Common problems in CNC machining of metal parts

• 1
  Symptom, likely cause, and what to do
• 2
  Why CNC machining of metal parts problems cluster in five areas
• 3
  Chatter and poor surface finish
• 4
  Burrs on edges, holes, and threads
• 5
  Tool wear and unexpected breakage
• 6
  Warping and dimensional drift
• 7
  Hole size error and location error
• 8
  A step-by-step check when a part goes wrong

Almost every rejected metal part we see traces back to one of five areas: vibration, edge condition, tool life, residual stress, or hole accuracy. They are not separate subjects. Chatter raises tool wear. Tool wear changes hole size. Heat from a dull tool feeds residual stress. Fix one and the next one often gets better.

The short version: put stiffness into the setup, put the right edge geometry on the tool, and keep heat out of the part. Everything below is a way of doing one of those three things. If you are debugging a specific print right now, the table above is enough to start.

The material matters more than most people expect. A 6061-T6 bracket behaves nothing like a 17-4PH shaft. Aluminum cuts clean at high speed but galls and loads the flutes. Stainless work-hardens if you let the tool rub. Titanium TC4 moves under heat and burns tools that dwell. Inconel will work-harden under a light pass and destroy a carbide edge in minutes.

• 1
  Aluminum (6061, 7075)High speed, sharp polished flutes, air blast or mist, watch for built-up edge.
• 2
  Stainless (304, 316L, 17-4PH)Positive rake, steady feed, never let the tool dwell or rub.
• 3
  Steel (1045, 4140)Flood coolant, coated carbide, watch for thermal growth on long runs.
• 4
  Titanium (TC4) and InconelLow speed, heavy feed, sharp edge, generous coolant, expect short tool life.

Chatter is a self-reinforcing vibration between the tool and the workpiece. Once it starts, the tool marks the wall, the wall pushes the tool further, and the marks get deeper. You hear it before you measure it. On a 6061 part running Ra 0.8–1.6 μm you will see it as regular ripples spaced along the cut.

The most common cause is tool overhang. Every extra 10 mm of stick-out costs stiffness. If a Ø10 mm end mill hangs out 60 mm, it will sing on any decent radial depth. Shorten the holder, use a shrink-fit or hydraulic holder, and the same cut goes quiet without slowing the spindle.

Workholding is the second cause. A part held only in a vise with 40 mm of unsupported material above the jaws will vibrate no matter how good the tool is. Add a toe clamp, a jack under the overhang, or machine from a soft-jaw pocket that supports more of the part.

If the setup is already stiff, look at the cutting parameters. Radial depth of cut drives chatter more than speed does. Dropping radial engagement from 50% to 20% of tool diameter often clears the noise, but you must raise feed per tooth to keep the chip load. Otherwise the tool rubs and you trade chatter for wear.

• 1
  Check overhang firstKeep it under 4× diameter wherever the geometry allows.
• 2
  Support the partA jack or toe clamp under the overhang kills most low-frequency chatter.
• 3
  Reduce radial engagementTry 20–30% of tool diameter before you touch spindle speed.
• 4
  Keep chip loadIf you reduce radial depth, raise feed per tooth to avoid rubbing.

Burrs are the material that did not leave cleanly at the exit of the cut. Soft aluminum and ductile stainless burr the worst. Hardened steel burrs less, but what it leaves is harder to remove. A burr on a mating face is a fit problem. A burr inside a hydraulic hole is a contamination problem.

The cheapest fix is to design the edge so the burr never forms. A 0.3–0.5 mm chamfer on the exit edge makes the tool push material off instead of tearing it. On a hole, add a chamfer at both ends if the print allows. This costs nothing at the design stage and saves a manual deburring operation.

When the edge cannot be changed, control the exit. Reduce feed by 30–50% in the last 1 mm of the cut. On a drilled through-hole, back the feed off as the drill breaks through. On a milled edge, climb milling leaves a smaller burr than conventional milling on most steels.

For parts that cannot have any burr, plan a deburring pass as a real operation, not an afterthought. A chamfer tool run at 8,000–12,000 RPM with a light axial pass removes the burr and leaves a controlled edge. Threads get a similar treatment: a chamfer before tapping, and a tap with the right chamfer lead for the material.

• 1
  Add a chamfer0.3–0.5 mm on the exit edge is usually enough.
• 2
  Slow the exitCut feed 30–50% in the final 1 mm of the pass.
• 3
  Climb millSmaller burr on most steels than conventional milling.
• 4
  Plan deburringTreat it as an operation with its own tool and parameters.

Tool wear is normal. Tool wear you did not plan for is a problem. A carbide end mill in 6061 can run for hours. The same tool in 17-4PH may last 30 minutes. If your tool life changes suddenly, something in the process changed: material batch, coolant concentration, or the holder.

The single biggest cause of early wear is running too fast for the alloy. Aluminum wants surface speed in the 300–500 m/min range. Stainless and steel want 100–200 m/min. Titanium TC4 wants 40–60 m/min with heavy feed. Inconel wants even less. Running a stainless part at aluminum speeds will burn the edge in a few minutes.

Coating matters as much as speed. An uncoated tool is fine for aluminum. For stainless and steel, an AlTiN or TiAlN coating survives the heat. For titanium and Inconel, an AlCrN coating with a sharp edge holds up better. The wrong coating can make wear worse, not better.

Watch the chip. A chip that comes off blue or black means the cut is too hot. A chip that comes off as dust means the feed is too low and the tool is rubbing. A chip that comes off as a proper 6 or 9 shape means the parameters are close. Read the chip before you read the tool.

• 1
  Match speed to alloy300–500 m/min for aluminum, 100–200 for steel, 40–60 for titanium.
• 2
  Match coating to alloyUncoated for aluminum, AlTiN for steel, AlCrN for titanium.
• 3
  Read the chipBlue or dusty chips mean the parameters are wrong.
• 4
  Check coolantConcentration and flow both matter on long runs.

A part that measures correct in the machine and wrong on the bench has a stress problem. When you remove material from one side of a rolled or extruded bar, the internal stress balances out by bending the part. The thinner the part, the more it moves. A 2 mm aluminum plate can bow 0.3 mm after one face is machined.

The fix starts before the machine. Stress-relieved stock moves less. For critical parts, rough machine with 0.5–1 mm of stock left, then stress-relieve again, then finish. This adds a step but it is often the only way to hold ±0.005 mm on a thin plate.

In the cut, balance the material removal. Take equal stock off both faces. Do not machine one side to final size and then flip. If the part has a lot of pocketing on one side, do roughing passes on both sides before any finishing pass. This keeps the stress balanced as long as possible.

Heat is the other driver of drift. A spindle that has been running for two hours is longer than a cold one. On a run of 50 parts, the first part and the last part can differ by 0.02 mm just from thermal growth. Warm up the spindle for 15–20 minutes, and re-check your offsets halfway through a long run.

• 1
  Use stress-relieved stockAsk the mill for it on thin or critical parts.
• 2
  Rough, relieve, finishLeave 0.5–1 mm, relieve, then take the final pass.
• 3
  Balance both facesEqual stock removal keeps the stress symmetric.
• 4
  Warm up the spindle15–20 minutes before the first part, re-check offsets mid-run.

Holes are where tolerance stacks up fastest. A drill is a flexible tool. A Ø6 mm drill at 5× diameter depth will walk and cut oversize. Add thermal growth and the hole you get is not the hole on the print. This is why a reamed hole holds ±0.005 mm while a drilled hole often does not.

The first fix is the tool. Use a stub or screw-machine length drill when the depth allows. Spot drill to establish location, but keep the spot shallow, about the diameter of the drill point. A deep spot can pull the drill off-center instead of guiding it.

The second fix is the operation sequence. Drill undersize by 0.1–0.2 mm, then ream or bore to final size. On a hole with a tight location tolerance, bore it in the same setup as the mating face. Moving the part to a second machine adds a setup error that no amount of tool care will fix.

Coolant and speed also move hole size. A hot drill cuts oversize. On stainless and titanium, use through-tool coolant if the holder allows it, and keep the speed in the recommended range for the alloy. If hole size drifts across a run, measure the first, middle, and last part. The pattern tells you whether it is thermal or tool wear.

• 1
  Use a short drillStub or screw-machine length when depth allows.
• 2
  Shallow spotDepth about equal to the drill point diameter.
• 3
  Drill undersize, then reamLeave 0.1–0.2 mm for the reamer or boring bar.
• 4
  Keep one setupBore critical holes in the same setup as the mating face.