Precision Shims and Spacers: Stamped to ±0.01 mm

In thirty years of supplying precision stamped parts, I have watched more assemblies fail from a missing or wrong shim than from a broken one. The danger is rarely a single part. It is the slow creep of accumulated error. You spec a bearing clearance of 0.05 mm. Your housing is machined to ±0.02 mm. The shaft runs to ±0.015 mm. Now you drop in three stamped shims to set preload, and if each one lands on the wrong side of its ±0.01 mm thickness tolerance, the stack builds the wrong way and your clearance is gone before the unit ever sees a load. Flatness is the quiet killer. A shim that reads perfect on a micrometer can still rock under a bolt because one corner sits 0.03 mm proud of the rest, and that rocking converts directly into vibration, fretting, and fastener loosening. Get the thickness and the flatness wrong together and no amount of torque saves you.

That scenario is why engineers who have been burned once start treating shims and spacers as precision components instead of hardware-store commodities. A stamped shim is a deliberate, engineered gap. When you ask for ±0.01 mm across a 0.5 mm part, you are holding the tolerance band to 2% of the nominal thickness. That is not a casual request. It is a process decision, a material decision, and a measurement decision all at once, and the three of them have to agree or the part is scrap. This guide walks all three: what shims actually do in a BOM, which metals suit which job, how a ±0.01 mm band is held and proven on a real line, what flatness, burrs, and coatings add to the equation, and what the whole thing costs. The capability numbers referenced throughout come from our own floor: 21 presses up to 110 tons, piloted positioning of ±0.005 mm, strip from 0.05 to 3.0 mm thick and up to 650 mm wide, with a QC lab running CMM, optical measurement, and in-line vision under IATF 16949:2016.

What We Are Actually Making

Precision stamped shims and spacers fall into three jobs in a real BOM. Adjustment shims set a running clearance or preload and get selected or stacked to dial in a dimension that machining alone cannot hold economically. Insulating washers break the electrical path between a fastener and a conductive housing while still carrying the full clamp load. Locating spacers hold two faces a fixed distance apart and keep a sub-assembly positioned in X, Y, or Z. Same family of parts, completely different failure modes. The adjustment shim fails when the thickness drifts. The insulator fails when a burr punches through the coating or the flatness lets edge contact short the circuit. The locator fails when the bore or OD walks off-center and the whole stack skews out of square.

Three less common variants are worth knowing because they show up in the RFQ under different names. Laminated shims are stacks of thin layers bonded or packaged together, peeled to reach the target thickness on the assembly line - they trade shim cost for assembly flexibility. Tapered shims are ground or formed with a wedge profile to correct angular misalignment between two faces. Matched sets are shims measured, sorted, and shipped together so the stack total, not the individual parts, hits the print tolerance. If your drawing calls any of these, the inspection plan changes completely, because the unit of measurement is the set, not the piece.

Picking the Metal: Stainless, Brass, and Phosphor Bronze

Material choice drives everything downstream - springback, tool wear, flatness recovery, and how the part behaves once it is clamped. Three alloys cover most of what crosses my bench.

Stainless steel 301 and 304. Austenitic stainless is the default for corrosion-exposed shims. Cold-rolled 304 in the 1/2 hard condition runs about 180 HV and 520-720 MPa tensile. It springbacks harder than brass, so a die designed by thickness alone will throw parts long. We compensate by over-bending and by tuning the press tonnage and dwell. 301 full hard climbs toward 400 HV and takes a permanent set better, which helps flatness but beats up the tool steel faster and shortens die life.

Brass C260 (cartridge brass). At thin gauge, brass is the easy one to stamp. Half-hard C260 sits around 100-150 HV with 45-55% elongation, so it flows into the die without cracking and recovers flat better than stainless. The cost is lower strength and poorer wear resistance. For an adjustment shim that sees no sliding, brass is often the smartest dollar on the table.

Phosphor bronze C510 and C521. This is the spring shim and the wear shim. C510 hard reaches 250-320 HV with excellent fatigue life and natural anti-galling against steel. C521 adds a touch more strength and a finer grain. When a spacer has to survive millions of load cycles without taking a permanent set, phosphor bronze is the call, and we usually stress-relieve it after blanking to kill the residual curl. The spring stamping guide covers the same material logic for the broader spring part family.

Two more alloys deserve a mention even though they are less common. Aluminum 5052 gives a light, corrosion-tolerant shim for weight-sensitive assemblies where the load is modest; it forms easily and stays flat, but it cannot carry the clamp load of steel at the same thickness and it gall under repeated adjustment. Beryllium copper C17200, aged after forming, is the shim of last resort for duty that combines spring force, conductivity, and wear - an electrical contact spacer inside a switch or connector. Both are niche, but both are routine when the application demands them.

Material	Typical hardness	Flatness after stamping	Spring behavior	Best shim job
304 stainless (1/2 hard)	~180 HV	Good; springback to manage	Moderate	Corrosion-exposed adjustment shims
301 stainless (full hard)	~400 HV	Good with coining pass	Strong	High-load shims, wear resistance
Brass C260 (1/2 hard)	100-150 HV	Excellent	Weak	Low-cost adjustment shims, insulators
Phosphor bronze C510/C521	250-320 HV	Good after stress relief	Strong, fatigue-resistant	Spring shims, cyclic-load spacers
Aluminum 5052	~60-90 HV	Excellent	Weak	Lightweight spacers, low load
Beryllium copper C17200	200-380 HV (aged)	Good	Very strong	Conductive, spring-duty spacers

The 0.05 to 3.0 mm Thickness Window

Our standard strip window is 0.10 mm up to 3.0 mm, with strip widths to 650 mm. Below 0.10 mm the part stops being a stamping and starts being a foil - you are fighting wrinkling, oil trapping, and feed instability more than tolerances, and the coil itself becomes the tolerance problem. Above about 2.0 mm, plain stamping leaves too much shear break and a ragged edge to hold ±0.01 mm repeatably; those parts usually move to fine blanking or, for prototypes, waterjet or laser cutting.

Inside the window, thin stock (0.10-0.30 mm) is where the ±0.01 mm tolerance is hardest, because the absolute band is a huge fraction of the material and the strip itself needs to be supplied to a tight gauge. We typically buy strip to ±0.005 mm so the incoming material does not eat the whole budget before we cut a single part. A shim cannot be more accurate than the coil it came from, and a coil that wanders in gauge shows up as a wandering mean on the SPC chart, not as a one-off bad part. For the general rules behind thin-part tolerances, the precision stamping tolerances guide is the broader reference.

Tolerance Targets: Thickness, Flatness, and Geometry

A shim drawing carries five tolerance callouts, and they are not equal. Thickness is the headline, but the others decide whether the part works when clamped.

Parameter	Typical target	Why it matters	Common trap
Thickness	±0.01 mm on 0.10-0.30 mm stock	Sets the gap or preload directly	Coil gauge drift eats the budget
Flatness	≤ 0.05 mm TIR thin; ≤ 0.1 mm per 25 mm thick	Rocking under bolt = vibration, fretting	Micrometer passes, granite fails
Burr height	< 10% of thickness where specified	Pierces coatings, galls mating faces, crack starter	Checked visually, not measured
OD / ID concentricity	Concentric within 0.02-0.05 mm typical	Off-center bore skews the stack	Only OD checked on incoming
Coating thickness	Netted out of the bare dimension	Counts against the final gap	Coated part measured, not bare stock

The drawing also needs the edge condition stated: a shim with a sharp exit burr on the face that slides during adjustment will gall the mating part and wreck the very clearance you were setting. Put the burr-side note and the burr-height limit on the print, because "deburr" alone is not a specification.

Holding ±0.01 mm, and Proving It with Cpk

A drawing tolerance is a promise. Cpk is the evidence. For a ±0.01 mm band on a 0.5 mm shim, the spec width is 0.02 mm total. To run that at Cpk ≥ 1.33 - the floor most automotive and sensor customers demand - the process standard deviation σ must stay under about 0.0025 mm, because Cpk = (spec width / 2) / (3σ) and 0.01 / (3 × 0.0025) = 1.33. Push to Cpk ≥ 1.67, the medical and aerospace preference, and σ drops below 0.002 mm. That is two microns of process spread, measured part to part, every shift.

You get there with ground die sections, in-die thickness gauging or post-process air-gauge sorting, temperature-controlled gauging at 20 °C, and disciplined SPC. We log a thickness reading per cavity every few hundred strokes and chart it; a drifting mean shows up long before it breaches the band, and the chart is what the customer's quality auditor actually reads. The process system behind this - progressive dies on presses up to 110 tons with ±0.005 mm pilot positioning, in-die sensing, and SPC on critical features - is documented under our quality and capability page.

Measuring a Shim: Tools, Temperature, and the 10-to-1 Rule

Nobody talks about the measurement problem, and it is the reason shim jobs go sideways. A ±0.01 mm tolerance is a 0.02 mm band. To measure inside that band with any confidence you need instruments whose uncertainty is a fraction of it - the old rule of thumb is that measurement uncertainty should stay at or below a tenth of the tolerance band, which for this job means sub-micron resolution. A hand micrometer reading to 0.001 mm is the starting point, not the finish line; operator technique, anvil wear, and temperature all eat into that reading.

On our floor the measurement chain is: air gauges for thickness on high-volume runs (they read without contact and without operator squeeze), CMM or optical measurement for OD, ID, and concentricity, and a granite flat with a light probe for flatness. Everything is gauged in a temperature-controlled room at 20 °C, because a 0.5 mm steel shim changes dimension by roughly a micron for every 10 °C of thermal swing, which is a tenth of the whole tolerance band. If your incoming inspection measures at 28 °C on a shop floor, you are not measuring the part, you are measuring the room.

The last measurement trap is sample size. A first-article report proves the die was right on Tuesday. It proves nothing about the other 200,000 parts. For a critical shim, the plan needs frequency: a thickness reading per cavity per shift at minimum, with air-gauge sorting or 100% inspection where the application (matched sets, high preload) demands it. Ask your supplier for the SPC chart, not just the FAIR.

Stack Thickness Accumulation: The Math Buyers Forget

Here is the part that surprises purchasing. One shim at ±0.01 mm is tame. Stack ten of them and the question is whether the errors add or cancel. Worst-case linear addition says the stack tolerance is 10 × 0.01 = ±0.10 mm. Root-sum-square (RSS), which assumes independent random errors, gives ±0.01 × √10 ≈ ±0.0316 mm. Real production lands between the two, closer to RSS if the process is centered and stable.

The practical move: if your design stacks more than three shims, specify a matched-set shim where we measure and bin the lot so the customer receives shims whose thicknesses are correlated, not independent. We can deliver a stack total held to ±0.02 mm even when each piece is ±0.01 mm, simply by sorting. That is a trick the drawing never tells you about but the assembly line thanks you for, because a matched stack removes the worst-case panic from the tolerance loop.

Two more rules keep the stack honest. First, measure the assembled stack, not just the parts, because the stack total is the dimension that matters and it is cheap to verify with a height gauge on the assembly fixture. Second, remember that a laminated shim is a stack by design - if you peel layers on the line, the layers themselves carry thickness tolerance, and the peel logic needs the same RSS thinking or the field gets the surprise instead of the assembly bench.

Flatness Is Not Thickness

You can hold thickness to ±0.005 mm and still ship a shim that rocks. Flatness measures deviation from a true plane, and a stamped part wants to curl from the cut edge and from residual stress in the strip. We spec flatness on two scales. For thin shims under 0.3 mm we hold ≤ 0.05 mm total indicator reading (TIR) across the part on a granite flat with a light probe, and we often add a flattening or leveling pass - sometimes a secondary coining operation - to pull that curl out. For thicker spacers we hold flatness to ≤ 0.1 mm per 25 mm of length.

A locating spacer with poor flatness cams the assembly; the bolt pulls one side down first and the bore goes out of square. A spring shim with a wave in it delivers a different preload at every corner of the bolt circle. Flatness is why we stress-relieve phosphor bronze after blanking and why we never ship a shim straight off a high-speed press without a flat-check on the granite. If the part must stay flat through assembly and service, the drawing should also say which side faces the mating surface, because curl direction matters and the die can be tuned to throw the curl away from the functional face.

Packaging is part of the flatness story. Thin shims stored loose in a box bow from their own weight and handling; we ship thin parts in trays or on flat separators and keep the curl, oil, and edge damage out of the shipping container. A flat part that leaves the press and arrives bowed is a flatness failure nobody puts on the supplier scorecard.

Burrs, Edges, and the Coating That Makes an Insulator

An insulating washer is only insulating until a burr pierces the film. Plain stamping leaves an exit burr on the punch side, typically 5-15% of the material thickness - so on a 0.5 mm part you can expect a 0.03-0.075 mm burr. For an insulator we deburr or tumble, then apply a dielectric coating such as anodize on aluminum variants or a powder or film on steel, and we verify breakdown voltage rather than assuming the spec held. Edge condition also matters for adjustment shims that slide during adjustment; a sharp burr galls the mating face and wrecks the very clearance you were setting.

The coating thickness itself counts against your gap, so we net it out of the stamped dimension - a 0.02 mm film means the bare part is cut 0.02 mm thinner than the final target. That netting decision belongs on the drawing: state whether the tolerance applies to the bare metal or the finished coated part, because the two are not the same and the difference is exactly the film thickness. If the coating wears in service, the gap grows by the worn amount, so a coated adjustment shim that must hold a preload for the life of the unit needs a wear callout too.

Deburring methods have their own trade-offs. Tumbling rounds the edges but can also round the functional faces and change thickness at the edge. Brushing is directional and leaves a radius on the brushed edge only. Hand or vibratory finishing removes the burr but adds handling cost and part-to-part variation. For a critical burr spec, the right answer is usually to control the die - sharp punches, correct clearance, and regular tool maintenance produce a small, consistent burr that a single controlled pass removes, instead of a variable burr that takes three different operations to chase.

How Shims Are Made: Progressive Dies and Fine Blanking

Shims are made the same way every other precision stamping is made: strip in one end, finished parts out the other, with the tolerance decided by the tooling and the measurement plan, not by luck.

The progressive die route. For shims inside the 0.05-3.0 mm window, a progressive die blanks, pilots, forms, and cuts off in one continuous pass. Pilots at ±0.005 mm positioning keep every station registered to the same datum, so the bore and OD stay concentric run after run. Die sections are ground, and for high-volume thin shims the working sections run in tool steel with carbide where wear concentrates. In-die sensors watch for misfeeds and tool breakage, and the press stops before a damaged die starts making scrap. Our high-speed lines run to 300 SPM on small work; thin-shim jobs run slower, because feed stability and oil control matter more than speed. The capability and process detail behind this route is covered in the progressive die high speed stamping service page.

Fine blanking and the thick-part route. Above about 2.0 mm, or when the shear edge has to be clean and square for a seal or a press fit, plain stamping gives way to fine blanking, which uses a third action to squeeze the material so the cut is almost pure shear with a smooth edge. Fine-blanked spacers hold thickness and edge quality that a standard blank cannot. Below the stamping window entirely - prototypes, odd shapes, tiny quantities - waterjet and laser cutting take over, and for thin shims in low volume, etching is sometimes the right call. The fine blanking guide walks the process choice in detail.

Sorting and inspection. The final step is where shims earn their reputation. High-volume shims run through air gauges that classify thickness into bins; matched sets are built from the bins so the stack total lands inside ±0.02 mm; critical parts get 100% dimensional check plus the flatness probe. Every lot carries the SPC chart and the material certificate, because a shim program is a data program. If your supplier cannot hand you a thickness distribution for the lot, you are buying luck, not shims.

Applications That Fail Without the Right Shim

Shims fail quietly, and the application list reads like a tour of expensive warranty claims.

Bearing preload. A bearing runs quiet, true, and long when the preload is right, and it hums, skids, and dies young when it is not. Stamped shims set that preload in gearboxes, pump housings, and spindle assemblies, and the shim thickness tolerance transfers directly into bearing load. This is where the matched-set discipline pays: a bearing stack built from independent ±0.01 mm shims can land anywhere in a 0.06 mm window, and the bearing does not care about your RSS assumptions.

Motor stack height. In electric motors, rotor and stator stacks are built from laminations, and the shims and spacers between them set the air gap and the stack height. A shim that is off by a few microns shifts the gap, changes the magnetic circuit, and shows up as efficiency loss, cogging, or noise. Motor work is also where the flatness spec matters most, because a bowed spacer between lamination packs becomes a vibration source at speed. The motor lamination stamping guide covers the stack-building context these shims live in.

Gear backlash and endplay. A gear pair needs a controlled axial position and a controlled backlash. Shims under the bearing races set both, and the tolerance chain runs shim thickness to bearing position to mesh quality. If the shim supplier's process drifts, the gearbox gets louder, the teeth wear faster, and the fix is a teardown.

Sensor air gaps. Hall-effect and proximity sensors read a magnetic field across a precise air gap. The spacer that sets that gap is often a thin stamped shim, and its thickness tolerance is the sensor's calibration tolerance. A 0.02 mm drift is a false trigger or a missed trigger, and neither is acceptable in a safety circuit.

Insulating barriers. Where a fastener must not become a current path, the insulating washer carries the clamp load while blocking the circuit. The failure mode here is electrical: a burr through the film, a flatness failure that lets edge contact short the stack, or a coating that was never validated for the working voltage. The washer family, stamped and otherwise, is covered in the washer, clip, and spring stamping guide.

Cost Drivers: Yield, Sorting, and Inspection

Shims are cheap per gram and expensive per tolerance, and the price ladder is mostly set by three drivers.

Material yield. Thin strip is priced by the kilogram, and a shim that blanks 12 per row instead of 8 per row cuts material cost by a third before anything else happens. Nesting and strip width are where the die design earns its keep, and the layout decision belongs to the stamper who knows the press and the coil. Offal from thin-strip stamping is not scrap to be ignored; it is a line item on every quote.

Sorting and inspection. This is the hidden cost. A ±0.01 mm shim run with SPC sampling costs one price; the same part 100% air-gauge sorted and binned into matched sets costs another, because sorting is labor, gauging is capital, and both scale with quantity. Put the inspection level on the RFQ with the tolerance, and the quote stops being a guessing game.

Coating and finishing. Anodize, powder, film, and deburring passes each add a process step and a verification step. Breakdown-voltage testing on insulators and thickness checks on coatings cost real time, and they are non-negotiable on the parts that need them. The plating and surface finish guide maps the finish options to the service environments that justify them.

The economics have one more lever: batch. Tooling for a shim program is modest compared with a complex terminal die, so the tooling amortization per part drops quickly with volume. What does not drop is the inspection cost per part, which is why a high-volume shim program should push hard on in-line gauging and process capability - the SPC chart is the cheapest inspector on the line.

How We Actually Make the Call on Your Job

Every quote starts with the failure mode, not the drawing. Tell me what the shim is doing - carrying preload, blocking current, locating a gear - and I will tell you the material, the process, and the inspection plan. A 0.1 mm 304 adjustment shim for a sensor mount, supplied in matched sets, sorted to ±0.01 mm at Cpk 1.67, flat to 0.05 mm TIR, is a completely different animal from a 1.5 mm brass locating spacer with a 0.02 mm coating. Both are "shims." Only one of them will survive your line without a service bulletin.

Common questions we answer before quoting:

Can you hold ±0.01 mm on a 0.2 mm shim? Yes, with the right coil (bought to ±0.005 mm gauge), ground die sections, and air-gauge verification. The cost is in the measurement and the SPC, not the press.

Do matched sets cost more? They cost the sorting labor and the binning, but they remove the worst-case stack tolerance from your assembly line. Most programs find the set price cheaper than the rework it prevents.

What do you need on the drawing? Thickness with tolerance, flatness, burr limit and burr side, OD and ID with concentricity, material grade and temper, coating (if any) with its thickness, and the environment or function of the part. The more function you give us, the better the material call.

What about prototypes? We run first-article lots on production tooling or waterjet for geometry proof, measure on the granite flat and the air gauge, and hand you a Cpk report with the parts - not a promise, a number.

If your assembly is fighting a clearance or a preload problem and you suspect the shim is the variable, send the drawing and the failure symptom through the quote request page. We will stamp a first-article lot, measure it, and hand you the Cpk report with the parts. That is how thirty years of staying in tolerance reads on a spec sheet, and it is the fastest way to find out whether your gap is real or just accumulated error you can engineer out.