1 Introduction

The surface roughness of wood is one of the important parameters that can be used to operate the automatic wood cutting machinery, and also to evaluate the quality of products. With the development of computer technology, automatic measurement methods of the machined surface roughness have been proposed. In recent years, new non-contact methods have been studied to measure the roughness of machined surfaces. For example, ultrasonic sensor was used to measuring surface roughness(Blessing et al., 1989). AE(acoustic emission) was used to measure the surface roughness of wood during the automatic control of the router feed speed (Cyra et al., 1996). The optical profilometer was used to measure the surface roughness of MDF (medium density fiberboard) (Lemaster et al., 1996). Wood surface characteristics were used to measure 3-Dimensional Roughness parameter (Costes et al., 2001). Laser scanning was used to measure the surface roughness of coated wood (Yuko et al., 2001). Laser optical imaging technique was used to measure the surface roughness on dry veneer and plywood samples (James et al., 1992).

An automatic roughness measurement system of machined wood demands the following characteristics in general: high speed processing of data, steady and accurate measurement of signal, good liner and ease of computer processing, which have been realized by the researches mentioned above. But besides those functions, quantitatively and on-line high speed measurement is especially very important to wood machining, which can improve the quality of production and raise the efficiency of processing(Faust, 1987). And this paper describes a method using the laser displacement sensor to quantitatively and on-line high speed(6 m·min^-1) measure the surface roughness of wood after the planing operation(Lemaster et al., 1982).

2 Materials and methods 2.1 Materials

The specimens for the experiment were Douglas Fir (Pseudotsuga menziesii) grown in Japan, Hinok (Chamaecyparis), Padauk (Pterocarpus) and Beech (Fagus crenata) with moisture contents between 10%~12%, and specific gravity of 0.53, 0.48, 0.67 and 0.63 respectively. The specimens were in lengths of 150 mm and widths of 80 mm. The radial section was planed to measure surface roughness. The cutting conditions used in the experiment are as shown in Tab. 1.

2.2 Calibration experiment

Calibration is necessary for the laser displacement meter before use. In order to get an accurate measurement, the laser sensor was fixed and the light from the laser sensor was emitted in parallel direction to the feed direction of the specimens, as shown in Fig. 1. At first, the distance was adjusted by moving CNC working table from the laser sensor to the target, in order to obtain a position where the output signal voltage was zero. From this position, the specimens were moved around until 500 μm in every 25 μm, and the output signal data were recorded with a voltage meter.

2.3 Surface roughness measurement

Two methods were used to measure the specimens' surface roughness. One of them was the laser displacement meter, and the other was the stylus-type surface measuring instrument. The experimental setup of the laser displacement meter consists of a laser sensor, a FFT analyzer, a computer and a feed table as shown in Fig. 2. The laser sensor was fixed in a vertical direction to emit the laser light to the wood surface in vertical direction. The resolution of the laser displacement meter is 0.5 μm. The specimens were fixed on the table, which moved at a speed of 4 m·min^-1. The displacement signal data from the laser displacement sensor were recorded in the FFT analyzer at a rate of 1 024 points per 0.8 second. The signals were then transmitted to the computer and the ten-point mean roughness (Rz) was calculated.

A diamond tracer point with a tip radius of 0.04 mm was moved on the surface at 12 mm lengths for the measurement by the stylus-type surface roughness measuring instrument. The surface roughness of the same place was also measured by the laser displacement meter. The roughness profile was derived from the primary profile using a 2RC filer, and cutoff wavelength was 2.5 mm. Ten-point mean roughness (Rz) also was calculated (Yoo et al., 1990).

3 Results and discussion 3.1 Calibration

One of the characteristics of the laser displacement sensor is its ideal lineation, from which 1 μm of input displacement produces the 2 mV output voltages(Keyence, 1994). But the relationship varies with differences of wood materials.

Fig. 3 shows the relationship between the calibration coefficient C1 and the output voltage from the laser displacement sensor for the specimens investigated. The data for Fig. 3 were obtained under the range of 0 to 500 μm input displacement (range of surface roughness). C1 value was the ratio of the output signal voltage to the input displacement (before calibration C1 of ideal line equal to 2, C1=mV/μm) (Keyence, 1994). As Douglas Fir is a coniferous tree, it's early and late woods have large difference in the color and structure compared to the broadleaf tree. Because of the influence of early and late woods, the output signal voltages were different although the distance from the laser sensor to the early and late woods was equal. The range of C1 values were from 2.02 to 2.04 in the early wood, meanwhile the range of C1 values were from 1.94 to 1.97 in the late wood. Therefore, the linear relationship between the output signal voltage and the displacement can be considered to be good when measuring the early or late wood areas only. In this experiment, the two averages of C1 for the early and late wood were used to calculate the surface roughness of Douglas Fir, respectively. They were 2.03 for the early wood and 1.96 for the late wood. Hinoki is the coniferous tree. Beech and Padauk are broadleaf trees. As those samples are early and late woods have no obvious difference, they were found that C1 values were almost same for the early and late woods. The average value C1 of Hinoki specimen was 2.01 and Beech specimen was 1.94 and Padauk specimen was 1.97 for the calculated surface roughness(Jouaneh, 1989).

3.2 Softwood Douglas and Hinoki

Fig. 4 shows the output voltage signals from the laser displacement sensor for the early and late woods of Douglas under different cutting conditions. Fig. 4a and Fig. 4b shows the output voltage signals for the early and late woods of the specimenⅠ, and Fig. 4c and Fig. 4d shows the output voltage signals for the early and late woods of specimen Ⅱ. It was found that the behavior of output voltage signal values were different in the early and late woods, and the changes of output voltage signal for the early wood were larger than that for the late wood. This phenomenon was possibly caused by the difference of surface roughness between the early and late woods, which was caused by the difference in the hardness and the color of the early and late woods, which affects the machined surface. But in most situations, the hardness is the more important factor.

The surface roughness Rz was calculated by the ten-point mean values of output voltage using different calibration coefficients C1, as mentioned above. Tab. 2 shows the Rz values measured by the laser displacement sensor and the stylus instrument for Douglas under different cutting conditions. Rz values were not equal for the two measurement methods with contact and without contact. Compared to the Rz values measured by the stylus instrument, the Rz values measured by laser displacement sensor were larger. Based on the Rz values measured by the stylus instrument, a new calibration coefficient C2 (C2=S/L, S:Rz values of stylus; L: Rz values of laser) was calculated to change the value from the laser displacement sensor to be equal or nearly equal of the value Rz from the stylus instrument. C2 was calculated to be 0.90 for the early and late woods. The results of the Rz from the laser displacement sensor measurement hardly had any difference with these results from the stylus instrument measurement, as shown in Tab. 2.

Tab. 3 shows the Rz values measured by the laser displacement sensor and the stylus instrument for Hinoki under different cutting conditions. Because the analysis method and measurement result of Hinoki are similar to Douglas, the detailed instruction of Hinoki is omitted.

3.3 Hardwood Beech and Padauk

Fig. 5 shows the output voltage signals from the laser displacement sensor for Beech under different cutting conditions. Fig. 5a, Fig. 5b and Fig. 5c were the output voltage signals for the cutting conditions Ⅰ, Ⅱand Ⅲ of Beech, respectively. The measurements could not be carried out for the early or late wood separately, because the early and late woods of Beech have no distinct difference. The Rz values measured by the laser displacement sensor and Rz values obtained from the stylus instrument are shown in Tab. 4. From Tab. 4, it can be seen that the Rz values from the laser displacement sensor were still larger than the values from the stylus instrument. Further, the Rz values from both the laser and stylus measurement methods increased with the feed speed, due to the increase in the feed per knife.

The calibration coefficient C2 of the laser displacement sensor was calculated by the same method as described above, and C2 was 0.9 to equate the Rz values from laser displacement sensor for Beech. It can be observed from Tab. 4, that the Rz values from the laser displacement sensor by using coefficient C2 were almost the same to the Rz values from the stylus instrument.

The Rz values of Padauk measured by the laser displacement sensor and Rz values obtained from the stylus instrument are shown in Tab. 5. Because the analysis method and measurement result of Padauk are similar to Beech, the detailed instruction of Padauk is omitted.

The different Rz values from the laser displacement sensor, which are usually larger than the values from the stylus instrument can be due to several reasons. The diameter of the laser light spot was different from the stylus. Further, the spot area of the laser light was 45 μm×20 μm, while the diameter of the stylus was 40 μm. Therefore, the area of laser light spot was smaller when measuring Rz of the surface roughness and the resolution of the laser displacement sensor was relatively higher than the resolution of the stylus instrument. Hence, the Rz values from laser displacement sensor were larger (James et al., 1992; Yuko et al., 2001).

4 Conclusions

As the linearity of the laser displacement sensor varies for the wood specimens, as in the case of early and late woods, calibration for linearity is required when the laser displacement sensor is used for surface roughness measurement of woods.

The ten point mean roughness measured by the lasers displacement sensor is larger than those measured by the stylus instrument. However, it is possible to convert the Rz from the laser displacement sensor to the Rz from the stylus instrument by the proportion coefficient between them.

It is possible to on-line measure the surface roughness of woods using the laser displacement sensor for the high processing speed such as 6 m·min^-1.