Wire EDM Precision Cutting: How the Spark Removes Metal

• 1
  What actually removes material in wire EDM precision cutting
• 2
  Which features belong on this machine and which do not
• 3
  Where the process gets expensive or simply stops
• 4
  Holding ±0.005 mm and Ra 0.2–0.8 μm
• 5
  Job setup and the mistakes that cost a second run
• 6
  Wire EDM against milling: picking by feature
• 7
  Typical settings and what each one changes

There is no cutter touching the part. A thin wire, usually brass or molybdenum, travels vertically between two guides while a generator pulses voltage across a small gap filled with deionized water. Each pulse ionizes the water, a spark jumps, and a tiny crater of metal melts and vaporizes. The dielectric flushes the debris away before the next pulse.

Because removal depends on electrical energy rather than mechanical force, hardness barely matters. A 60 HRC tool steel insert cuts at the same rate as soft aluminum of similar thickness. That is the core reason this process shows up on hardened dies, punches, and already heat-treated parts that would distort under milling.

The gap is the part engineers must reason about first. Discharge craters form slightly outside the wire path, so the kerf runs wider than the wire itself, typically 0.02–0.05 mm per side depending on the number of passes. Toolpath compensation must account for that offset, or every feature lands undersize by the kerf width.

Thermal damage stays shallow. The recast layer on a finishing pass usually measures a few micrometers, and a multi-pass strategy removes most of what a roughing pass leaves behind. Stress-relieved stock and a skim cut after heat treatment keep the edge clean.

Blind pockets with sharp internal corners are a poor fit. The wire is a straight line between two guides, so it cannot plunge into a closed cavity the way an end mill does. Through-features are the natural territory: profiles, keyways, punch outlines, and slots that run the full thickness of the stock.

Tapered walls are easy. Most machines tilt the upper guide independently, so a die with a 2° relief angle or a progressive die with varying draft comes off in one setup. Taper capability is usually quoted as a maximum angle at a given thickness, and the two trade against each other.

Narrow slots reward the process. A 0.25 mm wide slot in hardened steel is routine, and the same slot milled with a 0.2 mm cutter would snap long before reaching depth. Small holes drilled by EDM first, then wire-cut to finished profile, is a common sequence for fuel injector and nozzle work.

Feature size has a floor. The smallest practical inside radius equals the wire radius plus the spark gap, so a 0.25 mm wire cannot produce a 0.05 mm corner. Designers who need tight internal radii should plan for a separate operation or accept a larger corner.

Conductive material is the hard boundary. Plastics, ceramics, and most composites cannot be cut this way. Graphite and some sintered carbides conduct well enough to machine, but the surface finish suffers from the porous structure. Non-conductive coatings on a conductive core are fine as long as the cut reaches the core.

Cutting speed scales with thickness, and the curve is not linear. Doubling stock thickness roughly quadruples the time because flushing becomes the limiting factor. A 10 mm plate might finish in minutes, while the same profile in 150 mm stock can run for hours. Deep cuts need higher flush pressure and often a second roughing pass.

The wire is a consumable and it does break. Thermal shock, poor flushing, and worn guides all contribute. A break mid-cut means rethreading, and on a tall part that can cost more time than the cut itself. Operators compensate by lowering energy on tall sections and checking guide wear on schedule.

Cost is driven by machine hours, not by material removed. A part that takes 40 minutes of spark time costs roughly the same whether the blank weighs 200 g or 2 kg. That is the opposite of milling economics, where bulk removal dominates the quote.

Tolerance and finish come from the last passes, not the first. A roughing pass might leave 0.05 mm of stock and a Ra 3.2 μm surface. Skim passes, each removing a few micrometers, walk the surface down to ±0.005 mm and Ra 0.2–0.8 μm on the finest settings. Buying the finish means buying the time.

Thermal stability matters more than most shops admit. The machine, the dielectric, and the part should sit at the same temperature. A shop that runs a cut right after the part came off a warm truck will fight dimensional drift for the first hour.

Verification is done on the machine with a probe or off-line on a CMM. For tight bores, a plug gauge or an air gauge gives a faster pass or fail than a full CMM run. Reports can be supplied when the drawing calls for them.

Positioning accuracy of the machine itself sets the floor. Ball screws, linear scales, and guide alignment all contribute. A used machine with worn ways may still cut a clean surface but drift on position, which shows up as a tolerance band that widens across a long part.

Start with the datum. One edge and one hole as the XY zero, and top of stock as Z. Everything downstream, including the wire offset and the skim pass offsets, depends on that zero being real. Clamp the part so it cannot shift when the wire breaks.

Thread the wire through the start hole manually or with the auto-threader, then verify the offset direction. Cutting on the wrong side of the offset is the single most common scrap cause on a first article. Run a dry pass on graphics before energizing.

Stress relief before the finish pass is worth the extra day. Rough cutting releases locked-in stress and the part moves. If the tolerance is tight, rough, stress relieve, then take the skim passes on the settled part.

Keep the dielectric clean and the filters changed. Conductivity drifts as the resin depletes, and the spark energy changes with it. A conductivity meter reading outside the machine's window is a finish problem waiting to happen.