CNC for electrical components: what the drawing has to say

• 1
  What CNC for electrical components actually has to control
• 2
  When to machine the part and when to form it
• 3
  Materials and finishes that carry current
• 4
  Burrs, chips and cleanliness: the electrical risks
• 5
  Where the electrical and mechanical requirements fight
• 6
  How we run an electrical part from drawing to shipment
• 7
  Conductor material and finish by application

An electrical component is asked to do two jobs at once. It has to carry or interrupt current, and it has to hold a mechanical position for the life of the product. CNC for electrical components is mostly about the second job, because the first one depends on it. A contact that sits 0.05 mm too deep loses normal force. A busbar hole that is 0.1 mm off center puts a bending load on a terminal screw.

Three variables decide whether a machined electrical part works in the field. The first is conductor geometry: contact width, hole position, the flatness of a mating pad. The second is surface condition: burrs on a contact edge, plating thickness on a terminal, foreign metal embedded in an insulator. The third is material choice, which sets both the current path and the stiffness of the part around it.

Copper and brass move when you cut them. A 30 mm copper busbar can spring 0.02–0.05 mm after a heavy face milling pass because the rolled-in stress releases as material is removed. Rough the part, let it rest, then take the finishing cut. On thin-walled connector shells, the same effect shows up as ovality in the bore, not as a length error.

Tolerances follow function, not habit. A connector shell bore that locates a pin needs plus or minus 0.01 mm. A mounting flange that only has to clear a screw boss can sit at plus or minus 0.1 mm and cost a third as much. Splitting a drawing into functional and non-functional features is the single biggest cost lever we see on electrical parts.

Insulators are a different problem. PEEK, POM and PA machine cleanly, but they hold a burr at the edge of a slot and they are sensitive to coolant. Filled grades such as carbon fibre reinforced PEEK wear tools quickly and leave a fuzzy edge that has to be trimmed by hand. If the drawing calls for a sharp internal corner in a filled plastic, plan for a small radius instead.

• 1
  Conductor geometryContact width, hole position and pad flatness set the electrical interface.
• 2
  Surface conditionBurrs, plating thickness and embedded particles cause most field failures.
• 3
  Material behaviorCopper and brass release stress during cutting and move after the finish pass.

A lot of electrical parts start life as sheet metal. Busbars, brackets, shielding cans and terminal strips are punched and bent in seconds, and that is the right process when the profile is 2D and the thickness is under about 4 mm. Machining enters the picture when the part stops being a flat shape.

The crossover usually comes from one of four features. A pocket that has to sit at a controlled depth. A boss with a threaded hole on a face that is not normal to the sheet. A contact surface that has to be flat within 0.02 mm. A cooling passage that runs inside the body. Any one of these pushes the part toward milling rather than stamping.

Volume decides the rest. Stamping needs a die, so the tooling cost only pays back above a few thousand pieces. Below that, a machined version is cheaper per part and can be revised between builds. We run both, and the honest answer for a 200-piece electrical enclosure is almost always machining, even though the unit price looks higher at first glance.

Five-axis work matters here because electrical housings are rarely open on all sides. Angled connector exits, deep pockets around a PCB, and O-ring grooves on a sloped face are all single-setup jobs on a simultaneous 5-axis machine. Every extra setup adds a re-datum error of roughly 0.01–0.02 mm, and that error lands on the electrical interface.

One boundary is worth stating plainly. If the part is a pure flat conductor with two bends and 20,000 pieces a year, machining is the wrong answer and we will say so. The process should follow the geometry, not the other way around.

• 1
  Machining wins3D pockets, angled faces, flat pads, internal channels, low volume.
• 2
  Forming wins2D profiles under 4 mm thick at volumes above a few thousand pieces.
• 3
  Avoid extra setupsEach re-datum adds about 0.01–0.02 mm of position error.

Copper C101 and C110 are the default for busbars and contact bodies. They machine soft, they weld easily to a terminal, and they conduct at about 100 percent IACS. The catch is gumminess. Sharp tools, high rake angles and generous coolant flow keep the edge clean. Beryllium copper C17200 is the choice when the contact has to act as a spring, and it must be machined in the annealed state, then age hardened to reach its final strength.

Brass C36000 is the free-machining grade and gives the best surface finish per minute of cycle time, which is why it dominates connector pins and threaded shells. C27400 and C28000 turn up in terminals where a bit more strength is needed. Aluminium 6061-T6 covers housings and heat spreaders, and 7075 handles brackets that see vibration. For high-current work in tight space, C101 with a silver plate is common.

Plating is where electrical parts quietly fail. Electroless nickel at 3–8 μm gives a hard, uniform layer on a complex shape. Silver plating keeps contact resistance low but tarnishes, so it belongs on mating surfaces rather than exposed frames. Gold plating is used on low-signal contacts at 0.5–2 μm over a nickel strike, and the nickel layer is not optional. Without it, gold diffuses into the copper base and the contact resistance climbs over time.

Anodizing on an aluminium housing is an insulator, and hardcoat is a better one. If a housing has to be anodized and also provide a ground path, mask the ground pads before the bath or press in a separate contact. Conductive anodizing exists but its resistance is higher than bare aluminium, so treat it as a shielding surface and not as a current path.

For housings that must not conduct at all, PEEK and POM are machined directly. PEEK holds its shape at 250 °C, which is why it shows up in power electronics. POM is cheaper and stiffer but softens much earlier. Neither one tolerates a sharp internal corner well, so design with a radius.

• 1
  C101 and C110Busbars and contact bodies. Soft, gummiest to cut, best conductivity.
• 2
  C36000 brassConnector pins and threaded shells. Free machining, good finish.
• 3
  Plating stackNickel strike under gold. Silver on mating surfaces only, not exposed frames.

A burr is a mechanical defect that becomes an electrical one. On a contact edge, a 0.05 mm burr changes the contact area and can bridge two adjacent terminals inside a housing. On a threaded shell, a rolled-over thread flank raises the torque needed to seat a mating connector and can damage the plating on the mating part.

Chips left inside a cavity are worse than burrs because they move. A single aluminium chip in a sealed housing can migrate to a PCB pad and create a short that only appears after thermal cycling. We treat chip removal as a process step, not a cleanup task: high-pressure coolant through the tool, an air blast before the part leaves the vise, and a final wash before inspection.

Deburring method depends on the feature. Hand deburring with a ceramic blade works on external edges but cannot reach an internal cross-hole. Thermal deburring removes material from every edge at once and suits small brass and stainless parts with many intersections. Electropolishing smooths stainless contact surfaces and removes the micro-burrs that hand work leaves behind.

Where a burr is genuinely not allowed, say so on the drawing with a note rather than a general tolerance. We inspect the edge under magnification and report the result. A blanket note such as all edges deburred is ambiguous, and it usually means the critical edge gets the same treatment as a cosmetic one.

Cleanliness also covers the coolant. Some coolant chemistries leave a residue on copper that shows up as a dark film after a few days in storage. Parts that ship to a customer who will solder or weld them get a final rinse and a dry pack, not an oil film.

• 1
  Chip migrationA loose chip inside a sealed housing causes shorts after thermal cycling.
• 2
  Thermal deburringRemoves burrs from every edge at once on small brass and stainless parts.
• 3
  Note critical edgesFlag the contact edge specifically instead of writing a blanket deburr note.

Two requirements pull against each other more than any other pair on electrical parts: contact flatness and part stiffness. Making a conductor pad flat within 0.02 mm usually means a light finishing pass, which leaves a thin section. A thin section flexes, and flexing changes the contact force over temperature. The fix is to leave a stiffening rib away from the contact area rather than thickening the whole part.

The second conflict is plating thickness against thread fit. A 10 μm nickel plate on an M3 threaded hole shrinks the minor diameter enough to change the torque curve. Threads that will be plated should be cut slightly oversized, or the plating should be masked out of the thread entirely.

The third is thermal expansion. Aluminium moves about 23 × 10⁻⁶ per °C, copper about 17 × 10⁻⁶, and PEEK about 47 × 10⁻⁶. A housing and an insert that fit at 20 °C can interfere at 85 °C if the expansion is ignored. In a press-fit insert, that difference decides whether the joint loosens or cracks.

None of this is exotic. It just has to be decided before the first chip is cut, because a fix after plating is expensive. This is the part of the job where a drawing review pays for itself: an hour of DFM work against a scrapped batch of plated copper.

• 1
  Flatness versus stiffnessAdd a rib away from the contact pad instead of thickening the whole part.
• 2
  Plating versus threadsCut plated threads oversized or mask them out.
• 3
  Thermal expansionAluminium, copper and PEEK move at very different rates between 20 °C and 85 °C.