How to Calculate Cycle Time in CNC Machine Work

• 1
  Key takeaways
• 2
  What cycle time actually counts
• 3
  Cutting time formula and feed inputs
• 4
  Non-cutting time that gets forgotten
• 5
  Setup, load and unload time
• 6
  Checking the estimate against the machine
• 7
  When the calculation stops being reliable
• 8
  How to calculate cycle time in cnc machine work: 6 steps
• 9
  Cutting speed and feed starting points

Cycle time is the elapsed time from the moment the operator starts the cycle to the moment the part is released. It is not the same as spindle-on time, and it is not the same as the time a machine is occupied. On a 3-axis job with one setup, those three numbers are close. On a 5-axis job with probing and two fixtures, they can differ by 30% or more.

When you calculate cycle time in cnc machine planning, split the number into four segments: load and unload, cutting time, non-cutting machine moves, and tool changes. Add probing, dwell, and in-process measurement if the part needs them. Each segment has its own error sources, so estimating them together hides which one is wrong.

The cutting segment is the one most engineers trust. It is also the one most often overstated, because CAM defaults and catalog feeds rarely match the feed the operator dials in after the first article squeals. Treat the first calculation as a planning number, then correct it from the actual run.

One more boundary: cycle time per part only makes sense with a stated batch size. Setup time spread over 1 part and over 500 parts gives two very different numbers. Always write the batch size next to the cycle time, or the figure means nothing.

For each toolpath, cutting time equals path length divided by feed rate. Path length comes from the CAM model in millimeters. Feed rate comes from feed per tooth multiplied by tooth count multiplied by spindle speed. If you skip the multiplication and read the feed straight off a tool catalog, you are calculating a number no machine will run.

Spindle speed itself follows surface speed: rpm = 1,000 × cutting speed ÷ (π × tool diameter). In 6061 aluminium, carbide tools run at 300–500 m/min surface speed. In 304 stainless, drop to 90–150 m/min. In Ti-6Al-4V, 40–70 m/min. Those ranges are planning values, not promises; rigidity, coolant, and tool overhang move them.

Feed per tooth depends on radial engagement. A 10 mm carbide end mill in 6061 at full width might run 0.05 mm/tooth. At 25% radial width, 0.08–0.12 mm/tooth is realistic. Deep axial cuts pull the number down. If your chip load looks optimistic, the cutting time will look optimistic too.

A worked example: a 12 mm end mill, 4 flutes, 6061, 4,000 rpm, 0.06 mm/tooth gives 960 mm/min. A roughing path 3,200 mm long takes 3.3 minutes. Two finishing passes at 600 mm/min over 2,400 mm add 8.0 minutes. The same part in 304 stainless at 180 mm/min would take more than four times longer for the same paths.

Rapid moves look fast in the machine manual and slow in a real program. A 30,000 mm/min rapid sounds like nothing until you count 60 Z retracts at 150 mm each. That is 9,000 mm of travel, or 18 seconds of pure motion, before any chip forms. On a small part with many pockets, rapid and retract time often lands between 15% and 30% of the cycle.

Tool changes are the other silent cost. A 10-second change sounds harmless. Multiply by 14 tools and you lose 140 seconds per part. On a 5-minute cycle that is 45% overhead. Reduce tool count by combining features, or move low-usage tools to a second operation if the tolerance allows.

Dwells, spindle ramp-up, and spindle ramp-down add up on short cycles. A 2-second dwell after a face mill is 2 seconds every part. Over 5,000 parts that is nearly three hours of spindle time doing nothing. Delete dwells you cannot justify with a drawing note.

Probing and in-process measurement belong here too. A spindle probe touch is typically 2–5 seconds per point. If a drawing requires 12 probed points per part, that is 24–60 seconds before cutting time is added. Budget it separately, not inside the cutting estimate.

Load and unload time is the part of the cycle that changes most with the workholding method. A single vise with hard jaws on a 150 mm part might take 20–40 seconds. A custom fixture with eight clamps and a torque sequence can take 3–5 minutes. Same part, same spindle, very different cycle.

For pallet changers, the load happens off the machine. The machine-side number drops to the pallet swap, often 8–15 seconds. The operator-side number does not disappear; it moves to a separate station. If you quote only the machine-side number, you will understate labor even when the spindle is fully used.

Zero-point verification matters on first articles and on any part with a tight datum. Touching off a tool or confirming a fixture stop can add 1–3 minutes to the first piece and a few seconds to each following piece. Separate first-article time from steady-state cycle time in your plan.

Batch size changes this segment more than any other. Setup and first-article checks are fixed costs. They dominate a one-part run and nearly vanish on a 10,000-part run. This is why the same geometry can look expensive as a prototype and cheap as a production part.

CAM simulation is a useful second opinion, not the answer. It knows the toolpath geometry and the programmed feeds. It does not know that the operator reduced the feed by 20% because the tool chattered at full depth, or that the fixture needs a manual clamp check every 50 parts.

The cleanest check is the machine timer on the first article. Record the total, then record spindle-on time separately. The difference is your non-cutting and load time. If the difference is under 10%, the process is lean. If it is over 35%, look at tool count and Z retracts first.

Track the number over the first 20 parts. Break-in effects, chip clearing and operator rhythm change the result. A stable number after 20 parts is a fair planning figure. A number from part one is a guess.

Keep a simple record per part number: material, tool count, cutting time, total time, and batch size. After a few jobs, your estimates for similar geometry get much closer, and quoting stops being a coin flip.

Thin walls and tall features change everything. A part with a 0.8 mm wall in aluminium cannot be roughed at full chip load. The feed drops, sometimes by half, and the finishing passes multiply. Your formula still works, but the inputs must reflect the real strategy.

Hard materials and hard coatings have their own rules. Inconel and hardened tool steel run at 25–45 m/min and wear tools fast. Tool life, not feed rate, sets the cycle. Plan for more tool changes per part and shorter runs between offsets.

Five-axis simultaneous motion adds rotary acceleration and can force slower feeds through corners. A 4,000 mm path on a 3-axis machine and the same path on a 5-axis machine do not take the same time. If the part needs continuous 5-axis motion, add 15–30% to the cutting estimate and confirm in CAM.

Long parts and deep cavities stretch the numbers the other way. On a 4,000 mm travel machine, rapid positioning across the table can dominate the cycle. Plan your tool order to keep travel short, or the calculation will look fine on paper and slow on the floor.