1. Problem Overview
Machined parts produced on a 5-axis SINUMERIK 840D-controlled machine tool show visible chatter marks specifically around the rounded corners produced by circular interpolation or G1/G2/G3 corner-rounding segments in the NC program. The same NC program runs cleanly on an identical machine in the fleet, eliminating the part program and post-processor as the root cause. The defect is mechanical-optical, not geometrically out-of-tolerance at the corner, and the symptom typically traces to one of four families:
- Trajectory jerk on the B/C rotary axes at small radius transitions
- Inadequate or contaminated axis lubrication (linear guides, rotary axes, ballscrews)
- Ballscrew wear or preload loss on one of the five axes
- Drive-controller tuning mismatch (current loop, speed loop, Kv factor) following a recent axis swap
This reference walks through the diagnostic logic in the order Siemens field service typically applies it: NC-side smoothing first (cheapest change), then lubrication, then mechanical, then drive tuning last.
HMI / Setup / Save data or transfer the series-commissioning archive to the service PG so a parameter can be rolled back in seconds. Sinumerik 840D sl documentation is available on the official Siemens Industry Online Support portal.2. Symptom Analysis: What Chatter Patterns Reveal
Before touching any parameter, record the chatter pattern with the spindle stopped (a sharp marker laid on the workpiece traces the waviness directly). The wavelength and orientation localize the source.
| Chatter wavelength | Direction relative to feed | Likely source |
|---|---|---|
| Fine, periodic (~0.2-0.5 mm pitch) | Perpendicular to feed at corner | B/C rotary axis jerk; trajectory discretization |
| Coarse, regular | Aligns with one linear axis travel | Sticking-stick-slip on linear axis ballscrew or guide |
| Random, irregular | Any | Toolholder/runout, spindle bearing, fixture resonance |
| Localized to one face | Coincident with B-axis tilt | Rotary axis bearing preload or worm-gear backlash |
If the waviness appears only when the B or C axis is repositioning under load at the corner, the most probable causes are NC-side smoothing and rotary-axis dynamics. If the waviness persists on straight segments when only linear axes are moving, ballscrew and lubrication rise to the top of the list.
3. Root Cause Categories
Each root-cause family below has its own verification path. Walk through them in this order to minimize unnecessary disassembly.
- NC trajectory smoothing (jerk limits, compression tolerances, smoothing mode)
- 5-axis transformation parameters and orientation smoothing
- Lubrication system (volume, interval, oil grade, distribution block)
- Ballscrew health (preload, backlash, nut wear)
- Drive loop tuning (axis Kv, jerk filter time, acceleration limits)
- Mechanical resonances (linear guide rollers, rotary axis bearings)
4. SINUMERIK 840D Machine Data Parameters for Smoothing
The following machine data (MD) directly govern corner smoothing and jerk behavior on a SINUMERIK 840D sl. All identifiers use the long form with the softkey alias for the HMI Setup menu.
| MD identifier | Description | Typical default | Field-action |
|---|---|---|---|
| MD32400 $MA_AX_JERK_ENABLE | Enable axis-specific jerk limiting | 1 | Verify =1 on all five axes; set to 1 if previously 0. |
| MD32402 $MA_AX_JERK_MODE | Jerk filter mode (filter/characteristic) | 1 | Switch from characteristic to filter mode for smoother transitions. |
| MD32410 $MA_AX_JERK_TIME | Jerk filter time constant [s] | 0.01 | Increase in 0.005 s steps; typical window 0.015-0.040. |
| MD32420 $MA_JOG_AND_POS_JERK_ENABLE | Jerk limiting in JOG / positioning | 1 | Enable to match JOG and AUTO dynamics. |
| MD32430 $MA_COMPRESS_POS_JERK_ENABLE | Jerk limit active during compression | 1 | Must =1 when using COMPCAD/COMPCURV. |
| MD20150 $MC_GCODE_RESET_VALUES | Default G-code after reset | — | Confirm G64 / SOFT settings match the program header. |
| MD20480 $MC_SMOOTHING_MODE | Smoothing mode for path | 1 | Switch between 0=OFF, 1=mean-value, 2=soft. |
| MD20482 $MC_SMOOTHING_VECTOR_LENGTH | Smoothing vector length [mm] | 0 | Set to 1-3 mm for tangential smoothing. |
| MD20170 $MC_COMPRESS_BLOCK_PATH_LIMIT | Compression minimum path length | 1.0 | Reduce to 0.5 to merge short corner blocks. |
| MD32300 $MA_MAX_AX_ACCEL | Max axis acceleration [m/s²] | Axis-specific | Verify not reduced below drive capability. |
| MD32310 $MA_MAX_ACCEL_OVL_FACTOR | Acceleration override factor | 1.2 | Range 1.0-1.5; lower reduces corner overshoot. |
| MD24100 $MC_TRAFO_TYPE_… | 5-axis transformation type | 148 / 148 | Re-validate; mismatch between transformation MD and kinematic is a frequent corner-error source. |
4.1 Example parameter change procedure
- Switch to
Commissioning / Machine dataon the HMI. - Locate MD32410 ($MA_AX_JERK_TIME) for axis B, increase from 0.010 to 0.020.
- Apply, perform NCK reset, run a test part.
- If improvement insufficient, repeat on axis C, then re-evaluate feedrate.
5. Jerk Limiting and 5-Axis Trajectory Smoothing
In 5-axis contouring, the position controller commands simultaneous motion of three linear and two rotary axes. At a rounded corner, the path velocity is held while the tool vector rotates; the rotary axes must execute large angular deltas over a short tangential path. Without jerk limiting, the demanded acceleration derivative exceeds what the drive can supply, and the rotary axis lags — manifesting as chatter traces on the part.
Theoretical background on jerk-constrained feedrate profiling for NURBS and G1 corner-rounding is documented in 5-Axis tool path smoothing based on drive constraints (Senker et al.) and Feedrate interpolation with axis jerk constraints on 5-axis NURBS and G1 tool path (Tulsyan et al.). Both confirm that drive-limited jerk is the dominant excitation mechanism for chatter at corner transitions.
5.1 Verifying orientation smoothing (ORISON, SWIVEL)
SINUMERIK 840D sl provides orientation interpolation modes. Confirm the active mode matches the part program:
| Orientation mode | NC syntax | Behavior at corners |
|---|---|---|
| Linear interpolation of vector | ORIVECT, A3=0 | Sharp corner in tool vector; high rotary jerk |
| Great-circle interpolation | ORIVECT, A3=1 | Smooth tool vector; lower rotary jerk |
| Smoothing of orientation | OSCTRL / OSNSC | Active tool-vector smoothing around corners |
For a chatter complaint at corners, switch ORIVECT to A3=1 or activate OSCTRL with an angular tolerance of 0.5-2.0°.
6. Mechanical Inspection: Lubrication System
Field experience on 5-axis horizontal boring machines and 5-axis machining centers shows that inadequate or contaminated grease/oil is the single most common non-electrical cause of chatter on a single axis. Verify lubrication before opening any axis housing.
6.1 Verification procedure
- Pull the lubrication system diagnostics page on the HMI (Setup / Lubrication).
- Confirm the interval timer (typical 50-200 m of axis travel) is appropriate for the actual duty cycle.
- Confirm the dispensed volume per pulse matches the manufacturer specification; a clogged distributor will starve one axis while leaving the panel reading normal.
- Sample the lubricant and inspect for metal swarf (magnetic plug) and color change.
- Run the machine unloaded for 30 minutes and observe each lubrication point visually — every linear-guide block and ballscrew nut must show a fresh grease ridge.
7. Mechanical Inspection: Ballscrew and Bearing Health
Once lubrication is confirmed, measure mechanical condition. A ballscrew with lost preload or pitted nut flanks produces a regular chatter that aligns with screw rotation, not with feed direction.
| Check | Method | Acceptance |
|---|---|---|
| Ballscrew backlash | Dial indicator at the axis table, reverse direction at mid-stroke | < 0.02 mm for finishing, < 0.05 mm for roughing |
| Preload drag torque | Measure motor current at 1 mm/s in unloaded axis | Within ±15 % of commissioning baseline |
| Nut temperature | Pyrometer after 30 min unloaded axis travel | < +10 °C above ambient |
| Bearing preload (rotary) | Measure stick-slip at slow jog reversal | No audible clicking; smooth reversal |
| Linear guide rollers | Hand-feel carriage after lubrication cycle | No grittiness |
If backlash on any of the five axes exceeds the table values, schedule a ballscrew preload adjustment or replacement. On machines with cam-grinding or precision 5-axis milling duties, replacing the ballscrew assembly on the suspect axis is often faster and more cost-effective than a full rework of the nut.
8. Drive and Control Loop Optimization
If chatter persists after lubrication and ballscrew checks, the drive control loop on the affected axis is suspect. The original procedure reported that the B/C axis drive card had already been swapped — confirm the new card is paired with the correct firmware and that the axis commissioning data was transferred.
8.1 Drive tuning sequence
- Open the SINUMERIK
Commissioning / Drivesmenu and verify the drive object is recognized with the correct motor and encoder type. - Run the automatic commissioning routine (
Drive optimization) for the suspect axis only. - Compare the resulting Kv factor (MD32200 $MA_POSCTRL_GAIN) against the value on the working machine. If the local machine reads 30 % higher or lower, override it to the reference value.
- Re-run the circularity test (G1 circle in JOG with feed-rate override 100 %) and measure the roundness deviation with a dial indicator.
8.2 Jerk filter for closed-loop dynamics
The jerk filter time constant MD32410 affects the command feedforward. A small value (fast filter) places high demand on the speed loop bandwidth; a value larger than 0.030 s often allows a lightly-tuned drive to track the corner without lag, at the cost of slightly increased cycle time. Pick the smallest value that produces a clean corner trace.
9. Tool Path and CAM-Side Considerations
Once the controller side is clean, look at the part program. SINUMERIK handles small block sequences poorly because the position controller executes each block independently.
| CAM-side issue | NC-side mitigation |
|---|---|
| Many short linear blocks at the corner | Activate COMPCAD / COMPCURV (MD20480) |
| Sharp tangential transitions | Add G645 or G642 smoothing |
| Toleranced polygon corners | Use SOFT (G64 with look-ahead) and increase look-ahead frames (MD20170) |
| Excessive tool-vector rotation at a small linear path | Insert a small flat between two arcs to give the rotary axes time |
10. Diagnostic Procedure (Step-by-Step)
- Capture machine backup and save the current MD archive.
- Run the same NC program on a known-good machine and compare the surface finish. Document the result.
- Inspect the chatter wavelength and orientation; map to the table in Section 2.
- Verify lubrication: HMI diagnostics + physical flow check at each lubrication point, including rotary unions.
- Measure ballscrew backlash and nut temperature on all five axes.
- Inspect MD32400, MD32402, MD32410 on all five axes; harmonize jerk filter times.
- Inspect MD20480, MD20170; consider enabling compression (COMPCAD) if the program is dense.
- Validate ORIVECT / OSCTRL orientation smoothing settings for the active program.
- Run drive auto-commissioning on the suspect axis; harmonize Kv factor with reference machine.
- Re-machine the test part; document surface finish (Ra, visual).
11. Verification Checklist
| Item | Acceptance criterion |
|---|---|
| Visual chatter | None visible at corner under 10× magnification |
| Surface roughness Ra | Within specification for the part |
| Roundness on circularity test | < 0.01 mm deviation |
| Axis temperatures after 30 min | < +10 °C above ambient for linear axes; < +15 °C for rotary |
| Lubrication intervals | Interval timer and dispensed volume logged |
| MD archive | Saved with date and machine serial number |
| Drive tuning log | Kv and filter times recorded for all axes |
12. FAQ
Which MD has the largest impact on chatter at 5-axis corners on a SINUMERIK 840D?
MD32410 $MA_AX_JERK_TIME for the rotary axes (B and C). Raising it from 0.010 s to 0.020-0.030 s softens the rotary axis demand at corner transitions and removes the dominant excitation. Always pair this with MD32400 = 1 (jerk limiting enabled) and MD32430 = 1 (jerk limit active during compression).
How do I tell whether chatter is from the NC program or from the machine mechanics?
Run the identical NC program on a second, known-good 840D machine. If chatter follows the program (wavelength identical on both machines), it is tool-path / CAM / NC. If chatter appears only on the suspect machine, it is mechanical or control-side on that machine.
What lubrication interval is typical for a 5-axis machine tool axis?
For linear guides and ballscrews under intermittent cutting load, the typical range is 50-200 m of axis travel between lubrication pulses, with a dispensed volume per pulse of 0.05-0.15 cm³. Rotary axes are usually grease-lubricated every 500-2000 hours of operation. Always confirm against the machine builder's lubrication chart; do not extrapolate.
Can a swapped drive card on a B/C axis cause chatter?
Yes, if the replacement card is not loaded with the matching firmware or if the commissioning data was not transferred from the old card. The Kv factor (MD32200) and the jerk filter time (MD32410) on the drive side may revert to defaults, producing a different closed-loop response. Re-run automatic drive optimization and harmonize Kv with a reference axis.
What is the difference between G642 and G645 for chatter reduction?
G642 maintains the tolerance band parallel to the programmed contour (best for finishing surfaces). G645 uses an axis-specific tolerance and can run faster but introduces small orientational deviations. For 5-axis finishing passes where chatter is concentrated at corners, prefer G642 and combine it with G64 (look-ahead) for the smoothest result.