MFG

Manufacturing Glossary

Plain definitions for the manufacturing terms used across this site: machining, molding and tooling, sheet metal, 3D printing, materials, and file-format terms.

A glossary is only useful when each term is grounded in the standard or process it comes from. This page defines the manufacturing terms used across the site, grouped by domain: machining and tolerances, injection molding and tooling, sheet metal and forming, additive manufacturing, materials and properties, finishing, and file format and quality. Each entry ties a word to the number, standard, or process behavior behind it, so a reader can move from the term to the decision it drives.

The terms overlap across processes. Tolerance applies to a CNC pocket and a laser cut; surface finish applies to a milled face and a printed wall; kerf and the heat-affected zone shape both cutting and welding. Read the entries as a shared vocabulary, not a set of isolated definitions, and use the cross-references to follow a term into the process where it matters most.

Machining and tolerances

Processes and general tolerances

  • CNC machining. Computer-controlled subtractive machining, milling and turning, that removes material from a solid blank to hit a programmed dimension. It is the default choice when a part needs a tight tolerance on a mating face or a fine surface that additive or cut sheet cannot reach.
  • Tolerance. The allowed variation in a dimension, the band inside which a part is accepted. For machined metals, general tolerances follow ISO 2768-1, with class f, or fine, at plus or minus 0.05 millimeters for a 0.5 to 3 millimeter dimension and class m, or medium, at plus or minus 0.10 millimeters. A tighter tolerance costs more time and finer tooling, so it is specified only where the function demands it.
  • ISO 2768. The general-tolerance standard for machined parts, split into ISO 2768-1 for linear and angular dimensions and ISO 2768-2 for geometric tolerances. It lets a drawing carry a default tolerance class, such as class f and grade K, instead of writing a tolerance on every dimension.
  • Feed and speed. The paired cutting parameters: cutting speed, how fast the edge passes through the material in surface meters per minute, and feed, how much material each edge takes per pass or revolution. Spindle speed follows from cutting speed and cutter diameter, n equals vc times 1000 divided by pi times Dc, and the two are set together because speed drives tool life while feed drives finish and power. The feed and speed table gives cited starting values.

GD&T and datums

  • GD&T. Geometric Dimensioning and Tolerancing, a symbolic system, governed by ASME Y14.5 and ISO 1101, that controls form, orientation, location, and runout rather than size alone. A flatness or position callout in GD&T says what a face or hole must do, not merely how wide it is, which captures functional intent that plus or minus tolerancing misses.
  • Datum. A reference feature, a plane, axis, or point, that everything else on a part is measured from. Datums are the anchors of GD&T, because a position or perpendicularity callout is meaningless until the datum it is measured against is fixed on the drawing and on the inspection setup.

Fits and threads

  • Hole-fit classes. The ISO 286 system that writes a fit as a letter plus an IT grade: H7 is a hole tolerance, 0 to +21 micrometers at 18 to 30 millimeters, and g6 is a shaft grade that sits just below nominal for a running clearance fit. The letter fixes the fundamental deviation and the number the tolerance-zone width, so H7/g6 is a precise statement of clearance, not a habit. The hole tolerance and fits table gives the deviations for H7 holes with f6 through s6 shafts.
  • Torque spec. The tightening torque that reaches a fastener’s target clamp load, set by size and grade under ISO 898-1. Lubricated threads reach the same clamp at about 25 percent less torque than dry ones, so the thread condition is part of the spec, not an afterthought. The torque spec tables list values by size and grade.

Surface and feature terms

  • Ra, surface finish. Roughness average, the standard measure of surface smoothness per ISO 21920-2, where a lower Ra is smoother. As-machined metal sits around Ra 3.2 micrometers, or 125 microinches, a fine-machined surface around Ra 1.6, and a ground surface around Ra 0.4 micrometers, or 16 microinches. Ra is set against what the face must do, since a smoother finish takes a finer tool and a slower pass.
  • Chamfer and fillet. A chamfer is a beveled edge that transitions between two faces at an angle, usually 45 degrees, while a fillet is a rounded inside radius that blends them. Chamfers break a sharp edge for handling and assembly, and fillets remove the inside corner that would otherwise concentrate stress and crack under load.
  • Pitch. The distance from one point on a thread or tooth to the same point on the next, the spacing that defines a screw thread or a gear. Pitch is paired with the tap or die that cuts it, so a hole tapped for an M8 by 1.25 thread takes a bolt with a 1.25 millimeter pitch.
  • Tap and drill. To tap is to cut internal threads in a drilled hole with a tap, and the drill size is the hole drilled first, sized just under the thread’s minor diameter. A tap drill chart gives the right drill for each thread, because an oversize hole leaves shallow threads that strip and an undersize hole breaks the tap.
  • Gauge. A standard thickness for sheet metal, where a lower gauge number means a thicker sheet, so 16 gauge steel is thicker than 18 gauge. Gauge predates metric thickness in much of North American sheet metal work, and because gauge differs between steel, aluminum, and stainless, the metric thickness, in millimeters, is the unambiguous value to put on a drawing.

Sheet metal and forming

Cutting terms

  • Kerf. The width of material a cutting process removes, the slot the beam or abrasive stream leaves behind. Fiber laser kerf runs about 0.15 to 0.30 millimeters and waterjet kerf about 0.75 to 1.15 millimeters, so nesting software offsets each cut path by half the kerf to land the part on size.
  • Heat-affected zone, HAZ. The band of metal next to a thermal cut, laser or plasma, whose metallurgy has changed from heat input, often softer or harder and less corrosion-resistant than the parent metal. Fiber laser leaves a HAZ of about 0.10 to 0.25 millimeters on steel, while waterjet, a cold cut, leaves none.

Bend calculation terms

  • Bend allowance. The length of the neutral axis through a bend, the arc length the metal travels as it folds, used to calculate the flat blank size that folds to the wanted flange. It is derived from the K-factor, the material thickness, and the bend angle.
  • Bend deduction. The amount subtracted from the sum of the two flange lengths to get the flat blank length, the correction for the material that compresses inside the bend. It is calculated as the K-factor times thickness times the quantity 180 degrees minus the bend angle, times 0.01745.
  • K-factor. The ratio that locates the neutral axis within a bend as a fraction of the material thickness, the value that links bend allowance and bend deduction. It is typically 0.40 to 0.45 for most materials (PC-070).

Bend behavior and cut quality

  • Springback. The elastic recovery of sheet metal after a bend, the amount the flange springs back toward flat when the press releases, so the tool must overbend to hit the final angle. Soft aluminum springs back about 1 to 3 degrees, carbon steel 3 to 10 degrees, and stainless steel 5 to 12 degrees.
  • Minimum bend radius. The smallest inside radius a material will fold through without cracking, typically at least half the material thickness for ductile metals. A sharper radius concentrates strain at the outside of the bend and risks a tear, so the radius is set by both the material and the die.
  • ISO 9013-1. The standard that classifies thermal cut quality, from Level 1 for aerospace and medical precision up to Level 5 for flame-cut thick plate. It grades the cut edge on perpendicularity, dross, and surface roughness, so a Level 2 or 3 cut is the typical fiber-laser target for industrial work.

Welding and joining

  • Distortion. The bending of a welded assembly from uneven heating and cooling, where the weld and its surrounding metal expand and contract against the rest of the part. Clamping, balanced weld sequencing, and lower heat input control it, and the welding overview covers the techniques in depth.
  • Filler metal. The metal added to the weld pool to fill the joint, classified under AWS A5.9 for stainless (ER308L for 304, ER316L for 316) and A5.10 for aluminum (ER4043, ER5356). Matching the filler to the base metal is what makes the weld deposit sound and corrosion-resistant, and the welding materials guide maps the pairings.
  • Porosity. Gas trapped in the weld pool as it solidifies, leaving voids in the finished weld, usually from oil, moisture, or mill scale on the joint, or from shielding gas lost to wind. Cleaning the joint and protecting the gas stream prevent it, and the welding defects guide covers its causes and detection.
  • Weld joint. The geometry in which two pieces meet for welding: butt, lap, tee, corner, and edge. The joint type sets the weld, the preparation, and the symbol on the drawing, all covered in the weld joints guide.
  • Weld symbol. The drawing callout, defined by AWS A2.4, that specifies a weld: the arrow points to the joint, the reference line carries the weld type and size, and the side of the line tells whether the weld sits on the arrow side or the far side. The weld joints guide shows how to read the common symbols.
  • Weldability. How easily a metal can be fusion welded into a sound joint that keeps its service properties. Low-carbon steel welds easily, stainless welds well with heat control, and 6061-T6 aluminum loses roughly half its strength in the welded condition, which the design must size for; the welding materials guide works through each family.

Injection molding and tooling

  • Overmolding. Molding a second material, typically a soft TPE, over a first molded part so the two leave the tool as one piece. The bond is chemical, mechanical, or both, and it depends on substrate compatibility, so the resin pair is chosen together rather than one at a time. The overmolding and insert molding guide works the design rules.
  • Insert molding. Encapsulating a pre-made component, usually a metal insert, that is placed in the open mold before injection, so the plastic locks it in mechanically rather than bonding to it. A threaded boss or contact pin is then molded in place instead of fastened on afterward, which is the practical difference from overmolding. The overmolding and insert molding guide contrasts the two.
  • Two-shot molding. The automated form of overmolding, in which one machine molds the first material, indexes the part to a second cavity, and shoots the second, so both materials finish in one cycle. It needs a second mold or a multi-barrel press, which is why it pays off at volumes where hand-loaded overmolding does not. The overmolding and insert molding guide covers when two-shot fits.
  • Parting line. The seam where the two mold halves meet, present on every molded part. Dimensions held within one mold half hold tighter than dimensions that cross the parting line or pass through moving mold sections, so critical features belong on one side of it, and a stepped parting line needs a shut-off angle of about 7 degrees, 5 degrees at the minimum. The design rules guide places features against the line.
  • Draft angle. The taper on a molded wall that lets the part release from the steel during ejection. Published guides treat 0.5 degrees per side as the safe minimum on favorable geometry and 1 to 2 degrees as normal practice, and a textured surface adds about 1 degree per 0.025 millimeters of texture depth, from about 3 degrees under 15 micrometers to 6 to 8 degrees on coarse grain. The design rules guide tabulates draft by texture depth.
  • Gate. The opening through which melt enters the cavity. Every gate leaves a witness mark, so gates stay off cosmetic faces, and the gate is placed so melt runs from thick sections into thin ones, which walks weld lines away from appearance- and strength-critical spots. The design rules guide covers gate and ejection decisions.
  • SPI mold classes. The industry convention, Class 101 through Class 105, that grades a mold by intended shot life: Class 101 covers hardened-steel production tools built for 1,000,000-plus cycles, down through Class 103 P20 tools at up to 500,000 cycles to Class 105 prototype tools at about 500 cycles. The classes are attributed to the former Society of the Plastics Industry, now PLASTICS, as a convention rather than a published standard. The rapid tooling guide lays out the bands by class.
  • Bridge tooling. The middle tier of mold construction between rapid tooling and full production steel, built to run real parts while the production tool is made. Usage varies, with some suppliers treating bridge and rapid tooling as interchangeable, so the intended shot life is confirmed with the supplier, and everything learned on the bridge tool, from gating to ejection, feeds the production tool design. The rapid tooling guide covers the tiers and their trade-offs.
  • Shrinkage rate. The percent a molded dimension contracts from cavity size as the resin freezes, which the tool steel must be cut oversize to compensate: a 100 millimeter PP dimension at 1.5 percent takes about 101.5 millimeters of steel. Amorphous resins such as ABS, polystyrene, and polycarbonate run about 0.4 to 0.8 percent, while semi-crystalline PP runs about 1.0 to 2.0 percent and acetal about 2.0 to 2.1 percent. The shrinkage rates table lists flow and cross-flow values by material.
  • Bridge production. Running sellable parts while production tooling is still being built, a timeline decision that fills the gap between prototype and full-rate production rather than a process choice. It is sized against the premium of the interim tool versus the cost of waiting, and it ends at a planned switchover rather than running until it stops. The bridge production guide works the trigger and the sizing.

Additive manufacturing

Process and layer terms

  • Layer height. The thickness of each slice in a fused-deposition print, the value that trades surface quality against build time. A 0.1 millimeter layer prints smooth but slow, a 0.3 millimeter layer builds fast but shows ridges, and most engineering parts land around 0.15 to 0.2 millimeters.
  • Support. The sacrificial structure a printer builds under an overhang or bridge, removed after the build to leave the part behind. Supports cost material, cost time, and leave marks, so a part is oriented or chamfered to need as few as possible, often none below a 45-degree overhang.
  • Anisotropy. Direction-dependent properties, the trait that makes a printed part weaker across its layers than within them. A fused-deposition part is about 20 to 30 percent weaker along the Z, or build, direction than in the XY plane, so load-bearing parts are oriented with the load along the layers.

Materials and standards

  • Sintering. The process of bonding powder into a solid by heating it below the melting point, used in metal powder-bed fusion and in selective laser sintering of nylon. Sintering leaves a slightly rough, porous surface that usually needs post-machining on a mating face.
  • Photopolymer. A resin that cures from liquid to solid under light, the feedstock for SLA and similar vat-printing processes. Photopolymers give a smooth, detailed surface but tend to be brittle and to weaken in sunlight and heat, so the resin is chosen against the part’s working environment.
  • ISO/ASTM 52900. The standard vocabulary for additive manufacturing, the source of the process names, fused deposition modeling, selective laser sintering, and powder-bed fusion, used consistently across the site. It fixes the terminology so a term means the same process on every page.

Materials and properties

  • Tensile strength. The maximum stress a material takes before it breaks, the peak of the stress-strain curve. Aluminum 6061-T6 reaches about 310 megapascals, stainless 304 about 500 to 600, and titanium grade 5 about 895 to 1105, so tensile strength sets the ceiling a part can approach before failure.
  • Yield strength. The stress at which a material stops deforming elastically and takes a permanent set, the practical design limit for a loaded part. Aluminum 6061-T6 yields at about 276 megapascals, well below its 310 megapascal tensile strength, because a bracket that yields has already failed in service even if it has not broken.
  • HDT, heat deflection temperature. The temperature at which a polymer bends under a defined load, the practical upper temperature for a plastic part. PLA deflects around 55 degrees Celsius, PETG around 70, and ABS around 95, so HDT sets which filament a warm-environment part can use.
  • Shore hardness. A durometer scale for soft polymers and elastomers, where Shore A runs from flexible rubber at the low end to hard rubber at the high. TPU filament, the common flexible print material, spans Shore A 60 to 95, and the grade is chosen against how stiff or rubbery the part must feel.
  • Anisotropy in metals and polymers. Direction-dependent behavior that appears in rolled sheet, drawn tube, fiber-filled filament, and printed layers. A fiber-filled nylon is stiffer along the fiber direction, and a cold-rolled steel sheet bends differently with and across the grain, so the material’s direction is part of the design.
  • Hygroscopy. A material’s tendency to absorb moisture from the air, which shifts its dimensions and, in printing filament, weakens the part and causes stringing. Nylon PA6 absorbs about 9 percent water by weight and PA12 about 1 percent, so nylon is dried before printing and stored dry after.
  • Passivation. A chemical treatment, usually acid, that removes free iron from a stainless steel surface and lets the chromium oxide layer form fully, raising corrosion resistance without changing the dimensions meaningfully. It is standard for stainless parts in food, medical, and marine service. The passivation guide covers the nitric and citric processes and their spec points.
  • Anodizing. An electrolytic conversion that grows a hard aluminum oxide layer on an aluminum surface, Type II for color and corrosion resistance and Type III, hardcoat, for wear. Anodizing removes about 10 to 15 micrometers from the surface while building the oxide, so a tight fit is adjusted for the net change. The anodizing versus powder coating comparison works the choice between the two finishes.
  • Bead blasting. Spraying fine glass media at a surface to leave a uniform matte finish around Ra 1.5 to 3 micrometers that hides machining marks with minimal dimensional change, about plus or minus 0.01 millimeters. It is a common pre-anodize texture on aluminum and a final finish on stainless, and the bead blasting guide covers media and pressure.
  • Plating. Depositing a metal layer onto a surface, zinc for corrosion on steel, nickel for appearance, hard chrome for wear, with thickness classes set by standards such as ASTM B633 for zinc. Plating builds outward, so threads and tight fits must allow for the added layer or be cut to a pre-plate size, which the plating guide covers.
  • Powder coat. A dry-polymer finish electrostatically sprayed onto a metal part and baked to a tough, continuous film about 60 to 120 micrometers thick. It gives color and corrosion protection in one coat, but because it adds thickness, a powder-coated dimension must allow for the film or the parts will not assemble. For aluminum parts with tight fits, the anodizing versus powder coating comparison covers the thinner alternative.
  • Electropolishing. A reverse-plating process that removes a thin, uniform layer from a stainless steel surface to leave a mirror finish around Ra 0.1 to 0.2 micrometers. It smooths, cleans, and passivates in one step, which is why it is common on medical and food-contact parts. The electropolishing guide covers the process and its 5 to 25 micrometer removal budget.
  • Standard stock sizes. The thicknesses, widths, lengths, and diameters that mills and stockholders actually produce: plate in 2000 by 1000, 2500 by 1250, and 3000 by 1500 millimeter sheets, round bar from 10 to 250 millimeters. Parts sized to nest on stock cut cost and lead time, twelve 600 by 400 blanks fill about 92 percent of a 2500 by 1250 sheet, while a non-stock size can force a mill order with its own minimum weight and lead time. The standard stock sizes table lists the common sizes.
  • REACH and RoHS. The two EU substance regimes: RoHS restricts 10 substances in electrical equipment, lead to 0.1 percent and cadmium to 0.01 percent in homogeneous material, while REACH requires communication down the supply chain above 0.1 percent of a candidate-list substance. Both are assessed per homogeneous material, not per part, which is what makes a declaration specific. The REACH and RoHS guide covers what each requires.
  • DFM, design for manufacturing. The practice of shaping a part, choosing its features, and writing its drawing so that it can be made reliably and economically by the chosen process. DFM is the umbrella over every entry on this page, because a tolerance, a radius, or a file format chosen without the process in mind becomes scrap.

File format and quality

  • STEP. The STandard for the Exchange of Product data, the file format, .step or .stp, that carries boundary-representation geometry with tolerances and assembly structure. STEP is the primary 3D format for CNC machining and sheet metal, because it preserves the surfaces and datums a CAM system needs.
  • STL. The stereolithography mesh format, a tessellated surface with no tolerance, unit, or assembly data, accepted for 3D printing but not for CNC. STL is universal and simple, but because it carries no units, a file without explicit units is the most common and most expensive file-format error in custom manufacturing.
  • DXF. The Drawing eXchange Format, the standard 2D vector format for laser, waterjet, and plasma cutting, carrying the cut profile as lines, arcs, and polylines. A DXF cut path must use continuous lines, because hidden or dashed lines are ignored by the cutting software.
  • DWG. The AutoCAD native 2D and 3D format, accepted by many suppliers but less universally interoperable than DXF. For cutting, DWG is converted to DXF at the supplier, so sending DXF directly removes a conversion step.
  • Units and the 25.4 scale error. The units, millimeters or inches, that a file is drawn in, which must be stated in the file or the filename because STL and some DXF files carry no unit metadata. A millimeter file read as inches, or the reverse, produces a part at 25.4 times the intended scale, so confirming the units in the order is the single most effective file-format check.

How to use this glossary

Treat these definitions as the shared vocabulary behind every process and material page on the site. When a page mentions a tolerance class, a surface finish, or a file format, the entry here gives the standard and the number it rests on, and the related pages below carry the term into the process where it applies. Read across the groups rather than down one, because the same term, tolerance, finish, kerf, anisotropy, appears in machining, cutting, bending, and printing, and the decision it drives is the same wherever it shows up.

Frequently asked questions

What does Ra mean?
Ra, or roughness average, is the standard measure of surface smoothness per ISO 21920-2. A lower Ra is smoother. As-machined metal is about Ra 3.2 micrometers, or 125 microinches, while a ground surface reaches about Ra 0.4 micrometers.
What is the difference between tolerance and finish?
Tolerance is the allowed variation in a dimension, governed for machining by ISO 2768. Finish, expressed as Ra, is how smooth the surface is. Both tighten at the cost of cycle time and price, so they are set together against what the part must do.
What is kerf?
Kerf is the width of material a cutting process removes. Fiber laser kerf is about 0.15 to 0.30 millimeters, while waterjet kerf is wider at about 0.75 to 1.15 millimeters. Nesting software offsets each cut path by half the kerf so the finished part lands on size.
What is the difference between yield strength and tensile strength?
Yield strength is the stress at which a metal stops springing back and takes a permanent set. Tensile strength is the higher stress at which it finally breaks. A bracket is designed against yield, because it must not bend permanently under load.
What is springback and why does it matter?
Springback is the elastic recovery of sheet metal after a bend, so the press must overbend to hit the final angle. Soft aluminum springs back about 1 to 3 degrees, carbon steel 3 to 10 degrees, and stainless steel 5 to 12 degrees, so the amount is set per material.
Why do STL files cause unit errors?
STL carries geometry as a mesh but stores no units, so a file without explicit units is read against the supplier default. A millimeter file read as inches, or the reverse, prints or cuts at 25.4 times the intended scale, so the units must always be stated in the file or the filename.

Sources