figure 16: Dry G4S vs KD gauged MoEp using nominal dimensions & 474kg/m3 nominal density.

Comparing Figures 13, 14, 15 & 16 shows they the effect of assuming a density for the entire sample as opposed to determining the individual densities. For improved grading accuracy it is beneficial to measure individual densities.
The following Figures 17, 18, 19 & 20 show the differences in joist and plank stiffness from Green RS to Dry RS to Dry gauged using the test machine stiffness data. In the joist direction (Figure 17) the R2 value is 0.91 this compares with 0.88 in (Figure 4), ie.. the sonic stiffness is comparable with the test machine in terms of accuracy.

Figure 17: Green vs Dry G4S LMoEj by test machine.

Figure 18: Green vs Dry G4S LMoEp by test machine.

Figure 19: Green vs Dry RS LMoEj by test machine.

Figure 20: Green vs Dry RS LMoEp by test machine.

The lower R² values in the plank direction are interesting and are thus further discussed in Appendix A.

Other Acoustic Stiffness Data

• Acoustic grading was used in the recent Forest Research Machine Stress Grading multi-client. This study used kiln dried, gauged radiata pine 90x45 in 3 & 4.8m lengths and 190x45 in 4.8m length. The acoustic grading techniques were some of the best performers when used to grade for structural grades.
• Acoustic MoE’s were taken on a sample of 90x45 kiln dried, gauged, Douglas fir these compared very well with bending stiffness.

Conclusions

1. Accuracy prediction of kiln dried gauged stiffness is possible from acoustic MoE measurements on the green rough-sawn timber.
2. Using nominal green rough-sawn dimensions over the actual dimensions has a very small negative effect on the acoustic MoE measurements.

3. It is recommended that individual piece densities be used in the calculation of acoustic MoE, however using an assumed overall sample density may provide an acceptable solution in certain cases.

4. Forest Research has sufficient data and experience to warrant the development of a commercial acoustic stiffness grading system that can be applied to:

• Green or dry timber .
• Rough-sawn or gauged timber.
• Variable length lumber including finger-joint shooks.
• A range of timber species.

Currently there is potential for the development of a New Zealand based acoustic grading technology. However Forest Research must act with some urgency if it wants to hold some leadership in this area.
Appendix A:

The following Figures A1, A2 & A3 plot the joist vs plank stiffness (by test machine) for the green rough-sawn state (Figure A1), the dry rough-sawn state (Figure A2) and the dry gauged state (Figure A3). It is commonly accepted that the 1:1 ratio (slope of the regression line) would occur between plank and joist stiffnesses, this appears to be the case in the Figures A1 & A3 but not Figure A2. The Dry rough-sawn state shows the plank stiffnesses to be 1.12 times (12% higher) than the joist stiffnesses, also the correlation coefficient is lower (data spread is greater) than the other states.

Figure A1: Joist vs Plank Green RS

Figure A2: Joist vs Plank Dry RS.

Figure A3: Joist vs Plank Dry Gauged.

This anomaly as yet has not been explained however the following factors may well ply a role in this difference.

• In the drying kiln the timber was stacked as a plank on fillets butted together on the joist faces (Figure A4) thus only the plank faces were subjected to the full transverse airflow.
  Potentially the plank surfaces properties have been changed in some way to increase their stiffnesses. After gauging (3mm removed from each face) the difference between plank and joist stiffness disappeared.

Figure A4: Timber orientation and airflow within the kiln.

To generate an increase in plank stiffness 12% higher than the joist stiffness the outer 3mm thick plank surfaces need to have stiffness 60% higher than the core as shown in Figure A5.

Figure A5: Timber section showing suggested differential MoE