How Studio cuts a size to a fit, how it allows for the cutter's runout, and how it measures a part with the touch probe. The twin's numbers here are estimates from its physical models; Studio claims no positioning accuracy for the machine.
A size can carry a fit: a hole's diameter ("d": "Ø12 H7"), a shaft's ("fit": "g6" on an outside contour), a slot's width, or a rectangle's X and Y sizes. Studio takes the fit's limit deviations from the ISO 286-2 tables for the size, or the limit deviations as you write them ("tol": "+0.015/0", in millimetres).
The wall is roughed with an allowance, then finished in a pass of its own to the middle of the fit's zone. Fits of grade IT6 and finer get a spring pass: the finishing lap again, at size, to take off what the tool's deflection left. The closed loop adds spring passes where the twin's estimate of the deflection asks for them.
The answer says, for each fit, the zone, the diameter it was cut to, and why.
| Name | Pair | Use |
|---|---|---|
sliding | H7/g6 | Turns or slides, small clearance |
locating | H7/h6 | The closest clearance fit: put together by hand |
transition | H7/k6 | A little clearance or a little interference: locating, light press |
press | H7/p6 | Interference: pressed in |
A pair names both parts: the hole takes the first, the shaft the second.
A cutter that runs out cuts a hole larger than itself. The twin models the runout of the tool in its holder as an envelope. For a size with a fit, Studio moves the finishing path inwards for a hole, outwards for a shaft, by that envelope, so that the size the twin estimates is still the middle of the zone. A reamed hole is not changed: it is the reamer's size.
The envelope is the twin's estimate for its model of the tool and holder. Measure the first part.
Each touch is a sequence of probing moves (G38):
The trigger point is the ball's centre. Studio adds the ball's effective radius in the touch's direction. The effective radius and the probe's lobing come from its calibration on a ring gauge of a known diameter.
A size without its own tolerance takes the general tolerance of the class asked for (f, m, c, v), from ISO 2768-1's table for its size range: linear sizes, and the heights and angles of chamfers.
When the twin's grid is coarse for a tolerance, the twin's estimate for that check is said to be too coarse to judge; the measured value is judged all the same.
The class H, K or L after the ISO 2768-1 class (for example "mK") adds form checks, judged by ISO 2768-2's tables:
| Check | Touches | Table |
|---|---|---|
| Flatness | A floor on a 3 × 3 grid; the spread of the points from their plane | 1 |
| Straightness | A long wall, 3 to 8 points along it; the spread from their line | 1 |
| Perpendicularity | A wall at a second depth; its lean over the wall's height, against the machine's Z | 2 |
| Circularity | A hole or boss at 8 points; the spread of the radii | 4 (the diameter's tolerance) |
Two touches on a wall can only be told apart when they are far enough from each other. A wall too short for that, for example under a chamfer of its edge, is listed with the reason instead of a value.
A probe calibrated only on its effective radius has lobing of its own. It shows in the circularity: the report says how large the twin's probe model takes it to be.
The twin cuts its simulated workpiece with the same program, then runs the same touches on it with the same probe model. Its estimate is what that calculation gives. It shows the program's own effect on each size, as far as the twin models it: the tool's deflection, its runout, the cutting forces.
The estimate is not the machine's measuring uncertainty. The report says nothing about how precisely the machine measures or positions.