Estimating Cycle Time for New Parts

Every shop knows the sinking feeling: a new RFQ lands, the geometry looks straightforward, but you have zero history on this exact part. Guess the cycle time wrong and you either lose the job or lose money running it. For digital manufacturers running CNC mills, lathes, or powder-bed fusion machines, cycle time estimation is where quote accuracy lives or dies.

Why Cycle Time Estimation Breaks Down on New Parts

Historical data is the gold standard. When you’ve run a part family before, you know the actuals — tool changes, rapids, coolant delays, probe cycles, and the inevitable “wait, let me check something” moments. New parts have none of that. Estimators fall back on CAM simulation times, rule-of-thumb multipliers, or the classic “it looks like a 45-minute part” gut feel. All three fail differently.

CAM simulations often ignore non-cutting time: tool changes, pallet swaps, in-process inspection, operator interventions. Rule-of-thumb multipliers (“2x CAM time” or “add 30%”) don’t account for machine-specific quirks — a 20-tool ATC vs. a 60-tool chain, a high-pressure coolant system vs. flood, a probe routine that adds 3 minutes per setup. Gut feel is just bias wearing a shop coat.

Decompose the Operation Into Measurable Segments

Start by breaking the cycle into discrete, estimable buckets. This works whether you’re quoting a 5-axis impeller or a DMLS bracket:

• Setup & workholding: Fixturing time, indication, probing, first-article inspection
• Roughing: Material removal rate (MRR) based on tool engagement, stepover, depth of cut, machine power curve
• Semi-finish/finish: Smaller tools, lighter cuts, more passes, tighter stepovers
• Non-cutting machine time: Tool changes, rapids, spindle accel/decel, pallet swaps, chip clearing pauses
• In-process inspection: Probe cycles, CMM moves, operator gauge checks
• Post-process: Deburring, stress relief, heat treatment queue time, surface finishing

Each bucket can be estimated with physics-based inputs rather than feelings. MRR calculations use tool manufacturer data, machine torque curves, and your proven feeds/speeds library. Tool change time comes from your machine’s spec sheet (or a stopwatch on your floor). Probe cycles are deterministic — count the touches, multiply by your probe routine duration.

Build a Feeds & Speeds Library You Trust

If your estimators are still digging through tooling catalogs or guessing SFM for 17-4 PH at 0.125″ DOC, stop. Build a living feeds/speeds database tied to your actual tool holders, spindle interfaces, and coolant delivery. Include:

• Tool description, holder, and stick-out
• Validated SFM, IPT, DOC, WOC ranges for each material group
• Achievable MRR (in³/min or cm³/min) at those parameters
• Tool life expectations (minutes or parts per edge)
• Notes on chatter zones, vibration limits, thermal concerns

When a new quote arrives, your estimator selects the tool-assembly from the library, plugs in the engagement geometry, and gets a defended MRR number. Roughing time = stock volume / MRR. No guessing.

Account for Machine-Specific Non-Cutting Overhead

Two shops running the same toolpath on different machines will see different cycle times. A horizontal with a 2-pallet changer and 60-tool matrix behaves differently than a VMC with a 30-tool arm and no pallet. Document your machine constants:

• Tool change time (chip-to-chip)
• Pallet swap / workpiece load time
• Rapid traverse rates and acceleration profiles
• Spindle ramp-up/down to cutting speed
• Probe cycle duration (per touch point)
• Coolant spin-up / high-pressure delay

These are measurable. Run a test program. Time it. Enter the numbers once. Now every quote inherits your reality, not the CAM software’s idealized simulation.

Use Parametric Templates for Recurring Feature Types

Most shops see the same feature patterns repeatedly: pocket families, hole patterns, face milling operations, turned profiles, lattice structures for additive. Build parametric time templates for each. A pocket template takes inputs — volume, depth, corner radius, floor finish requirement — and outputs a time range based on your library data. A hole pattern template counts holes, diameters, depths, and tapping requirements.

Templates turn “estimate this pocket” into “select pocket template, enter 4 numbers, get 12.3 minutes.” They also force consistency across estimators. The senior guy and the new hire should produce the same cycle time for the same feature geometry.

Add Contingency Based on Uncertainty, Not Fear

Every estimate carries uncertainty. New material? Add 15-20%. Unproven workholding? Add 10-15%. First time running this machine configuration? Add 10%. But apply these to specific buckets, not the total. A 20% contingency on roughing MRR is defensible. A 20% contingency on “the whole job” is a fudge factor.

Track actual vs. estimated by bucket. Over time, you’ll see which buckets consistently run over and can tighten the contingency or fix the root cause (better workholding, updated feeds/speeds, operator training).

Close the Loop With Actuals Capture

Estimation improves only when you compare predicted vs. actual — by operation, by machine, by material, by estimator. Capture real cycle times from your machine monitoring or MES. Tag them with the quote number. Review weekly. The shops that reduce quote-to-cash cycle time and win more profitable work are the ones treating estimation as a feedback loop, not a one-way guess.

From Spreadsheet to System

You can start with a shared spreadsheet for feeds/speeds, machine constants, and feature templates. But as quote volume grows, the maintenance burden and version-control headaches compound. This is where a quoting engine with built-in process logic — like Solvi — pays off. It stores your validated process data, applies it automatically to new geometries, and learns from actuals captured on the shop floor. The result: cycle time estimates that hold up, quotes that convert, and jobs that run to margin.