7 Critical CNC Controlled Mistakes That Are Costing Your Factory Thousands
Seven shop-floor errors that turn a good drawing into scrap, rework and late shipments. Written for process engineers, CAM programmers and buyers who sign off on machined parts. Read it to judge which of these problems already sits in your process, and what to change first.
Where the money actually leaks
Most losses in a CNC controlled process are not dramatic. They show up as one extra setup, one scrapped batch, one inspection report that nobody reads until the customer calls.
Tolerance treated as a suggestion rather than a spec
A drawing that calls ±0.01 mm on every feature is not safer. It is slower. The machine still has to hold the tightest number on the print, so cycle time, setup count and inspection load all track the worst tolerance on the part, not the average one. On a part with 40 features, one over-tighted bore can add a second operation and a CMM pass.
The opposite failure costs more. When a shop quotes to a loose tolerance but never verifies it, the parts arrive at assembly and do not fit. That is not a machining problem at that point. It is a schedule problem, and it usually gets solved with air freight.
Functional tolerances should be tied to what the part does. A bearing seat needs a real fit class. A clearance hole for an M6 bolt does not. GD&T lets you say which is which without tightening everything to prove you are serious. We run DFM analysis on every quote and flag features that are tighter than the function requires. That is part of the 12-hour quote turnaround.
Picking the wrong alloy, not just the wrong material
6061-T6 and 7075-T6 are both aluminum. They behave nothing alike on a machine. 7075 machines clean and takes a good finish, but it is more notch-sensitive and less weldable. 6061 is easier to hold flat on thin walls and cheaper to replace when a setup goes wrong.
The grade matters more than the family. Free-machining 303 stainless cuts with short chips and good surface finish, but it has lower corrosion resistance than 316L. A medical bracket that needs passivation and chloride resistance should not be quoted in 303 just because the shop has bar stock on the shelf.
Temper and condition also move the part after machining. Stress-relieved plate holds flat better than as-rolled. A part that measures good on the machine and bows two days later usually has residual stress, not a bad operator. Material certification should travel with the job, and we check incoming stock against the drawing before the first cut.
Workholding designed after the setup sheet is written
Every inconsistency in a batch traces back to how the part was held. A vise with 2 mm of jaw lift will tilt a part enough to throw a bore out of position on the second operation. The operator compensates, the next part runs clean, and the pattern never repeats. That is the worst kind of scrap, because you cannot see it in the CMM data.
Thin-wall parts need support, not more clamp pressure. Soft jaws bored in place, vacuum plates, and low-melt fixturing all exist for this reason. Choosing the right one is a function of wall thickness, material and how many parts you are running.
For production runs above a few hundred pieces, a dedicated fixture usually pays for itself in the first week. It removes the operator variable, shortens load time, and makes the process repeatable enough to trust the in-process check. We design workholding as part of the process plan, not as an afterthought on the shop floor.
Letting the CAM system pick the tool path
Default tool paths are safe, not efficient. A constant-engagement path on a deep pocket can cut cycle time by a third compared with a conventional offset pass, and it puts less heat into the tool. The CAM programmer has to know that, because the software will not volunteer it.
The same logic applies to entry and exit moves. Plunging straight into a corner work-hardens stainless and burns the insert. Helical entry and a lead-out arc cost nothing in programming time and add tool life you can measure.
Rest machining matters on parts with small internal radii. If the roughing tool cannot reach the corner, the finishing tool does the roughing work, and it is not built for that. Identifying those areas before the program runs is cheaper than finding them on the machine.
Tool path decisions also affect surface finish. A path that changes direction at every pass leaves visible witness marks. A consistent stepover with the same lead angle produces a uniform Ra that may let you skip a finishing operation entirely.
Coolant treated as a cost line instead of a process variable
Coolant does three jobs: it removes heat, it clears chips, and it lubricates the cutting edge. A shop that only worries about the first one ends up with chip recutting, poor finish and short tool life. Through-spindle coolant with the right pressure solves all three in deep-hole and deep-pocket work.
Concentration matters. A refractometer reading of 6 percent is not the same as 10 percent, and the difference shows up in tool life on titanium and stainless. Checking concentration weekly is a five-minute task that prevents a week of scrapped parts.
On aluminum, high-pressure coolant can be the difference between a clean chip evacuation and a re-cut chip welding to the flute. On cast iron, dry machining with air blast is often better, because the fines mix with coolant into a paste that clogs the sump. Matching coolant strategy to material is part of the process plan.
Thermal growth ignored until the last part of the shift
A machine does not hold ±0.005 mm all day by default. Spindle growth, ballscrew heat and ambient temperature all move the tool point. The first part of the morning and the last part of the afternoon are not the same part.
The fix is not to run slower. It is to warm up the machine, use a stable coolant temperature, and re-check the datum after the first hour of cutting. On long parts, a 2 °C change in shop temperature can move a 1,000 mm dimension by more than 0.02 mm.
Compensation on the control helps, but it only works if the machine is measured. We run a warm-up cycle before precision work and check critical dimensions at intervals through the run, not just at the end. That is how the first part and the last part both land inside tolerance.
Surface finish planned as a separate job
A perfect machined surface can still fail if the finishing step is treated as an errand. Anodizing adds a few micrometers and can round a sharp edge. Bead blasting hides tool marks but changes the dimension on a thin wall. If those steps are not in the process plan, the part that passed inspection leaves the shop out of spec.
Masking decisions belong in the drawing review. A conductive anodize on a grounding pad, a hardcoat on a wear surface, and a clear anodize on a cosmetic face are three different processes on one part. They also have different lead times.
The practical answer is to machine for the finish you actually need. A Ra 0.8–1.6 μm target can be reached with the right stepover and a sharp tool, no polishing required. A Ra 0.2–0.8 μm target may need a separate operation. Knowing which one the drawing calls for before the job is quoted prevents the rework that shows up after the parts come back from the finisher.
Material and finish choices that drive cost
Common pairings and where they go wrong.
| Material | Typical use | Watch for | Finish note |
|---|---|---|---|
| 6061-T6 | Brackets, housings | Thin-wall bowing | Clear anodize adds ~5 µm |
| 7075-T6 | Aerospace fittings | Notch sensitivity | Hardcoat for wear |
| 303 stainless | Shafts, fittings | Lower corrosion resistance | Passivation after machining |
| 316L stainless | Medical, marine | Gummy chips, tool wear | Electropolish for cleanability |
| Ti-6Al-4V | Implant, aerospace | Heat at the edge | No anodize on load faces |
| POM (acetal) | Insulators, bushings | Thermal growth | No plating, machine dry |
Questions engineers ask before releasing a job
How tight a tolerance can you actually hold in production?
Our stated capability is ±0.005 mm on critical features, with the right material, workholding and thermal control. That is a process capability, not a claim that every dimension on every part lands there.
Dimensions that need to sit at that level should be identified on the drawing so they get their own inspection plan. Loose features do not need the same attention, and treating them the same only adds cost.
When should I choose 5-axis machining over 3-axis?
When the part has features on multiple faces that would otherwise need two or three setups, or when the geometry has compound angles and contoured surfaces that a 3-axis machine cannot reach in one orientation.
For a simple prismatic part with features on two faces, a 3-axis machine with a good fixture is often faster and cheaper. We have 16 simultaneous 5-axis centers and 27 three-axis machines, so the choice is made on the part, not on what is available.
Do you review the design before quoting?
Yes. Every quote includes a DFM analysis. We flag tolerances that are tighter than the function needs, features that will be hard to hold, and material choices that will cause problems downstream.
Quotation and the DFM review come back within 12 hours. Production can start within 24 hours after that.
How do you handle confidential drawings?
Uploads are secure and confidential. An NDA is available on request, and we can work under your NDA template if you prefer.
We hold ISO 27001:2022, which covers information security across the quoting and production process.
What is the smallest order you will take?
There is no minimum order quantity. We run from a single prototype to 10,000+ part runs on the same process plan.
Prototype work is a good place to catch tolerance and workholding problems before they get expensive in production.
How is surface finish specified and verified?
Finish is specified as a Ra value, not as a description like smooth or matte. We work to Ra 1.6–3.2 μm as-machined, Ra 0.8–1.6 μm for high-finish surfaces, and Ra 0.2–0.8 μm where the drawing calls for it.
Verification is done with a surface roughness tester on the specified features, and results can be included in the inspection report.
Send the drawing, get a process review
We review tolerances, material and workholding before quoting, so the cost estimate reflects the part you actually need. Quote and DFM analysis back within 12 hours.
12-hour quoteFree DFM analysis100% inspectionNo minimum order
