Quality Scan: Temperature Influences Measurement Accuracy
During my career, I've noticed that many companies, and the actual users of measurement equipment, do not understand the effect that temperature has on accurate measurement.With today's tighter measurement requirements, one must eliminate as many variables as possible—including temperature.
For instance, some of my customers had the idea their measurement equipment had to be kept in an air-conditioned room. When I walked into the room where they kept the equipment, it felt like a meat locker! There are a number of reasons why this is not good.The primary reason, though, is that the temperature is still very much out of control. It's a common mistake to interpret air conditioning as a means of temperature control.
Additionally, I've seen measurement-equipment users at companies not take temperature into account at all. One customer I visited could not understand why, on a piece of equipment that I recently installed, their two-inch triple-X ring gage measured too large by 0.0003" (0.0076 mm). To be sure the specs for the measurement were clearly understood, I asked for the specification certification for the ring gage. Once I had the certification, I measured the room temperature where the measurement equipment was located. The reason for the discrepancy in measurement became clear. The ring gage was certified at 68°F (20°C), and the room temperature where the measurement was performed was at 78°F (25.6°C).
Depending on the measurement accuracy required and the material measured, temperature can be a critical variable. Different materials have different coefficients of linear expansion, but materials also have different heat transfer coefficients. Simply put, different materials react by expanding or contracting more rapidly than other materials. Material thickness must also be considered. If you were to take two pieces of aluminum of the same length but different thicknesses, the thinner piece of aluminum will respond to temperature change more rapidly than the thicker piece. This behavior is caused by the greater thermal mass of the thicker component.
The only way to completely eliminate temperature as a variable is to ensure that all manufacturing and inspection are performed at 68°F.
Why 68°F? In most, if not all, standards traceable to NIST the measurements or certifications are performed at 68°F. This temperature is the internationally accepted standard measurement temperature for metrology and measurement- certification organizations such as NIST. It became the standard reference temperature in the International Standard ISO recommendation number 1 for Geometrical Product Specifications in 1951.
Performing all manufacturing and quality control in a temperature-controlled environment may seem, and probably is, an unrealistic objective. But it's the only way temperature can be completely ruled out as a variable in quality control and the manufacturing process.
The best way for most companies to determine an acceptable temperature change in the manufacturing process and in quality control is to review the coefficients of expansion for the specific part material, and the part tolerances.Then, determine an acceptable temperature change, along with an acceptable measurement uncertainty for the parts being manufactured.
For example, a number of companies have determined that an acceptable temperature for manufacturing and quality control is 74°F (23.3°C).They ensure that manufacturing and the QC lab are at the same temperature within an acceptable variation. This approach makes sense, providing the temperature is kept under control within acceptable limits. Doing so eliminates temperature as a variable between manufacturing and quality control.Then, using the known standards for the specific measurement process, accurate measurement can be calculated and adjustments made with regard to temperature differences.
Another approach I've seen is to employ no temperature control in manufacturing, but maintain temperature control in the QC lab. Parts are 'acclimatized' for a period of time prior to inspection, so they are at the same temperature as the quality lab.The main problem with this sort of operation is that one has to estimate machining offsets in the manufacturing process to keep production parts in specification, and then there's the amount of time required to inspect parts, because you must wait for the parts to acclimatize.
One needs to look at the coefficients of expansion and coefficients of heat transfer for the part material as well as the part tolerances, and determine an acceptable temperature-driven measurement uncertainty. Remember, this uncertainty stacks up with that of the process or equipment being used in inspection.
This article was first published in the June 2008 edition of Manufacturing Engineering magazine.