A granite surface plate’s flatness limit depends on its size, grade and whether the measurement is local or covers the entire work surface. Mitutoyo’s bulletin lists common rectangular plate tolerances in accordance with ASME B89.3.7-2013. The chart below preserves both flatness measures for Grades AA, A and B, so a local repeat-reading limit is not mistaken for the overall plate limit.

Surface plate flatness tolerance grades

Overall flatness and local flatness

Overall flatness is the distance between two parallel planes that contain all measured points across the work surface. Local flatness, also called repeat reading, applies to a smaller area and therefore has a tighter tolerance. The values in the dataset are in microinches; multiply by 0.0254 to convert a microinch value to micrometres. For example, a 36-by-48-inch Grade A plate has a 400 µin overall limit and a 70 µin local limit in Mitutoyo’s table.

Grade AA is the highest accuracy of the three grades, Grade A is commonly used for inspection areas, and Grade B is a shop-floor grade. A grade designation does not replace a calibration report. The 2013 standard’s decision rule also considers test uncertainty: Mitutoyo notes a test uncertainty ratio of at least 2 for Grade AA and at least 4 for Grades A and B when stating conformance.

Worked size selection

For a 24-by-36-inch plate, first match both dimensions rather than choosing only the longer side. The Grade AA overall limit is 100 µin and the local limit is 45 µin. Grade A is 200 µin overall and 70 µin local. Grade B is 400 µin overall and 120 µin local. If the inspection requirement concerns the entire plate, compare against the overall column. If it concerns a small region checked with a repeat-reading gage, use the local column and its procedure.

Metric sizes are not included in this selected dataset. Do not infer metric limits by rounding the nearest inch-size row; consult the governing standard or the calibration provider for the exact dimension and grade. Temperature gradients, support placement, loading, granite behavior and measurement method can affect the result. The plate should be installed on its specified supports and evaluated using a suitable calibrated instrument.

Common mistakes

First, do not compare surface roughness Ra with flatness; they describe different surface properties. Second, do not use the local flatness limit as the whole-plate limit. Third, do not claim a plate meets grade based on instrument resolution alone. The decision depends on uncertainty and the conformity rule. Fourth, do not treat an old calibration result as proof of the plate’s current condition after damage or relocation.

These are common-size rectangular granite plate limits, not a complete replacement for ASME B89.3.7 or the calibration certificate. Source: Mitutoyo’s technical bulletin on accuracy and calibration.

The dimensions in the source table are width by length, not a diagonal measurement. Do not substitute a nearby plate size unless the standard or calibration provider allows that interpolation. For metric work, this report does not list rounded conversions: the standard’s metric values and exact plate-size categories should be consulted directly. A unit conversion can change the displayed precision, so retain enough digits for the calibration decision and do not round a limit more generously than the source.

Mounting and use also matter. A granite plate can sag or distort if it is supported at the wrong points, subjected to a large temperature gradient, or loaded beyond its intended use. Clean the surface before measuring, let the plate and instrument stabilize, and follow the calibration method. If a plate is chipped, moved, relapped or exposed to unusual loading, arrange an appropriate recalibration before relying on the old certificate.