The RTCP Function of Five Axes, From Principle to Shop Floor

• 1
  Key takeaways
• 2
  What the RTCP function of five axes actually corrects
• 3
  Which five-axis layout needs which RTCP variant
• 4
  The numbers that decide whether RTCP holds tolerance
• 5
  Turning RTCP on: G43.4, TCPM and what follows
• 6
  Where RTCP stops helping
• 7
  How to prove RTCP is working before you cut the part
• 8
  RTCP on vs RTCP off: what changes on the floor

A five-axis machine has three linear axes and two rotary axes. The rotary axes carry the workpiece or the spindle. When either one turns, the tool tip drifts away from the point the CAM system wrote. Without correction, the tip swings along an arc, and the part comes out with a gouge or a step where the surface should be smooth.

RTCP, short for Rotation Tool Center Point, makes the control hold the tip on the programmed point while the rotary axes move. The controller solves the forward kinematics for the current A, B or C angle, then offsets X, Y and Z by the amount the tip would otherwise shift. That offset is recalculated every interpolation cycle, not once per block.

The effect is easiest to see on a table-table machine. Spin the C table 90° with RTCP off and the tool tip sweeps a curve through the part. Turn RTCP on and the tip stays at the commanded point. The rotary axes still move. Only the linear axes pick up the difference.

This is why the function is often called the center point function rather than a compensation option. It does not correct tool wear or thermal drift. It corrects the geometry that the rotary axes introduce by design.

• 1
  Forward kinematicsThe control computes where the tip is, given the current axis positions.
• 2
  Inverse offsetIt then moves X, Y, Z so the tip returns to the programmed point.
• 3
  Per-cycle updateThe offset is refreshed at the interpolation rate, not per G-code block.

Not every five-axis machine uses the same correction. The layout decides what the control has to solve. A head-head machine tilts the spindle, so the pivot point sits inside the spindle head. A table-table machine tilts the workpiece, so the pivot point sits in the rotary table. A mixed layout has one rotary axis on the spindle and one on the table.

On a head-head machine the tool tip moves when the head tilts. RTCP offsets X, Y, Z to hold the tip at the programmed point. The rotary center is a fixed machine constant that you measure once with a dial indicator or a test bar. If that number is wrong, every tilted cut is wrong by the same amount.

On a table-table machine the workpiece rotates, so the tool tip appears to move relative to the part. Here the control has to work in two frames at once: the machine frame and the part frame. The pivot distance from the table face to the intersection of the two rotary axes is the critical value. Measure it with a gauge block stack or a ball bar, then store it in the machine parameters.

Mixed layouts are the most sensitive. The spindle-side rotary axis and the table-side rotary axis each carry their own offset. Both must be correct before RTCP can do anything useful. A 0.02 mm error in either offset shows up as a taper on a tilting cut.

• 1
  Head-headPivot inside the spindle head. One rotary center to measure.
• 2
  Table-tablePivot inside the rotary table. Pivot distance is critical.
• 3
  MixedTwo offsets, one on each side. Both must be correct.

RTCP only works as well as the data you feed it. Three values matter most: the tool gauge length, the pivot distance, and the rotary center offsets. Each one is a measured value, not a nominal one. A printed tool length from the holder catalog is a starting point, not an answer.

The gauge length is the distance from the spindle gauge line to the tool tip. Measure it on a presetter or with a tool setter on the machine. A 0.05 mm error in gauge length becomes roughly a 0.05 mm error at the tip on a tilted cut, and more when the tilt angle is large. On a 45° tilt the error projects directly into the surface.

The pivot distance is the distance from the rotary axis intersection to the mounting face. On a trunnion table it is often given by the builder, but it changes if the table is rebuilt or if a riser is added. Re-measure after any mechanical work. Store it with at least three decimal places.

Rotary center offsets are the X, Y and Z positions of the rotary axis intersection in machine coordinates. Find them with a dial indicator swept around the axis, or with a ball bar on a circular test. Write them into the work offset or the machine parameter, depending on the control. Do not carry them over from a different machine.

• 1
  Gauge lengthSpindle gauge line to tool tip. Measure on a presetter.
• 2
  Pivot distanceRotary intersection to mounting face. Re-check after rebuilds.
• 3
  Rotary offsetsAxis intersection in machine coordinates. Sweep with an indicator.

Fanuc controls use G43.4 for tool center point control, with G49 to cancel it. Some builders map it to M-code pairs instead. Siemens uses TRAORI, and Heidenhain uses TCPM or M128 depending on the generation. The name changes, the job does not. You enable the function, hand it the tool data, and it holds the tip.

The order matters. Call the tool length offset first, then the RTCP activation, then the rotary moves. Cancelling RTCP before the rotary axes return to zero leaves the tip in the wrong place for the next block. On most controls the safe sequence is: retract Z, cancel RTCP, index the rotary axes, re-enable RTCP if needed.

Feed rate behaves differently once RTCP is on. The control now moves linear axes to compensate for rotary motion, so the programmed feed is applied to the tip, not to each axis. That is usually what you want. It also means a large rotary move at a high feed can demand axis speeds the machine cannot reach. Watch the axis load meters on the first run.

Look-ahead matters too. Controls with a five-axis look-ahead buffer can pre-read the rotary motion and smooth the linear compensation. Without it, tight corners on a tilting cut show faceting. Check the block processing time and the look-ahead depth in the machine parameters before blaming the CAM output.

• 1
  FanucG43.4 on, G49 off. Tool offset first.
• 2
  SiemensTRAORI, with the tool data in the active frame.
• 3
  HeidenhainTCPM or M128, depending on control generation.

RTCP corrects the geometry of the rotary axes. It cannot correct the machine. If the A axis is not square to the Z axis, the error appears in the part whether RTCP is on or off. The function assumes a rigid, aligned machine. On a worn machine it simply moves the error to a different place.

Backlash is the classic case. RTCP commands small linear moves to compensate for rotary motion. If the linear axis has 0.01 mm of backlash, those small moves are partly lost. The result is a surface that looks fine in one direction and rough in the other. Check backlash before you tune RTCP parameters.

Thermal growth is the other limit. The spindle grows as it warms, so the gauge length changes during a long cycle. RTCP uses the gauge length you entered. If the machine has no thermal compensation, a long roughing cycle followed by a finishing pass can drift outside ±0.005 mm even with RTCP active.

Tool deflection is a third one. RTCP keeps the tip on the path, but if the tool bends under load, the cutting edge is not where the tip is. On thin ribs and deep pockets, keep the radial depth of cut modest and the tool as short as the geometry allows. RTCP cannot see deflection.

• 1
  SquarenessAxis alignment errors pass straight through RTCP.
• 2
  BacklashSmall compensation moves get lost in the lash.
• 3
  Thermal driftGauge length changes as the spindle warms up.
• 4
  DeflectionThe cutting edge moves even when the tip is correct.

Run a circular test. Mount a ball bar between the spindle and the table, command a circular path with the rotary axes tilting through the motion, and read the roundness and squareness values. A healthy machine holds the circle within a few micrometres. A large lobing pattern points at a wrong pivot distance or a bad rotary offset.

Cut an air pass first. With the part offset well clear, run the full program and watch the tip position on the control display. The commanded tip point and the actual tip point should track within the machine's stated accuracy. If they diverge during rotary moves, the setup values are wrong.

Then cut a test block. A simple 100 mm × 100 mm block with a 45° face and a contoured pocket shows most RTCP problems. Measure the face with a height gauge and the pocket with a CMM. Compare the measured profile to the CAD nominal. A constant offset means a setup value is wrong. A varying offset means a mechanical problem.

Keep a record. Note the gauge length, pivot distance and rotary offsets for each machine, plus the date they were measured. When a job drifts, you can check whether the data changed or the machine did. That record saves more time than any parameter tweak.

• 1
  Ball bar testCircular path with rotary motion. Read roundness and squareness.
• 2
  Air passWatch commanded vs actual tip position on the display.
• 3
  Test block45° face plus contoured pocket. Measure on a CMM.
• 4
  Setup logRecord the three key values and their measurement dates.