What Is CNC EDM Machine?

• 1
  Key takeaways
• 2
  What is CNC EDM machine erosion actually doing
• 3
  Sinker and wire: two answers to the same question
• 4
  When EDM is the wrong process
• 5
  What tolerance and finish you can expect
• 6
  Materials that suit a CNC EDM machine
• 7
  How EDM fits into a production route
• 8
  Sinker vs wire EDM vs milling

A CNC EDM machine removes metal by spark erosion. A shaped electrode and the workpiece sit in a dielectric fluid, usually deionized water or hydrocarbon oil. The controller pulses voltage across the gap, roughly 0.01 to 0.05 mm, until the dielectric breaks down and a spark jumps. Each discharge lasts microseconds and reaches temperatures near 12,000 °C, melting or vaporizing a tiny volume of both electrode and workpiece.

The fluid then does two jobs at once. It quenches the melted crater, and it carries the eroded debris out of the gap before the next pulse. If flushing fails, the debris re-solidifies on the surface and the cut short-circuits. That is why every EDM cycle alternates on-time, off-time, and a retract move that lets fresh fluid enter.

Nothing touches the workpiece. There is no cutting force, no chatter, and no tool deflection. Hardened tool steel at 62 HRC erodes at the same rate as annealed 4140, because hardness is not the mechanism. Conductivity is. The workpiece must conduct electricity, which is why EDM handles metals and conductive carbides but not plastics or ceramics.

The controller tracks the gap voltage and adjusts servo position thousands of times per second. When the gap opens, the electrode advances. When a short is detected, it retracts. This closed loop is what separates a CNC EDM machine from a manual EDM, and it is what holds a cavity depth to ±0.005 mm across a long burn.

Sinker EDM, also called ram EDM or die sinking, uses a formed electrode that mirrors the cavity it burns. The electrode is usually graphite or copper, machined on a separate CNC mill to a size smaller than the finished cavity, because the spark gap adds material removal on all sides. This is the process for blind pockets, sharp internal corners, and deep ribs that no end mill can reach.

Wire EDM uses a thin brass or coated wire, typically Ø0.1 to Ø0.3 mm, that travels through the workpiece like a band saw. The wire never touches the part. It cuts a 2D profile with a kerf equal to wire diameter plus spark gap, and the upper and lower guides can be offset to cut tapered or ruled surfaces. This is the process for punch dies, extrusion dies, and thin slots.

The choice usually comes down to geometry. If the feature is a cavity with a floor, use sinker. If the feature is a through-profile with a constant or tapered wall, use wire. Some parts need both. A forging die might be wire-cut from a hardened block first, then sinker-burned for the deep impression.

Both machine types run unattended for long stretches. A wire EDM can cut overnight through 200 mm of hardened D2, and a sinker can burn a multi-cavity mold insert with automatic electrode changes. That unattended time is a real cost advantage, but it does not make EDM fast. It makes EDM predictable.

EDM is slow at removing bulk material. If a part starts as a 100 mm block and ends as a 20 mm plate, you should mill or saw that stock away first. Burning it off with a wire machine can take days. The practical rule is to rough with milling or waterjet, then use EDM for the final 0.2 to 0.5 mm of material and the features milling cannot reach.

Surface finish comes at a cost. A roughing burn leaves a recast layer, sometimes called the white layer, that is harder and more brittle than the base metal. It can measure 2 to 20 μm deep depending on the energy setting. For fatigue-critical parts such as aerospace brackets or medical implants, that layer must be removed by a finishing pass or by post-machining, or it becomes a crack initiation site.

The workpiece must conduct. Plastics, most ceramics, and glass cannot be EDM'd. You can sometimes burn a conductive coating, but that is a specialty process, not a production route. If your part is PEEK or alumina, EDM is off the table and you should be looking at diamond grinding or laser.

Electrode cost matters on sinker work. A complex cavity may need three or four electrodes, each machined on a separate mill, each consumed during the burn. For a one-off part, that tooling cost can exceed the EDM time. For a 10,000-part run, it amortizes to almost nothing. That is the volume threshold to check before you commit.

A well-controlled EDM process holds ±0.005 mm on position and cavity depth. That is the same figure we hold across our CNC floor, and it is what makes EDM viable for mold inserts, stamping dies, and medical tooling. The limit is not the spark. It is thermal drift in the machine and the accuracy of the electrode or wire path.

Surface finish depends on the energy per pulse. Low energy gives a fine finish, Ra 0.2 to 0.8 μm, but removes material slowly. High energy gives Ra 1.6 to 3.2 μm and cuts faster. Most shops run a roughing pass, a semi-finish pass, and two or three finishing passes to reach Ra 0.8 to 1.6 μm on a production part.

The recast layer is the hidden variable. A single roughing pass can leave a heat-affected zone that alters fatigue life. If the drawing calls for a specific surface integrity, we add a finishing pass that removes 5 to 10 μm and inspect the cross-section on a sample. That step is not optional on aerospace and implant work.

Inspection follows the same logic as our milling work. We check the electrode or wire path against the model, measure the first article, and monitor the burn in process. Final inspection is 100% before shipment, with reports on request. If a cavity has to hold ±0.005 mm, it is measured, not assumed.

Any conductive metal can be EDM'd. Tool steel and hardened die steel are the classic use case because they are difficult to mill after heat treatment. We routinely burn 4140, 4340, D2, and A2 at full hardness, along with 17-4PH stainless and Inconel for aerospace and energy parts.

Carbide is another good fit. Tungsten carbide is hard and electrically conductive, so EDM shapes it where grinding would be slow and costly. This is common for progressive die inserts and wear plates. Copper and beryllium copper also burn cleanly, which matters for electrode and mold work.

Aluminum is conductive and can be EDM'd, but it is rarely the best choice. The thermal conductivity pulls heat away from the spark, the cut is unstable, and milling is faster on almost every aluminum feature. We reserve EDM for aluminum only when the geometry demands it, such as a deep narrow slot in a soft fixture plate.

Titanium sits in the middle. It conducts well enough and EDM avoids the work hardening that makes titanium milling difficult, but the recast layer is a real concern for fatigue parts. If the application is critical, we plan a finishing pass and verify the surface before the part ships.

EDM rarely runs alone. On a typical tooling job, we mill the block, heat treat it, then wire-cut the outside profile and sinker-burn the internal cavities. The milling removes the bulk, the heat treat sets the hardness, and the EDM delivers the final geometry without disturbing the temper of the whole part.

The process sequence changes the tolerance stack. If you wire-cut before heat treat, the part may move during quenching. If you wire-cut after, you are cutting hardened material, which is exactly what wire EDM is designed for. The correct order is almost always: rough machine, heat treat, then EDM.

Lead time follows the same pattern as our other services. Quotation and free DFM analysis come back within 12 hours, production can start within 24 hours, and parts ship in 3 to 5 days. EDM adds time only when the burn itself is long, such as a deep cavity or a thick wire cut.

The engineering value is in the planning. A part that looks impossible to mill usually has one or two features that need EDM and a dozen that do not. Separating those features keeps the cost down and the lead time short. That is the conversation we have before the first chip is cut.