← Back to Field Notes

November 19, 2025 · The Bloomfield Team

The Top 5 Causes of Scrap in CNC Machining

The average CNC job shop scraps between 3% and 5% of produced parts. On a shop doing $10 million in annual revenue with a material-to-revenue ratio of 35%, that scrap rate represents $105,000 to $175,000 per year in wasted material alone. Add the labor, machine time, and opportunity cost of parts that never ship, and the true cost of scrap easily doubles.

The frustrating part is how concentrated the causes are. Across dozens of shops we have worked with, five root causes account for over 80% of all scrapped parts. Fix those five and the scrap rate drops by half.

For a broader look at quality metrics and how they connect to shop floor performance, see our guide to production visibility for manufacturers.

1. Setup Errors (28% of Scrap)

The largest single source of scrapped parts in CNC machining is setup-related: wrong offsets, incorrect work coordinates, mislocated fixtures, and datum reference errors. These mistakes produce parts that are dimensionally wrong from the first cut, and they often go undetected until the first piece comes off the machine and hits the CMM.

Setup errors concentrate in two scenarios. First runs on new jobs where the operator is working from a drawing and a process sheet without historical reference. And repeat jobs where the setup was modified since the last run and the change was never documented. Both scenarios share a common root: the operator lacks complete, current information about how the job should be set up.

Shops that maintain detailed setup documentation with photos, offset values, and fixture locations from the last successful run reduce first-piece scrap on repeat jobs by 40 to 60%. The documentation takes 15 minutes to create after a successful run and saves hours of rework on the next one.

2. Tool Wear and Breakage (21% of Scrap)

Cutting tools wear predictably. A carbide end mill in 4140 steel has a predictable life curve based on feeds, speeds, depth of cut, and material hardness. When the tool is changed at the right interval, parts stay in spec. When it runs past its useful life because the operator is busy and the tool life counter is not set, dimensions drift and surface finish degrades.

Tool breakage is more dramatic and more preventable. A broken tool in the middle of a finishing pass usually means a scrapped part. The causes are almost always documented somewhere in the shop's history: a specific tool in a specific operation at a specific feed rate that fails when the material hardness runs at the upper end of the spec range. That pattern exists in the data. Capturing it and surfacing it before the next run prevents the breakage from recurring.

3. Programming Errors (18% of Scrap)

G-code errors, wrong tool paths, missed operations, and incorrect speeds and feeds. Programming errors tend to produce catastrophic scrap rather than gradual degradation. A rapid move into the part surface. A tool change at the wrong Z height. A missing finish pass that leaves a feature oversized.

The standard defense against programming errors is simulation and first-article verification. Both work. The failure mode is rushing: the programmer is under time pressure, the simulation step gets skipped, and the first part becomes the simulation. On a $400 Inconel blank, that is an expensive test run. Shops that track which programs produced scrap on first run and why can build a review checklist that targets the specific error categories their programmers make most frequently.

4. Material Defects (10% of Scrap)

Hard spots in castings. Inclusions in bar stock. Material that meets the mill cert specification but machines differently than the last lot. Material-related scrap is the hardest to prevent because the root cause originates outside the shop. What the shop can control is incoming inspection discipline and historical tracking of material performance by supplier and lot.

A shop that tracks which material suppliers produce lots that machine consistently and which produce lots that cause problems has data to make purchasing decisions that directly reduce scrap. Over time, the correlation between supplier quality and scrap rate becomes clear enough to justify paying a premium for material that machines predictably.

5. Fixturing and Workholding (8% of Scrap)

Parts that move during machining. Insufficient clamping force. Fixtures that have worn beyond their original precision. Soft jaws that were cut for a different revision of the part. Fixturing scrap is almost entirely preventable because the variables are known and measurable. Clamp force can be specified. Fixture runout can be checked. Soft jaw dimensions can be verified against the current revision.

The failure mode is institutional. The fixture was built three years ago. The person who built it has left. The documentation does not specify clamp torque or indicate which revision of the part it was designed for. The current operator sets it up based on what looks right, and the part shifts 0.002" during a heavy roughing cut. That is $600 worth of material and four hours of machine time in the bin.

Turning Data Into Prevention

Every one of these five causes is preventable with the right information delivered at the right time. Setup documentation from the last successful run. Tool life data from historical production. Program verification records. Material performance tracking by supplier. Fixture specifications and revision history.

That information exists in most shops. It lives in job travelers, operator notebooks, the ERP, the quality system, and the memory of the people on the floor. The shops that pull it together into a system where operators can access the relevant data before they start a setup are the shops that run scrap rates under 2%. The shops that leave it scattered run scrap rates of 4 or 5% and accept it as the cost of doing business.

It does not have to be. The data is there. The patterns are clear. The question is whether the information reaches the operator in time to prevent the next scrapped part.