Micro Stamping: Precision Parts Thinner Than a Hair

A 0.05 mm phosphor bronze coil, 180 mm wide, running at 250 strokes per minute on a battery tab job, will produce 400,000 miniature stamped parts before the shift ends. It will also produce 400,000 rejects if the punch sharpness interval was set from a 0.3 mm stainless playbook. On that job, the burr height limit was 5 microns, the punch band started rubbing past roughly 350,000 hits, and nobody had written a regrind trigger into the control plan. The parts looked perfect to the eye. The vision station caught them at 11 to 14 microns of rollover, right at the contact interface where the downstream weld sits. Micro stamping fails this way almost always: not in the dramatic crash of a broken die, but in the quiet drift of a burr, a curl, or a feed register that moved 0.01 mm over eight hours. If you specify miniature stamped parts, the numbers below are the ones your supplier needs to hear you say.

The Snapshot

• Micro-class parts at this plant sit in the under 1 g mass band, typically 1 to 15 mm across, formed from strip in the 0.05 to 0.3 mm gauge range, inside the plant's full 0.05 to 3.0 mm window.
• Punch and die clearance at thin gauge is specified as a percentage of material thickness, commonly about 8 to 15 percent total depending on material, which at 0.05 mm means single-digit micron gaps per side.
• Burr height discipline on micro work follows published practice of roughly 10 percent of material thickness or less, so a 0.05 mm foil carries a limit near 5 microns.
• The microstamping machine side of the house: 21 high-speed presses from 25T to 110T, up to 300 SPM, with feed register held to +/-0.005 mm.
• Metrology support: CMM resolution at 0.001 mm, image measurement systems at +/-0.005 mm, optical comparators, and profilometer-class surface checks across 20 measurement, vision, and gauge instruments.
• Volume economics: nested or series progressive layouts on hair-thin strip reach 500,000+ pcs/day class output, which is what makes per-part cost on sub-gram stamped components viable.
• Finishing such as selective gold in the 2 to 8 micron class used across the connector industry runs on managed partner lines, coordinated and qualified by us rather than claimed in-house.

What Actually Qualifies as Micro Stamping

There is no ISO line that says a part becomes microstamping at a defined gram weight, so buyers should treat the label as a band, not a badge. In practice, three things move a job into micro territory at the same time: part mass, strip gauge, and the tolerance stack. A terminal that weighs 0.02 g, punches out of 0.08 mm copper alloy, and holds a 0.02 mm positional call is micro regardless of what anyone calls it. A 5 g bracket from 2 mm plate is not, even if it fits in your palm.

The working definition we run to: miniature stamped parts under the 1 g class, most of them well under 0.1 g, produced from strip between 0.05 and 0.3 mm thick. That sits at the very bottom of the plant's 0.05 to 3.0 mm material thickness window, and it is the reason the same shop can run battery tabs beside kitchen-appliance brackets without pretending the physics is identical. Typical product families in this band include battery tabs and carriers, leadframes for MEMS packages, pins and contacts for RF connectors, miniature connectors and sensor parts, and small housings for watches and monitoring devices, where the enclosure itself is a stamped shell no larger than a grain of rice.

The second qualifier matters more than mass. Below about 0.1 mm gauge, the behavior of metal changes. Strip tears differently, springs back differently, holds burrs differently, and responds to tension like fabric rather than like sheet. A punch/die gap of 0.03 mm sounds trivial on paper, yet at 0.05 mm thickness it is 60 percent of the material, which is closer to shearing paper than blanking plate. Everything downstream of that physics, clearance rules, burr limits, press choice, coil handling, tool steel, inspection methods, comes from this one fact: at thin gauge, the absolute numbers are tiny even though the percentage behavior is normal for stamping.

Gauge, Material, and Where the Applications Land

Materials for micro work favor copper alloys and austenitic stainless foils because they combine thin-gauge availability, clean shearing edges, and plating adhesion. C7025, a copper-nickel-silicon alloy, earns its place where you need high strength plus reasonable conductivity, roughly 40 to 45 percent IACS in the aged condition, for connector beams that must hold force over thousands of mating cycles. C5191 phosphor bronze is the classic spring alloy, cheaper, fatigue-friendly, the default for contacts and clips. C17200 beryllium copper sits at the top of the heap for aged yield strength and about 22 percent IACS conductivity, used where the section is so small that only a beryllium copper can carry the required spring load. On the stainless side, 304 and 316L foils cover corrosion-driven parts, with 316L preferred where sterilization or body-fluid-adjacent exposure appears in miniature connectors and sensor assemblies. Kovar-type iron-nickel-cobalt alloys show up as industry materials where a glass-to-metal seal needs a matched expansion coefficient, a recurring requirement in hermetic packages and sensor housings. The full alloy mechanics, tempers, and selection trade-offs are covered in our guide to lead frame stamping tolerances, alloys, and process, which maps the same families onto frame-level specs.

Strip gauge	Typical materials	Representative miniature stamped parts	Notes for specifiers
0.05 to 0.08 mm	C5191, C7025, 316L foil	Battery tabs, MEMS leadframe tie bars, RF contact beams	Handle like fabric; tension control and vacuum part transport become mandatory
0.09 to 0.15 mm	C17200, C7025, 304	Connector terminals, sensor leads, tiny contact springs	The workhorse band for small metal stampings in electronics
0.16 to 0.25 mm	C5191, 304, Kovar-type	Watch and device housing blanks, leadframe pins, shims	Form features like small bends become realistic with tight inside radii
0.26 to 0.30 mm	304/316L, C7025	Micro clips, thin lids, cover plates	Top of the micro band; clearance percentages still dominate tool behavior

One caution before you pick the thinnest foil to save mass: gauge drives stiffness as the cube of thickness. Dropping from 0.15 to 0.10 mm cuts bending stiffness by almost 60 percent. Many specification problems that look like material problems are actually people asking a 0.08 mm part to behave like a 0.15 mm part.

Clearance and Burr Discipline at Thin Gauge

Die clearance is the single most mis-specified parameter on micro jobs. Published practice expresses clearance as a percentage of material thickness rather than an absolute gap, and thin strip is no exception to the rule, it just makes the arithmetic unforgiving. Typical total clearance lands around 8 to 10 percent of thickness for soft copper alloys and 12 to 15 percent for austenitic stainless, which shears with a wider rollover band. At 0.3 mm on C5191, the total gap is about 0.024 to 0.03 mm. At 0.05 mm, the same percentages produce 4 to 7.5 microns total, which is one to two microns per side after you split it. That is grinding-and-EDM territory where toolmakers talk in tenths of a millimeter and measure in tenths of a micron.

Too little clearance at thin gauge does not give you a cleaner edge, it gives you a secondary shear, burnished rollover on both faces, and accelerated punch wear because the strip is being squeezed rather than cut. Too much clearance produces large rollover and a wide, soft burr that a vision system flags instantly on 0.05 mm material. When register precision matters more than edge perfection, many fine-blanking-style setups deliberately run near-minimum clearance, but on a progression of a hundred plus stations, you optimize the cut for the stations that feed plating and welding, then accept the rest.

Burrs on micro parts are specified the same percentage way. Common published acceptance practice caps burr height at roughly 10 percent of material thickness, so 0.05 mm strip yields a 5 micron cap, 0.1 mm yields 10 microns. Some connector specs push to 15 to 20 percent for non-contact geometry, and tight down toward 5 percent where the burr sits in a mating interface or a weld zone. Insist on burr direction, too: on hair-thin parts, a burr leaning into the strip plane tatters, a burr leaning into the thickness behaves predictably. Write the percent, the measurement method, and the location into the drawing, or you are auditioning for the defect table later.

The Microstamping Machine: What the Press Has to Do

Buyers often assume micro parts need the smallest press. Mostly correct, but for the wrong reason. A 0.05 mm blanking force on a 2 mm square is trivial, a few kilograms. Tonnage is not the constraint. Stability at the bottom of the stroke window is. The press has to deliver repeatable ram behavior across millions of hits so that the punch entry point never wanders, which is exactly why our low end sits at a rigid 25T frame rather than a flimsy bench machine: frame deflection, slide guidance, and bearing wear all get proportionally expensive when your whole tolerance story is 0.01 mm. The 21 high-speed presses on the floor, 25T through 110T, cover micro work through mid-size small metal stampings without moving a job onto a machine whose clearances swallow your band.

Speed is the second requirement. Up to 300 SPM is where thin-gauge copper-alloy microstamping earns its volume numbers, and high speed only works when the feed, tool, and part-handling all keep pace. Three subsystems decide whether a fast press produces parts or produces dust. Feed precision: servo feeders on micro strip hold register within +/-0.005 mm, the same number as our general tolerance band, and pitch error at 300 SPM is what the vision station sees as a drifting pattern. Stroke control: the press must run with minimal vertical play so punch entry timing never shifts relative to the strip. Rigidity under lateral load: at high SPM, even a hint of punch drag on a 1 mm wide carbide punch sends a stress wave that snaps it.

The third requirement is what nobody lists in a brochure: machine cleanliness and part flow. At 0.05 mm, a stray burr fragment on the die face is a rock on a highway. Air blow-off, vacuum pickup transport, and enclosed strip paths are not optional accessories on micro jobs; they are the difference between a 99.5 percent yield and a 96 percent yield, and at 500,000 pieces a day that gap buys or loses money in one shift.

Coil Handling and Tension Control for Hair-Thin Strip

Everything about a micro coil is fragile. A 0.05 mm, 150 mm wide stainless foil weighs about 3 kg per 100 meters of length, stores almost no bending stiffness, and will telescope, edge-damage, or stretch into a wavy ribbon if you treat it like a normal 0.5 mm coil. Good micro stamping starts at the decoiler with four disciplines. First, tension: back tension on the decoiler and controlled payoff tension, held in a narrow band, typically managed with a dancer roller or servo pay-off loop, so the strip is taut enough to register consistently but never taut enough to elongate. Yield strength in thin tempers of C5191 can sit near 400 MPa; a 150 mm wide, 0.05 mm strip yields at roughly 300 N of tension, which is less force than a firm hand pull. Second, alignment: the coil, feed, and die centerline must be dead straight within a couple of tenths over the whole run, or the pilot holes drift and your +/-0.005 mm register is gone. Third, edge care: any nick from a bad slit edge propagates during shearing and shows up as tears at predictable stations. Fourth, surface respect: oil choice and roll marks matter when parts go on to plating, because a foil that looks fine can carry a residue that gold will not wet onto.

Ask a supplier what their decoiler does at low tension, not at high speed. The honest answer separates shops that have run foils from shops that have run sheets.

Tooling for Miniature Stamped Parts

Progressive tooling is the backbone here, and the same logic that governs any progressive die stamping job applies, scaled down until the details bite. A micro progression for a connector terminal might run 40 to 80 stations across a strip 100 to 200 mm wide, blending blanking, piercing, staking, bending, trim, and cutoff. Station count climbs because every feature wants its own cut at thin gauge: combined operations smear edges and bend fragile geometry.

Punches are solid carbide almost without exception. At 0.5 to 2 mm diameters, tool steel bends, rubs, and wears; carbide holds its edge and its size across long runs, at the price of brittleness that punishes misalignment with instant fracture. Die inserts are carbide too, usually with changeable wear parts so a tired insert swaps in minutes rather than pulling the tool apart. Fabrication stays conventional precision machining: wire EDM, sinking EDM, and surface grinding produce the punch profiles and die gaps in the micron band, with the toolmaker lapping final clearance rather than trusting any single process. We do not claim exotic laser micro-machining in the die shop; edge quality at this scale comes from controlled shearing and careful grinding, and that is a feature, not a limitation, because it keeps the entire edge in a known, repeatable metallurgical condition rather than a heat-affected zone.

Strip layout design is where micro tooling earns or burns money. Pilot patterns go early and often, because they are the datum for every station after them. Carrier design keeps parts tied until cutoff while avoiding scrap locks. And knockouts versus edge stands: at thin gauge, a blank that tries to stay put in a die plate will scratch, so strippers and engineered knockouts do the pinning.

Nesting, Series Layouts, and Per-Part Economics

Sub-gram parts have no business costing money, but they can, when the layout wastes strip. Two nesting strategies fix that. Wide multi-lane nesting runs the same part in 6 to 20 lanes across the strip, multiplying output per stroke; the cost is that any tool defect repeats in every lane, so lane count is a bet on your inspection discipline. Series layout interleaves two or more different parts on one strip, a terminal lane beside a tab lane beside a shim column, so one press, one tool, one feed system serves an entire product family. Our throughput at the 500,000+ pcs/day class comes from this arithmetic: a 300 SPM press with 10 lanes produces roughly 180,000 parts per hour on paper, and a well-run three-shift week turns that into numbers that let a 0.03 g contact sell for fractions of a cent.

The offsetting cost is scrap behavior. Hair-thin skeletons are springy, they break up in the shifter, and they jam chutes if the layout leaves slivers. A specifier who demands the thinnest, narrowest web to save material sometimes discovers the press cannot run at full speed until the web comes up. Balance the strip economy and the run speed together, and insist the supplier's quote states the width, the web dimensions, and the SPM they intend to run.

Inspection and Metrology on Parts You Can Barely See

Hand gauges end around 0.3 mm features; below that, measurement belongs to machines. The stack at this plant covers three layers. CMMs with 0.001 mm resolution carry the datum work: position of contact windows, critical widths, true-position of tie-bar remnants, evaluated against the same GD&T framework you would use on any stamping. Image measurement and optical comparator systems, at +/-0.005 mm, do the high-throughput dimensional checks on 2D geometry and can profile an entire terminal field in a single capture. Vision stations run inline as well, watching burr, edge condition, and feed register stroke after stroke, because the only defect pattern that stays quiet on a 0.05 mm job is the one nobody photographs.

Surface work gets its own layer: profilometer-class checks for Ra and for the tiny edge radii that determine whether a contact mates cleanly or digs into a housing. Flatness and curl are the eternal thin-gauge fights, measured on optical flats or by vision with a defined datum, and specified the right way, which means calling a sweep or a bow value, not a vague "flat" note. The table below shows the band we hold as routine production capability on micro work, remembering that capability claims without a Cpk method behind them are marketing.

Characteristic	Routine capability (micro band)	Method	Specifier note
Feature position / true position	+/-0.005 mm	CMM at 0.001 mm resolution, image measurement cross-check	Tie every call to a pilot-hole datum, not to the fragile part edge
Blank size / outline	+/-0.01 mm	Vision system at +/-0.005 mm, sampled on CMM	Thermal and feed drift move outline more than position on long runs
Flatness / sweep	0.01 to 0.05 mm class depending on gauge	Optical flat, shadow moire, or vision profiling	Call sweep as a number with a free-state condition after stress relief
Burr height	Under roughly 10 percent of gauge (5 microns at 0.05 mm)	Vision at magnification, profilometer edge trace for disputes	Specify location and direction, not just a height cap
Surface texture on functional faces	Ra class checks at profilometer resolution	Contact profilometer	Name the faces that matter; micro parts are mostly non-functional surface

Plating and Finishing of Micro Parts

Thin-gauge copper parts almost always end up plated: tin for solderability, silver or gold for contact interfaces, nickel as a barrier, and the whole stack has to land on a part that weighs 20 milligrams without gluing itself to its neighbor. That is why finishing of miniature stamped parts runs on managed partner lines, qualified and coordinated by us, rather than pretending a stamping plant owns every chemistry. For RF and signal contacts, the connector-industry band of selective gold at 2 to 8 microns on mating surfaces over a nickel underplate is the familiar specification, with the thicker end reserved for high-cycle sockets. On micro geometry, selective plating means fixtures, masking discipline, and racking that survives parts light enough to be moved by air.

Two specification habits save plate jobs. First, tell the finisher the part mass and bulk density: 500,000 tiny parts in a barrel behave like fluid, and retention, drainage, and tumble marks are real engineering decisions, not cosmetic ones. Second, specify plate location and minimum thickness with a measurement method, XRF on sample coupons at defined sites, so acceptance is data rather than appearance. And if the part will be welded after plating, flag the weld zone now: gold and tin migrate, and a weld button over a plated edge is a different conversation than a welded post-plate sequence.

The Defect Playbook: What Actually Goes Wrong

Defect	Root cause	Fix
Punch breakage	Carbide at 0.5 to 2 mm is brittle; misalignment, lateral load at entry, double-punch from a sliver, or pilot wear letting stations shear off-axis	Recut entry timing, tighten guide clearances, add punch sensors on every critical punch, scrap slotted so it cannot ride the die
Excess burr	Punch band worn past its sharpness interval; clearance percentage drifting as die faces abrade	Defined hit-count regrind schedule, burr trend from inline vision, not end-of-shift eyeball sampling
Curl and bow	Shear-zone asymmetry at thin gauge; improper stripper pressure; thermal stress relief before the part is flat	Optimize shear percentage, flat stock selection, press-side flattening stations or controlled free-state settling, then verify flatness after any bake step
Magnetic carry-over	Steel-class foils (304/316 work-harden to slightly magnetic states, Kovar-type is magnetic) stick to punch faces or each other; parts bridge and double-feed	Demagnetizer at the coil entry, air knives, non-magnetic die inserts where possible, part-on-punch detection
Register drift	Feed slippage on oily strip, dancer-tension swings, pilot hole tear-out at thin gauge	Servo feeder with registered pressure, larger or doubled pilot holes, tension loop tuning, vision alarm on pattern position
Edge tears and slivers	Nicked slit edge, worn die profile, scrap web too narrow at speed	Coil edge inspection at load-in, scheduled die polish, widen the web before widening your losses

Notice that five of the six fixes are process controls, not tool redesigns. On micro work, the supplier you choose is the supplier's control plan, and the punch list above reads like the questions section below.

RFQ Checklist for Micro Parts

Before a tooling decision on anything thinner than a hair, put these ten answers in the RFQ packet. They follow the same spirit as the DFM groundwork covered in 5 DFM questions before committing stamping tooling, but aimed straight at the micro band, with no free pass for vague drawings.

• State the gauge and alloy temper, not just "0.08 mm bronze." C5191 in spring temper shears differently than in half hard, and the toolmaker sizes clearance to the answer.
• Give mass and functional envelope: part weight, planar size, and where the part lives in the assembly, so stiffness and flatness requirements can be derived instead of guessed.
• Put every tolerance on a datum you can feed with: pilot holes or strip edge, and say which.
• Write the burr cap as a percent of gauge with a location and a measurement method, and separate contact zones from cosmetic zones.
• Define free-state flatness as a sweep or bow number with units, and say whether it applies before or after any bake or plating.
• Name the plating stack, placement, and minimum thickness with the acceptance method (coupon plus XRF is normal), and state whether parts plate before or after cutoff.
• Declare the volume band and whether the part shares an assembly sequence with siblings; a series layout quote looks very different from a single-part quote.
• Provide the mating interface geometry: weld lands, snap windows, contact beams, and any feature that cannot carry a 0.01 mm deviation.
• State the inspection expectation: CMM on FAI, vision inline on production, sample AQL, and who owns the gauge R&R on the micron-scale features.
• Ask the supplier for the punch-life and regrind trigger they intend to run on your tool, and whether their control plan monitors burr and register stroke by stroke.

Frequently Asked Questions

Is micro stamping cheaper than machining tiny parts?

At volume, yes, by two or three orders of magnitude per part. A progressive tool on a 25T press producing 10 to 20 lanes at speed delivers sub-cent contacts where Swiss machining charges cents. Below roughly tens of thousands of pieces, the tooling investment dominates and machining or etching can win; micro stamping is a volume process, and the tool is the business model.

How thin can strip actually go in a normal progressive press?

Production-credible territory for foil stamping starts around 0.03 to 0.05 mm with proper handling and tension control. Our window begins at 0.05 mm. Below about 0.05 mm, part stiffness drops so far that transport, plating racking, and singulation often define feasibility more than the punch press does.

Do micro parts need special packaging and handling?

Yes, and spec it early. Sub-gram parts sinter into a brick if plated or tumbled without separation, carry static and magnetic pickup, and bridge in chutes. Vacuum trays, separated bulk with cushioned interleaving, or pick-and-place-ready strip packaging are normal answers. If the part feeds an automated assembly, tell the stamping supplier how the assembly picks it up, because orientation control starts at the die cutoff.

Can you stamp magnetic materials without carry-over problems?

It is manageable, not magic. Work-hardened 304, 316L in some conditions, and especially Kovar-type alloys cling to tooling and each other. The standard mitigation is a demagnetizer at coil entry, air assist at cutoff, non-magnetic inserts at critical stations, and part-on-punch detection so a bridged part trips the press instead of doubling up.

What is the realistic tolerance story on miniature stamped parts?

With a modern servo-fed microstamping machine and fine-pitch progressive tooling, +/-0.005 mm on registered features is routine production territory, verified on CMM at 0.001 mm resolution and watched inline by vision systems. Burr and flatness, not position, are the parameters that separate a well-specified micro part from a rejected one, so spend your tolerance budget there.

The Bottom Line

The risk on micro stamping is quiet drift: burrs, register, and curl moving past micron-scale limits long before an operator notices. Progressive die stamping wins where those numbers stay under statistical control, servo feed at +/-0.005 mm, inline vision, and a regrind plan, at volumes from 500,000 pieces a day class upward.