Quality control over grading is important to ensure that products are able to deliver the performance that designers expect. This applies equally well to appearance grading and structural grading. In-grade testing of structural products can be used to verify structural properties.
Timber producers produce their own quality control systems and manuals and in some cases seek endorsement for their quality system under ISO 9002 or AS3902. This is not a requirement for every producer, but many of the larger producers see the value in having checks and balances on their quality systems to ensure that their grading is as effective as possible.
In some cases, third party auditing of the quality control systems can be used to issue a product certification. This gives potential purchasers confidence that the quality control and assurance programs are in-place and functioning well.
The main process affected by quality control techniques is the grading operation. In-grade testing is the ultimate test of the effectiveness of any grading operation. There are two types of quality control that can be applied:
• Process Control – checks that the process is being conducted in the way that will produce material of consistent quality. Typical of process control is the testing and training of graders used in the visual grading process . For machine stress grading, the calibration checks, consistency checks and verification using calibration sticks are all methods of implementing process control.
• Product checking – looks at the output from the process and establishes whether or not the properties of the product are consistent and within the specification for the product. It picks up things that may not be detected in process control, such as variations in raw materials. This primarily uses the process of “in-grade testing”.
Product checking is part of any manufacturing process. This is used for all timber products including plywood, glulam and manufactured items such as I-beams. There are three areas in which checking may be incorporated in product checking for quality control:
• Every piece can be checked by a different grading method. Eg after machine stress grading, timber can be checked using visual grading criteria. This will pick up problems in the grading process. (Note that it does not generally work to use the same grading technique for the check as was used in the production grading.)
• Shift checking is the process of testing one physical property a number of times during each shift. Invariably for structural timber, this is the MoE or stiffness. In this way, the timber is not broken, so can still be marketed, but the stiffness can be used to check that the grading is able to produce the product expected. The results are aggregated to give a picture of long-term drift in properties as production is continuing.
• Periodically (typically once per year) a sample may be collected and subjected to more strenuous testing. Usually two properties would be determined for each piece, ie, MoE (stiffness) and bending strength. This gives a better check on the properties of the entire population produced by that mill.
The results of the product checking can be used to track any slow drift in properties of the mill product. If this is observed, then steps can be taken to correct the processes that have caused the drift.
Machine Stress Grading QC
The machine stress grading process requires that quality checks be built into the operation of the machine so that its grading is uniform over time, and that the correlation between grading parameter and the properties of the product remains valid. These checks are on the performance of the machine, which can be found by running calibration sticks of known properties through the machine, and on the performance of the output, by periodically performing destructive tests on samples of the graded timber.
Quality control gives some indication of the success of the grading, and it is incorporated in the design of timber elements through the use of different capacity factors for different levels of control.
Normal quality control measures are appropriate for most normal structures. However, there are some structures that may require much more stringent quality control. This can be set in the product specification.
• For example, in the exhibition hall for the Sydney Showgrounds at Homebush Bay, each finger joint in the laminated beams used as the structural skeleton for the dome, was proof tested prior to lamination of the beams. This provision was included in the specification for the timber and contributed to the cost of the structure. However, it gave the client and the designers an assurance that the product would perform as expected.
• Similarly in appearance products, special jobs may require special checking. For example, panels used as lining in the Australian Parliament house had specifications that included very tight tolerances on moisture content, and very tight colour requirements. Colour matching cards were used at all stages of production to ensure that only products with acceptable appearance would be seasoned and machined.
These unusually demanding jobs ended up paying a premium for the timber and the quality control that was used, but the product was exactly what was required.
In moving towards a reliability based code in limit states format there has been a need to establish the true strength distribution of graded timber for some of the more commonly used structural species. This work is known as “in-grade testing” and involves the following:
Sampling the stocks of commercially available timber in a particular species and grade. (In some cases, this may require shipping timber from different parts of the country to ensure that the sample is as representative as possible.
The full sized lengths of commercially available timber are tested in a manner similar to the loading expected in service for the product. (For example bending strength is derived using a third point loading over a span of 18 times the depth of the timber. The bending moment diagram is similar in shape and scale to that produced by a uniformly distributed load, and the span-to-depth ratio is characteristic of bending members.)
The data from the tests is modified to allow for the sampling and can be used to represent the strength of the entire population of timber in that species and grade. The design properties can be chosen so that they are less than the characteristic values indicated by the tests. In many cases, different sized cross-sections will have been tested enabling a size factor to be derived.
The procedures for performing in-grade tests and appropriately presenting the data have been detailed in <AS/NZS 4063 - 1992>.
In-grade testing has been completed for the group of species known as Australian pine and Victorian Ash. With Australian pine, the data has been used to create a number of grades, where the properties have been derived from the in-grade test data. These classes are known as MGP10, MGP12, MGP15 with the “10” referring to 10,000 MPa, a rounded characteristic modulus of elasticity. The other MGP grades follow the same convention. All timber produced by MSG is required to be in-grade tested.
In-grade testing involves the careful sampling of commercial timber in the sizes and grades produced, and the testing of hundreds of pieces of timber to get as close as possible to the true strength distribution of a grade. The true strength distribution of timber in the marketplace is used as a starting point to assign or confirm a set of design values. These values can be tied to an existing grade or can be product specific.
Also, in-grade testing can be used to periodically test timber to ensure that the product is delivering the properties implicit in the grade stamp applied. This gives a higher level of reliability to the product properties.
In-grade Testing to Determine Design Strength
The following outlines the process of in-grade testing used to determine the design strengths of timber:
- Large numbers of pieces of commercial sized structural timber of a single species and grade are tested. The resulting distribution tends to follow a log-normal distribution rather than the normal distribution as for the small clear specimens. The fifth percentile strength from this distribution is used for all subsequent calculations.
- The fifth percentile of the strength distribution is modified to account for the sampling process. The result of the modification is an estimate of the fifth percentile strength of the entire population of the single size and grade that was sampled and tested. The modification algorithm allows for both sample size and the variance of the properties from the tests, and gives 75% confidence that the data reflects the 5%ile of the population.
- This figure is the characteristic strength of the timber, which is used as a basis for setting the design value.
Some structural timber products have most or all of their major characteristic properties (e.g. bending strength, stiffness, compression and tension) based on in-grade testing, whilst some secondary properties, such as compression perpendicular to grain, are based on strength grouping.
In-grade Testing to Determine Design Strength
A similar process can be used to monitor properties of continuing production of timber products. The process can be used to monitor strength or stiffness. It is illustrated here for strength.
- Sampling over a long period of time gives material that represents the long-term production of that product. Testing of this sample of material uses the same test method outlined in AS/NZS4063.
- The 5%ile of the test results is calculated.
- This value is compared with an acceptance criterion given in AS/NZS4490. Where the 5%ile is above the acceptance criterion, the production is said to match the design value within expected sampling error. Where the 5%ile is below the acceptance criterion, then there is doubt about the properties of the material and corrective action may be required.
In-grade testing is a versatile concept that has acceptance around the world. Different test protocols are used in different parts of the world, and analysis methods differ slightly too. However, the basic principle remains the same:
• Take a large sample of full sized commercial timber, and test it fairly.
• Use strength values representing the low end of the strength distribution as the basis for design.