Ask any metrology engineer who has spent a decade chasing sub-micron repeatability, and they'll tell you the base material matters more than most spec sheets let on. A machine can have the best linear encoders and the tightest servo loops in the world, but if the platform underneath it drifts with temperature or rings with every passing forklift, none of that precision survives contact with the real world.
For most of the 20th century, cast iron was the default choice for machine bases, and marble showed up occasionally as a "precision" alternative in lower-cost measuring equipment. Neither holds up well against natural granite once you look closely at the physics.
The problem with cast iron
Cast iron is easy to machine and cheap to source, which is exactly why it became the industry default for so long. But it has three properties that work against precision applications:
It has a relatively high coefficient of thermal expansion, so a base that's flat at 20°C can develop measurable distortion by the time a workshop warms up a few degrees over a shift.
It retains internal stresses from casting that release slowly over years, meaning a "flat" surface plate can quietly warp as it ages - a phenomenon well documented in machine tool literature going back decades.
It's magnetic, which disqualifies it outright for any application near sensitive sensors, MRI-adjacent equipment, or high-precision electromagnetic measurement setups.
Where marble falls short
Marble looks like a reasonable upgrade - it's non-magnetic, it's stable at room temperature, and it has a certain old-world credibility in metrology. The catch is that marble is a metamorphic rock built from calcite, which is significantly softer than the quartz-feldspar structure of granite. Surface plates made from marble wear faster under repeated contact from gauge blocks, probes, and workpieces, and calcite is also more reactive to acidic cleaning agents and coolant residue than granite is. Manufacturers who substitute marble for granite to cut cost are, in effect, trading long-term dimensional stability for a lower unit price - a shortcut that shows up years later as measurement drift nobody can quite explain.
Why granite wins on the numbers
Natural black granite, when properly selected, offers a rare combination of three things at once: dimensional stability, damping, and hardness.
Thermal stability: Granite's coefficient of thermal expansion is significantly lower than cast iron's, and it responds to temperature changes far more slowly due to its density and crystalline structure. Density is actually a meaningful proxy here - high-density black granite in the range of roughly 3,100 kg/m³ resists both thermal drift and the kind of internal micro-porosity that lets moisture in over time.
Vibration damping: Granite's crystalline grain structure absorbs vibration more effectively than the more uniform lattice of cast metal, which is part of why granite bases are the default choice under coordinate measuring machines (CMMs), profile projectors, and laser interferometer setups where even sub-hertz table resonance can show up in the data.
Non-magnetic, chemically inert: Unlike cast iron, granite introduces no magnetic field distortion, and unlike marble, it's far more resistant to acids, oils, and coolants common in machine shop environments.
Long-term flatness retention: Because granite isn't cast or forged, it has no internal casting stress to relax over time. A properly lapped granite surface plate, stored and used correctly, holds its calibrated flatness for years rather than requiring the periodic re-machining that cast iron sometimes does.
This is also why granite surface plate flatness tolerances are written into so many national standards - DIN 876 in Germany, JIS B 7513 in Japan, GGG-P-463c in the US, and China's GB/T 22095 all define grade-specific flatness tolerances (typically Grade 00, 0, 1, and 2) precisely because granite's stability makes those tolerances achievable and, more importantly, keepable over years of shop-floor use.
Not all granite is equal
It's worth noting that granite quality varies enormously by quarry source and grain structure. Denser, finer-grained black granite generally outperforms coarser or lighter-colored stone in both stability and surface finish potential - which is one reason serious metrology suppliers specify density and grain characteristics rather than just calling out "granite" as a generic material. In our own production, every batch of raw stone is checked for density and grain consistency before it ever reaches the CNC or the grinding shop, and finished surface plates are verified on Mahr and Mitutoyo comparators traceable to national calibration institutes before they ship.
What this means in practice
If you're specifying a base for a CMM, an air-bearing XY stage, a laser interferometer setup, or an optical inspection line, the material choice isn't cosmetic - it's the foundation the rest of your accuracy budget sits on. Cast iron still has its place in structural, non-precision applications. Marble should generally be treated with some skepticism when it shows up in equipment marketed as "precision." For anything where sub-micron stability actually matters, granite remains the more defensible engineering choice, and the data backs that up rather than just tradition.
If you're evaluating granite bases or measuring tools for an upcoming project and want to talk through tolerance requirements, grain specifications, or calibration traceability, our engineering team is happy to walk through the specifics with you.