A wrung stack of gauge blocks can be slid into a clearance gap as a Go check, and the absence of a fit at the next-larger size confirms a No-Go, mirroring the way machinists use the same blocks to set sine plates and bore gages [S3][S4].
The catch is the 10:1 rule of metrology: the reference used to verify a dimension should resolve one decimal place finer than the tolerance on the workpiece, so Grade 2 blocks with a typical ±0.0001 in (2.5 μm) deviation only support verifying clearances of roughly ±0.001 in or coarser [S3][S6].
What a gauge block actually is, and why that matters for gap work
NIST defines a gauge block as a length standard having flat, parallel opposing surfaces, and the published length is the interferometric length measured against the wavelength of light at 20 °C (68 °F), 101,325 Pa, 1,333 Pa water vapor, and 0.03% CO₂ [S1][S2].
Because the two faces are flat and parallel, a wrung stack behaves like a single rigid reference of the same nominal thickness as the sum of the individual blocks, which is exactly the geometry a clearance gap needs for a comparison-style Go/No-Go check [S1][S4]. The stack is also the main means of length standardization used by industry for calibrating micrometers, sine bars, calipers, and dial indicators in inspection roles [S4].
The 10:1 rule and grade selection for clearance checks
A shop-grade (Grade 2) steel gauge block carries a length tolerance on the order of ±0.0001 in for blocks up to 1 in, while Grade 0 blocks tighten that to roughly ±2 μin (0.05 μm) and Grade 00 calibration-grade sets are specified in the single-digit microinch range [S3][S6].
Applying the 10:1 rule, those tolerances only let you verify features that are 10× looser, so a Grade 2 stack of ±0.0001 in is appropriate for checking a ±0.001 in clearance, but inadequate for a ±0.0002 in press-fit transition, where a Grade 0.5 or Grade 00 stack and a calibrated report become mandatory [S3]. As one shop-floor respondent put it, unless you have the actual cert numbers, the "blocks used for comparator setup" introduce an error vector you cannot ignore, which is why the 1975 obsoleting of Grade B did not actually remove economy-grade sets from the market [S3].
Building a stack for a gap: the wringing and minimum-block rules
Two opposed, ultra-flat faces cling together by surface tension in a process called wringing; the resulting stack has the same nominal height as the sum of the stamped values, with each wrung joint contributing a wring film of about 25 nanometers (0.98 μin) that the manufacturer already subtracts from the nominal length [S2][S4].
Best practice is to use the fewest blocks possible: a .638 in target is built as a .500 plus a .138 rather than a four-block .200 / .149 / .151 / .138 stack, because every wrung joint adds parallelism and length uncertainty even though the wring film is sub-micron [S2][S4]. For clearance-gap work that stack is then paired with a thin oil film on steel blocks to protect the faces; ceramic (SiO₂-based) and tungsten carbide (WC) blocks skip the rust concern and the glove requirement, which is why calibration-grade sets are commonly ceramic or carbide while shop-floor sets remain hardened steel [S2][S4].
Comparison: gauge-block stack vs. feeler, pin, and air gages for the same job
On four shop-floor decision criteria, a wrung stack of gauge blocks is the highest-accuracy option but the slowest, while feeler stock is fastest but lowest accuracy, and pin/Class Z pin gages sit between them on both axes: [S3]
- Accuracy: Grade 0/00 stack at single-digit μin beats Class Z pin gages at ±0.0001 in and easily beats feeler stock, which carries ±0.001 in or worse on individual leaves [S3][S4][S6].
- Speed: feeler leaves and pin gages are picked from a tray; a wrung stack needs cleaning, wringing, and re-verification, so the stack loses on cycle time.
- Gap-shape compatibility: a stack is a flat, parallel reference, so it works for rectangular slots, keyways, and large bore clearances; round pin gages are mandatory for circular bores and feeler leaves can fold or shave when forced.
- Cost and reuse: a 112-piece metric set in a wooden case runs the highest upfront cost, but each block is a long-life standard that is also used for caliper and micrometer calibration, so the cost is amortized across every inspection job in the shop [S2][S4].
Use cases and failure modes for the stack-as-gage trick
For a one-off assembly fit, a machinist will wring a stack to the nominal clearance, try it in the gap, and try the next-larger stack; if the larger one will not enter, the gap is between the two values, with the resolution set by the difference between adjacent block sizes in the set, typically 0.001 in or 0.01 mm [S3][S4]. Hobby and home-shop machinists commonly use the same stack as a depth stop on a mill quill, as a Go/No-Go for splines and keyways, and as a height reference under a test indicator on a surface plate, all of which are accepted shop uses even though the last one is hard on the blocks [S3].
Failure modes show up fast: a hand-warmed stack or a hand-warmed part can drift by several tenths of a mil because steel expands at roughly 6 μin/in per °F, and a stack that has not been wrung flat will rock in the gap and give a false No-Go reading [S2][S5]. The NIST reference conditions exist precisely because the published length is only valid at 20 °C (68 °F), 101,325 Pa, 1,333 Pa water vapor, and 0.03% CO₂, and the temperature term is the only one that physically changes the block length, while the others only matter because the primary measurement is an interferometric comparison against the wavelength of light [S1][S2].
Verifying the stack before you trust the gap reading
You cannot definitively verify a wrung stack with a typical shop caliper or even a typical shop micrometer, because both are too coarse to confirm the parallelism of the top face against the bottom face of the stack [S5]. The correct reference for stack verification is an electro-mechanical gauge-block comparator with opposing measuring styli, which NIST describes as the way to read parallelism of the two faces rather than just total length [S1].
If a comparator is not available, the practical floor is to use the stack immediately after wringing, on a surface plate that is itself at 20 °C, and to limit the stack to a clearance check rather than a precision length transfer; for any higher-stakes decision the block set should be sent out with a cal date, because an out-of-cal shop set is a known source of false Go/No-Go calls in assembly fit work [S1][S3][S5].
For a process engineer spec'ing the next job, two trackable signals are worth watching in 2026: the practical drift between indicated and actual stack height under hand-warming, which multiple shop-floor reports put at 0.0005 in on 10 in bores during a 5 °F coolant swing, and the continued availability of Grade B and "economy" sets that the 1975 standard obsoleted but that suppliers still ship as inspection-grade product [S3][S5]. Picking the stack up to check a clearance gap is fine; trusting an uncalibrated economy set to do it is the most common way this method produces a wrong answer.
The underlying component specifications are covered under check valve, swing check valve, and gauge block.
Background reading: Cable-Extension LVDT vs Bare LVDT for Long-Stroke Position Feedback.