Windnc Fagor CNC Controller: 7 Critical Secrets for Faster Machining
The Windnc Fagor CNC controller is an open-architecture control, so most of its speed lives in parameters that ship at conservative defaults. This page explains what each setting actually changes on the machine, which parts benefit, and when touching it makes scrap instead of cycle time.

In this article
- 1
- 2
- 3
- 4
- 5
- 6
- 7
Key takeaways
What the Windnc Fagor CNC controller changes about cycle time
A machine tool is only as fast as the decisions its control makes between two blocks of G-code. The Windnc Fagor CNC controller exposes those decisions as parameters: how many blocks it reads ahead, how it rounds a corner, how tightly it follows a programmed path. On a closed control you get a preset answer. On this one you get a number you can move.
That matters most on 3D contoured work. A mold insert, an impeller, an orthopedic plate: the toolpath may be a million short moves, and the control spends real machine time reading and planning them. On a straight two-axis turning pass the same parameters change almost nothing. Recognizing which of the two jobs is on the table is the first skill.
The rest of this page walks through seven settings, in the order they usually matter on the floor. For each one: what it does mechanically, what evidence tells you to change it, and what normally goes wrong when you push too far.
- 1Long block countsThousands of tiny G01 moves expose the planner's read speed.
- 2Corner-heavy geometrySharp direction changes force a speed decision at every vertex.
- 3Constant-engagement roughingFeed can rise and fall with radial depth without hurting the tool.
Look-ahead depth and NURBS interpolation
Look-ahead is a buffer. The control reads a set number of blocks ahead, computes the deceleration needed for each upcoming direction change, and decides whether the current feed is achievable. If the buffer is too shallow, the machine arrives at a corner still at full feed and then has to slam the brakes. You hear it as a thud, or see it as a witness mark on the part.
Factory defaults for look-ahead are set for the widest possible range of machines and materials. That means safe, and safe usually means short. Raising the buffer lets the control plan a smoother velocity profile through a series of corners, so the machine carries more speed instead of stopping and restarting at every vertex. On a dense 3D finishing pass this is often the single largest cycle-time lever.
There is a limit. A very deep buffer costs memory and planning time, and if you set it far beyond the actual block density, the gain flattens. Watch the axis load and the surface finish, not just the clock.
NURBS interpolation is the second lever. Instead of feeding thousands of linear segments, the CAM system outputs a spline and the control generates the path internally. Block count drops by an order of magnitude, so read time drops with it. The visible effect is a surface without faceting on organic shapes: a smooth blend instead of a staircase. Tolerance for the spline is set in the CAM output, and that tolerance, not the machine, usually decides how close the surface sits to nominal.
- 1Raise look-ahead in small stepsChange one increment, then re-cut a known test part.
- 2Match spline tolerance to the drawing0.01 mm is common for cosmetic surfaces; tighter for sealing faces.
- 3Keep G01 for prismatic workSplines buy nothing on straight walls and flat floors.
Adaptive feed control and the electronic handwheel
Adaptive feed control adjusts the programmed feed while the cut is running. The control watches spindle load, or axis following error, and raises feed where the cut is light and lowers it where the tool bites deep. In a roughing pass with a constant radial stepover, AFC adds little. On a long tapered wall, a fillet, or a pocket corner where engagement spikes, it is the difference between a chipped insert and a complete pass.
The tuning target is spindle load, not feed rate. Set a load ceiling around 70 to 80 percent of the spindle's rated continuous power and let the control find the speed. Set it at 95 percent and you remove the safety margin that absorbs a hard spot in the casting.
AFC also has a boundary. On a finishing pass where surface finish is the deliverable, varying feed changes the tool mark pattern. Many shops disable AFC for the last pass and keep it for roughing only. That is a normal, defensible choice. For finish cuts we hold Ra 0.8–1.6 μm on aluminum and steel, and a steady feed is easier to defend.
The electronic handwheel, or MPG, is the setup tool most operators underuse. In handwheel mode the control scales one click of the wheel into a defined increment, so a 0.01 mm step becomes a precise jog rather than a guess. On a five-axis setup with a rotary table, that incremental control is how you dial in a work offset on a curved datum without re-indicating the part. Setup time is cycle time you never get back.
- 1AFC for roughingUse it where engagement varies and the tool is at risk.
- 2Fixed feed for finishingKeeps the tool mark uniform across a cosmetic surface.
- 3MPG steps of 0.01 mmGood default for picking up a datum on a machined face.
Spindle orientation, rigid tapping and thermal compensation
Rigid tapping needs the spindle and the Z axis to stay synchronized through the whole cycle, including the reversal at the bottom of the hole. If the spindle orientation offset is off, or the reversal ramp is set too aggressively, you get torn threads or a tap that snaps on the way out. Slowing the reversal and returning through the same path costs a fraction of a second per hole and saves the tap.
Choose the tapping method by hole count and thread class. Floating tap holders forgive small synchronization errors and are the safer choice for a handful of holes in a fixture that may shift. Rigid tapping without a holder is faster and holds pitch depth better on a production run where the thread class is tight. Neither is universally better.
Temperature compensation is the slowest-acting of the seven. A spindle grows as it warms, and so does the ballscrew. On a long boring operation the tool position drifts by a few microns over the first hour. With compensation active, the control applies a correction curve based on spindle and axis temperature, so the first part and the twentieth part sit at the same position.
This matters on bores and bearing seats with a ±0.005 mm tolerance, and on any run long enough for the machine to reach thermal steady state. On a short five-minute cycle the drift never develops, and the compensation is doing nothing you can measure. Warm up the machine before the first cut on any tight job, regardless.
- 1Rigid tapping for volumeFaster pitch control, but the fixture must be rigid.
- 2Floating holder for low volumeAbsorbs sync error on a few holes.
- 3Warm-up cycle firstFifteen to thirty minutes of spindle rotation before tight tolerances.
G61 blending logic: precision against speed
G61 tells the control to stop exactly at the end of each block before starting the next. It is the most accurate mode and the slowest. On a sharp internal corner with a tolerance callout, that stop is what keeps the corner square instead of rounded. On a long contoured surface, it turns a smooth pass into a series of hesitations.
The usual mistake is running G61 across an entire program because one feature needed it. The control has no way to know that the other 90 percent of the toolpath is free to blend. Set the mode per operation, not per program.
The practical split on a five-axis job is straightforward. Use exact-stop modes on datums, sealing faces, bearing bores and any feature with a true position callout. Use a blending mode with a defined tolerance on the swept surfaces between them. That tolerance is a real decision: at 0.01 mm the control can round corners enough to keep feed up, and the deviation stays inside most general machining tolerances.
The engineering question is not which mode is better. It is which features actually carry a tolerance that the blending would violate. Read the drawing, mark those features, and switch modes at the operation boundary.
- 1Exact stopDatums, bores, sealing faces, sharp internal corners.
- 2Blend at 0.01 mmSwept 3D surfaces where the drawing tolerance is looser.
- 3Mode per operationNever leave one mode active across the whole program.
Which setting to touch, by part type
Match the lever to the geometry in front of you.
| Setting | Best-fit part | When to leave it alone |
|---|---|---|
| Look-ahead depth | Dense 3D finishing passes with many corners | Short prismatic cycles with few blocks |
| NURBS interpolation | Organic surfaces from a CAM spline output | Straight walls, flat floors, drilled plates |
| Adaptive feed control | Roughing with varying radial engagement | Cosmetic finishing passes |
| Electronic handwheel | Five-axis setup on a curved datum | Production runs already proven out |
| Rigid tapping sync | High hole counts, tight thread class | A few holes in a shifting fixture |
| Thermal compensation | Long runs with ±0.005 mm bores | Cycles under ten minutes |
| G61 blending | Datums, bores, sealing faces | Swept surfaces with open tolerance |
The trade you are actually making
If the feature carries a tolerance callout, keep the control in exact-stop mode and accept the slower pass. If the surface is cosmetic or the tolerance is open, blend and raise look-ahead, because the deviation stays inside what the drawing allows.
Questions engineers ask about controller settings
Do these settings transfer between machines with the same control?
No. Look-ahead depth and blending tolerance depend on the machine's rigidity, the servo tuning and the toolholder. A setting that works on a small 500 × 500 × 450 mm machine can chatter on a larger one.
Treat every number here as a starting point for a test cut on the specific machine, not a value to copy across the shop.
How do I know look-ahead is actually the bottleneck?
Watch the axis load meter and listen at the corners. If the machine decelerates hard at each direction change and the block count in the program is in the tens of thousands, look-ahead is a candidate.
If the machine runs smoothly and the cycle is simply long, the limit is somewhere else: tool engagement, spindle speed or the toolpath itself.
Can NURBS interpolation be used on any surface?
It needs the CAM system to output a spline rather than linearized points. If the post processor only emits G01, there is nothing for the control to interpolate.
Spline tolerance is set in CAM. A loose tolerance gives a smoother, faster surface that sits further from nominal, which is fine for cosmetics and not fine for a sealing face.
Is thermal compensation worth enabling on a short job?
Rarely. The drift develops over the first hour of spindle running, so a ten-minute cycle never sees it.
A warm-up routine before the first cut does more for a short tight-tolerance job than the compensation curve does.
Does adaptive feed control wear tools faster?
It raises feed where the cut is light and lowers it where the tool is loaded, so the peak load is capped. That generally extends insert life in roughing.
The risk is setting the load ceiling too high. Without margin, a hard spot in the material has nowhere to go.
How does GreatLight control these variables on production parts?
We run 127 high-precision CNC machines, including 16 simultaneous 5-axis machining centers, and hold ±0.005 mm with 100% inspection before shipment.
Process parameters are set per part during first-article prove-out, then locked for the run. Reports are available on request.
Send the drawing. Get a machining plan back.
Quotation and free DFM analysis within 12 hours, no minimum order quantity, and an NDA on request before any file moves.
12-hour quote100% inspection±0.005 mmNo MOQ