During the past 30 years I’ve had the opportunity to measure quite a few manufactured parts, machine tools, fabrications, and large vessels. I’ve also reviewed hundreds of metrology surveys done by technicians. Unfortunately, on more than one occasion I’ve seen data that were scaled incorrectly or not scaled at all. The results of incorrect thermal scaling may cause inaccurate data, showing faulty parts even if the parts are not faulty. It is essential to understand the purpose and importance of thermal scaling.
Unless otherwise noted, a drawing or model is scaled at 1, or 68°F. In quality labs, temperatures are consistent even though that may not always be the case. Measuring different parts in varying temperatures may cause the material to expand and contract due to thermal change. This change can be calculated by using the coefficient of thermal expansion (CTE). Most measurement software packages include the capability to apply the CTE calculation to the measurement file based on a particular material type.
For a laser tracker, a weather station monitors air temperature and air pressure. This information generates a correction factor for the laser wavelength to maintain a consistent measurement. Also included is a material sensor, which allows the operator to acquire the temperature of the part. However, this does not compensate for thermal growth. The laser tracker provides measurements relative to the part, at the current temperature and at the time of the measurement.
As a training instructor I often use the following analogy. Imagine a piece of aluminum that measures 48 in. in a 68°F controlled inspection lab. The laser tracker will measure the bar at 48 in., but placing it in an environment that is 90°F would cause the bar to experience a thermal linear growth of approximately 0.013 in. The measurement will now read the bar at 48.013 in. Obviously, when compared to its actual design length, the measurement would be incorrect by 0.013 in. The correct approach would be to compensate the measured data based on the CTE of aluminum and report the part as meeting the design length.
For volumetric checks on CNC machine tools, I also use the laser tracker. Again, thermal compensation must be considered in the application files loaded on the CNC control. Most standards, like ASME 5.54 and ISO 230, require compensation data to be scaled at 68°F. The material’s CTE may differ depending on what material the machine tool builder uses in his machine construction. It is ideal to use the machine tool’s major material composition as the guide for determining CTE. Usually this material is steel. The steel in the machine will expand and contract with the parts. Some manufactures use the material being used to construct the linear scale as their metric for CTE.
The challenge is to decide how to approach the process because many industries are machining large composite parts that have a much lower CTE. This causes the parts to expand and contract at a different rate than the machine. The best approach is to calibrate the machine at its normal operating temperature and not to apply a CTE effect to the measured data. The differences will be accounted for in the data collected through compensation tables. Another good approach is to calibrate the machine to 68°F and scale the part programs based on the material type being measured.
Another dilemma is how to account for the temperature of a part as it expands and contracts during measurement. It is important to understand the amount of thermal growth variation that may occur with respect to the material part type and to determine the thermal changes that may impact the tolerances for the part. There are two options for this situation.
The first is to add a new instrument to your measurement job to compensate for a part’s temperate difference. Once the measurement is complete, scale each of the different instrument measurements based on the part’s temperature recorded during that particular instrument’s setup. The second option is to average the part’s temperature at the end of the measurement survey and use that temperature average for calculating the correct scale for your part.
Based on my knowledge and experience, these approaches prove to be accurate, although though I’ve heard a lot of different thoughts and opinions. I welcome any feedback on these approaches.