Introduction NcCycleSense reads your NC program, models how the control and the machine will actually move, and computes the cycle time.

• Drop in an NC program and get, in seconds, a cycle time estimate, where the time goes, and a report you can download.
• Your NC program is processed in your browser. It is not sent anywhere.
• Free, no registration.

Drop a program in and try it →

What is in the report

What is accounted for

A CAM system usually reports cycle time as path length divided by feedrate, assuming the machine runs at the programmed feed throughout. On the same machine, three different programs are dominated by completely different limits. The percentage is the share of cutting time in which that limit holds the speed down:

All three are built-in samples on the cycle time estimate page — one click runs them.

Axes and acceleration

• Per-axis velocity and acceleration limits: in a coordinated XY move, the faster axis is held back by the slower one.
• Acceleration profiles: trapezoidal, S-curve (jerk-limited) and snap-limited.

Path

• Look-ahead: like a real control, it reads hundreds of blocks ahead before deciding the current speed.
• Corner deceleration: judged per axis from the direction change between two segments.
• Arc speed limit: on a small radius, centripetal acceleration pulls the feed down.
• In-position check: the difference between G61 exact stop and G64 continuous cutting.
• Positioning: every G00 block decelerates to zero and confirms in-position — independently of G61/G64, which govern the interface between cutting blocks.

Control and machine

• Block processing time: how long the control takes to read and prepare one block.
• Non-cutting time: tool changes, spindle ramp-up, spindle orientation and dwells, counted where they actually happen.
• 5-axis simultaneous motion: when the tool tip runs a straight line but the rotary axes cannot keep up, the real feed drops. G43.4, G43.5, G68.2 and TRAORI are supported.
• Macros are actually evaluated: #100, WHILE, G65. In a family program the hole positions are computed, not written in the file.
• Canned cycles G73–G89 and CYCLE81–CYCLE85, including peck retract and clearance amounts (parameters 5114 / 5115, which are not visible in the program).
• Subprograms M98 / M99; on Siemens, calls by name or L number, with PROC parameters and MCALL. Including calls across files (across files needs the desktop version — the web version takes one file at a time).

3D toolpath

• Coloured by the dominant limit, so you can see which stretch is costing time. You can also switch to cutting / rapid.
• Click any segment to see its feed, its time and the matching NC line; click the NC source and the path jumps with it.
• Search for G02, M06, T3 or a line number.
• Playback runs to scale: the slow stretches really are slow, and a tool change holds for its own seconds.
• An STL can be loaded as the background part.

The tool diameter is indicative only; the tip position comes from the actual trajectory. No collision detection, no material removal simulation. The STL is not uploaded, exactly like the NC program.

Your own machine

The default is a configuration template: pick a configuration (3-axis, 4-axis, or a table-table, head-table or head-head 5-axis), then a class (compact and high speed, general, or large and heavy cutting). Every value in a template is inferred from a machine class — not one of them comes from your machine.

Put in what this machine actually does on the machine settings page. The first section there is air-cutting test programs: run them on the machine and copy back the cycle time the control shows. No parameter knowledge needed, and no stopwatch.

The numbers you can copy off a spec sheet or a parameter screen turn out to matter little; the ones that matter can only be measured. The rest of the parameters are in the downloaded .toml, each annotated with its Fanuc parameter number — edit the file directly if you want to change them.

Load the file on the estimate page. Configuration files are not uploaded, exactly like NC programs.

Scope

The support matrix for all three controls →

One row per thing a machine does, one column per control, computed or not at a glance. When a program uses a command that is not supported yet, the diagnostics section of the report says so. Need another control, or a lathe? Tell us — it affects the order we build things in.

Data handling

Only three separate pages send data: feedback, contact and account. None of them has a file upload field.

Accuracy

We do not guarantee the accuracy of the estimate. The result is useful for comparing options, finding where the time goes, and spotting programs that differ a lot from your existing estimate. Before you make it the sole basis for a quotation or an acceptance test, check it against your own machine.

Error comes from three places: machine parameters, shop-floor conditions (feed override, real tool changes and waiting), and variation in the program itself. The report states where this run's parameters came from.

The estimates are reconciled segment by segment against public measured machining records (the NIST manufacturing research platform in the United States), with cutting, rapid and non-cutting compared separately.

What is next

In progress

• Accuracy validation and improvement: continuing to find machines and measured data to compare against.
• More machine configuration templates.
• Feeding user-reported measured cycle times into the same reconciliation. They are collected today but not used yet.

Next

• More controls, and lathes. Fanuc, Siemens, and HEIDENHAIN series for now.
• Self-service calibration for the cutting side (acceleration, corners, arcs, spindle). The test programs exist; for now we work those numbers out for you.
• NC program optimisation suggestions.