What a Maryland CNC Processing Professional Actually Controls

• 1
  Why the Fifth Axis Changes Geometry, Not Just Speed
• 2
  Tolerance Stack-Up and What ±0.005 mm Really Means
• 3
  Material Behavior Drives Toolpath and Finish
• 4
  Workholding and Setup Count Decide the Real Cost
• 5
  Inspection Is Part of the Process, Not a Final Gate
• 6
  Where the Process Reaches Its Limits
• 7
  Which Process Fits the Part

A three-axis mill moves the tool along X, Y and Z. The tool axis stays vertical, so every feature must be reachable from above. A five-axis machine adds two rotary axes, usually A (tilt around X) and B (rotate around Y). The tool or the workpiece can now tilt and rotate while cutting, so the cutter approaches a surface from an angle instead of straight down.

That single change decides which parts are manufacturable in one setup. A deep cavity with drafted walls, an undercut, a port face at 40°, or a contoured blade all need the tool vector to follow the surface normal. On three axes you would reposition the part by hand, re-indicate it, and cut again. Each reposition adds setup error and floor time.

Five-axis simultaneous motion is not the same as 3+2 positioning. In 3+2 the rotary axes lock and the part is machined like a three-axis job from a new angle. It is faster to program and rigid, but the tool axis is fixed per orientation. Simultaneous motion keeps all five axes interpolating, which is what you need for ruled surfaces, impellers and sculpted molds.

The trade is stiffness and cost. A tilted tool creates a longer effective stick-out and side load, so chatter risk rises. A five-axis job that is not needed wastes money. The engineering question is always whether the geometry demands the extra two axes, not whether the machine is newer.

A stated tolerance of ±0.005 mm (±0.0002 in) is a capability limit under controlled conditions, not a default for every dimension. It applies to specific features, on specific materials, with stable temperature and a finishing pass. A 200 mm aluminium plate and a 20 mm hardened steel boss do not hold the same number.

Stack-up is where most arguments start. If a bore, its mating pin and the housing that carries both each carry ±0.01 mm, the assembly can drift 0.03 mm before any thermal effect. Good practice is to identify the two or three dimensions that govern fit, tighten only those, and leave cosmetic or clearance features loose. Tightening everything raises cost with no functional gain.

Thermal drift is measurable. Aluminium expands about 23 × 10⁻⁶ per °C, so a 100 mm part moves roughly 0.0023 mm per °C. A shop floor that swings 5 °C across a shift can eat half of a ±0.005 mm band. This is why inspection happens in a controlled room and why parts are measured at 20 °C reference.

The practical rule we use: tell us the fit, not just the number. If a bore must press onto a bearing, state the bearing class and the intended interference. We can then choose the process, the finishing pass and the gauge that actually control that fit.

• 1
  Tighten selectivelyOnly the dimensions that govern fit deserve the smallest band.
• 2
  Mind temperatureAluminium moves about 0.0023 mm per 100 mm per °C.
• 3
  State the functionA bearing fit tells us more than a bare ± number.
• 4
  Inspect at 20 °CMeasurements taken hot are not comparable to drawing limits.

Aluminium 6061-T6 cuts fast and holds fine detail, which is why prototypes and fixtures lean on it. The 7075 grade machines cleanly too but is less forgiving of poor chip evacuation, and thin walls spring back after the cut. We leave a finishing allowance and take a light pass to control that movement.

Stainless 304 work-hardens. If the tool rubs instead of cutting, the surface gets harder and the next pass is worse. The fix is a positive rake, a feed that stays above the work-hardening threshold, and enough coolant. Grades 303 and 17-4PH behave differently again: 303 is free-machining, while 17-4PH in the H900 condition is tough and rewards rigid setups.

Titanium Ti-6Al-4V (TC4) conducts heat poorly, so the cutting edge runs hot while the chip carries little away. Speeds drop, coolant flow matters, and tool wear is the pacing item. Inconel and Hastelloy push this further. These jobs are less about spindle speed and more about tool life planning and conservative depths of cut.

Plastics are a separate discipline. PEEK and POM machine well but move with temperature; ABS and PC can gum if the feed is too light. Delrin and PTFE cut cleanly with sharp tooling and air blast instead of flood coolant. Material choice sets the toolpath, the finish and the inspection plan together.

Machining time is visible. Setup time is where cost hides. Every new orientation means a new fixture, a new zero, and a new chance for error. On a five-axis machine a part with features on five faces can often be cut in one or two setups. On three axes the same part may need five. That difference usually outweighs the higher hourly rate.

Fixtures have to hold the part without distorting it. Thin-walled housings and long shafts are the classic problem: clamp pressure closes the part, the cut looks fine, and it springs back out of tolerance when released. Soft jaws, vacuum plates and sacrificial tabs are common answers. For a Ø400 mm rotary table, a well-balanced tombstone lets several parts run per cycle.

Access is the other constraint. A tool needs clearance for the holder, not just the cutter. Deep pockets and tall bosses can block the shank even when the tip fits. This is where DFM review pays for itself: we check the tool vector and holder envelope before quoting, not after the first article fails.

Our three plants run 127 high-precision machines, including 16 simultaneous five-axis centers, 16 mill-turn centers and 27 three-axis machines. The right answer is not always the biggest machine. A simple prismatic bracket belongs on a three-axis mill, and it will cost less there.

A dimension that is never measured is a hope. Inspection starts with the raw material certificate, continues with in-process checks on critical features, and ends with a final report before shipment. Every part we ship is inspected. Reports are available on request, and for medical and automotive work the documentation trail is part of the deliverable.

CMM programming follows the drawing's datum scheme. If the drawing names A, B and C, the inspection setup must use those same datums. When a shop measures from a convenient face instead, good parts can read as bad and bad parts can pass. Datum discipline is the quiet reason two shops disagree about the same part.

Surface finish is measured, not eyeballed. As-machined aluminium typically lands at Ra 1.6–3.2 μm. A finishing pass takes it to Ra 0.8–1.6 μm, and fine work reaches Ra 0.2–0.8 μm. If a drawing calls for a specific Ra, say which surface. Polishing the whole part is expensive and usually unnecessary.

First-article inspection closes the loop. We compare the report to the drawing, flag anything outside the band, and correct the process before the run continues. For regulated industries, the ISO 9001:2015, IATF 16949:2016, ISO 13485:2016 and ISO 27001:2022 systems define how that record is kept and protected.

Maximum part size is bounded by machine travel and by physics. Our largest envelope reaches 4,000 × 400 × 150 mm, and medium work fits 750 × 1,150 × 550 mm or 600 × 600 × 600 mm. Beyond that, accuracy per unit length starts to fight thermal drift and part weight. Long parts need support, and support can fight the toolpath.

Aspect ratio matters as much as size. A 300 mm deep pocket with a 6 mm cutter will deflect, no matter how good the machine is. Tool length-to-diameter ratios above roughly 5:1 demand reduced feed, a smaller radial engagement, or a different strategy such as helical interpolation. If the drawing allows a larger corner radius, say so — it can halve the cycle time.

Tolerance tighter than ±0.005 mm is possible on selected features with the right material and temperature control, but it should be justified by function. If a feature only needs to clear a mating part, a loose band saves money and reduces scrap risk. Tightening without a reason is the most common cost driver we see in incoming drawings.

Quantity changes the answer too. There is no minimum order quantity here; a single prototype and a 10,000-part run are both normal. Low volume favors machining. High volume with a stable design often favors casting or molding plus finish machining. We will say so when that is the cheaper path.