Siemens 840D Axis Motor Humming: Kv Factor and MD1407 Tuning

David Krause20 min read
Motion ControlSiemensTroubleshooting
Licensed PE Working through this on a live machine? A Maine-licensed engineer can take it from here — included with IMD hardware, by the hour for everything else. Book an engineer

Siemens 840D Axis Motor Humming: Kv Factor and MD1407 Tuning

A 2006-vintage SINUMERIK 840D (Powerline) with SIMODRIVE 611D drives developed a loud, sustained audible hum on a Nikken 4th/5th-axis rotary immediately after the backlash compensation was tightened. The motor had not been mechanically binding prior to the change, and the hum persisted even after the backlash value was relaxed. This reference documents the full root-cause map for that class of fault, the alarm codes it produces (21612, 25201, 300608), and the field-proven tuning sequence used to silence the drive without losing follow-on accuracy. The procedures apply to SINUMERIK 840D Powerline with 611D, 840D sl with 611D, and 840D sl with SINAMICS S120, with callouts where drive-MD semantics differ.

Safety: Before opening any drive cabinet or disconnecting motor leads, lock the main disconnect, wait five minutes for the DC bus to bleed below 60 VDC, and verify with a meter. Drive MD changes on a live machine must only be performed by a qualified commissioning engineer with E-stop within reach. Drive MDs that are not safety-locked become effective immediately after OK on the HMI, but the drive performs an implicit re-initialization that can drop torque for ~200 ms.

1. Symptom Classification

840D operators and service engineers consistently mis-label the noise class. The acoustic signature points directly at the loop that is mis-tuned, so classify the noise first.

Sound Frequency Range Most Likely Source First MD to Touch
High-pitched whistle 800 Hz to 4 kHz Speed controller P-gain too high; current-loop d-q decoupling issue MD1407 (611D) / p1460 (S120)
Low-frequency hum (~50/60 Hz) Mains-related PWM carrier interaction with mechanical resonance; loose lamination stack MD1407 + p1800 carrier frequency
Audible growl that varies with speed Proportional to RPM Mechanical: bearing, gearbox, coupling, brake drag Mechanical intervention
Squeal that disappears at standstill Velocity-dependent Encoder cable damage or shield break Encoder wiring
Loud hum that stops with axis enable removed Any Purely control-loop; mechanical is exonerated MD32200 + MD1407

If the hum is unchanged when the controller enable is removed via PLC (DB38.DBX2.1 = 0), the source is mechanical or electrical (cable, brake). If the hum dies within 200 ms of enable removal, the source is in the drive control loops.

2. System Identification and Drive Type

MD numbering and effective drive MD scope differ between the three relevant combinations. Confirm what is in the cabinet before you start changing values.

Control Drive Position Loop MD Speed P-gain MD Current P-gain MD Notes
840D Powerline SIMODRIVE 611D MD32200 $MA_POSCTRL_GAIN MD1407 $MD_SPEEDCTRL_GAIN_V MD1190 Drive MDs accessed under StartUp → Machine Data → Drive MD
840D sl SIMODRIVE 611D MD32200 MD1407 MD1190 Same numbering as Powerline
840D sl SINAMICS S120 MD32200 p1460[0] p1700[0] Drive parameters live on the SINAMICS CU; editing in the NCK has no effect on S120

For the 2006 vintage described, expect SIMODRIVE 611D modules (6SN1123-xAx0x-xHAx series for 1FE1 / 1FK6 motors; 6SN1145 / 6SN1146 for spindles). The 1FK6 feed motor with built-in holding brake is the most common cause of a post-backlash hum because the brake pad spring-pre-load is sensitive to axial shaft loading.

3. Alarm Code Reference

Three alarms typically appear in the alarm log when a hum is accompanied by instability. Each one points to a different layer of the enable chain and must be cleared in order; do not skip ahead to the drive parameter while a high-priority alarm is still pending.

Alarm Full Text Layer Clearance Number Corrective Action
21612 Channel %1 Axis %2 VDI signal 'Servo enable' reset during motion NC-PLC interface RESET Investigate why the PLC dropped pulse enable during travel. Check DB38.DBX2.1, ESTOP chain, and the brake-release acknowledgement.
25201 Axis %1 Drive %2 fault Drive hardware POWER ON / RESET on drive Drive module reported an IGBT, link, or encoder fault. Read r0947/r0948 on SINAMICS, or read drive MD 0900 / 0901 (611D) for last fault code.
300608 Drive 8: Speed controller output at limit Drive control loop RESET (clears when speed error returns inside window) Speed controller saturated. Reduce mechanical load, increase MD1408 (TN), or reduce MD1407 (Kp). Verify encoder feedback integrity first.
Alarms 300000–300999 are drive-specific alarm numbers assigned per drive slot. "Drive 8" in alarm 300608 refers to drive address 8, not axis 8. The drive address is shown on the HMI under Diagnosis → Drives. If you only have 4 axes, drive 8 is likely a spindle on a second NCU or an unused slot.

4. Enable Chain and Brake Verification (DB38.DBX2.1)

The 840D uses the DB31–DB38 (8 axes per NCU) block as the axis interface to the PLC. Each axis has 32 bytes of process data. The signals relevant to a hum that started immediately after a mechanical change are concentrated in the first six bytes.

Signal Address (Axis 8) Type Function
Pulse enable / Controller enable DB38.DBX2.1 Bool, from PLC → NC Enables torque-producing current. Drops on E-stop, follow-up error, or PLC logic.
Drive enable DB38.DBX21.7 Bool, from NC → drive Internal NC-side enable sent to the drive module.
Axis enable DB38.DBX0.6 Bool, PLC User-level axis enable from PLC application.
Position measuring system 1/2 DB38.DBX1.5 / DBX1.6 Bool Which encoder is active. 1 = motor encoder, 2 = direct measuring system.
Brake released feedback DB38.DBX62.5 (varies by version) Bool, drive → PLC Brake pad open. If this reads 0 in operation, the holding brake is dragging.

Procedure:

  1. Open the PLC online (Step 7) and force a VAT view on DB38.
  2. Apply axis enable, jog command, and confirm DB38.DBX2.1 goes to 1 within 100 ms.
  3. Read the brake feedback bit. On a 1FK6, this is DBX62.5 or the equivalent mapped to the drive telegram.
  4. If the brake is dragging, the motor current will rise 8–15% at standstill and the audible hum is the PWM struggling against the brake torque. Releasing or adjusting the brake shim is the mechanical fix; do not chase this with a Kv change.

5. Mechanical Root Causes

Mechanical binding produces a hum because the speed controller fights a constant disturbance torque. The Kv increase you try to compensate actually amplifies the audible result.

  1. Backlash over-tightened. The Nikken CNB and CBN rotaries use a worm-and-wheel or planetary reduction. A backlash value below the published minimum (typically 0.02 mm or 3 arc-min) pre-loads the gear train. Pull the backlash back to the manufacturer value and re-test.
  2. Coupling alignment. Concentricity and angularity tolerances for a typical 1FK6-to-Nikken input coupling are 0.05 mm. A lateral offset of 0.1 mm produces a sinusoidal speed disturbance at 1× mechanical revolution that the speed controller amplifies into a hum.
  3. Bearing pre-load. After backlash adjustment, the rotary table bearings (e.g., Nikken cross-roller) can be pre-loaded off their factory setting. Re-shim to published preload.
  4. Holding brake drag. Common failure mode on 1FK6 and 1PH8 motors after 15+ years of service. The armature plate is a clearance fit; a worn plate drags the rotor.
Decoupling test (most reliable mechanical check): Loosen the motor from the rotary, run the motor unloaded, and listen. If the hum persists, mechanical is exonerated. If the hum disappears, the mechanical path is the cause and no amount of drive parameter tuning will silence it.

6. Electrical Root Causes (Cable, Winding, Encoder)

If the decoupling test produces a clean motor, move to electrical isolation. A damaged cable or winding imbalance drives an asymmetric current that the controller cannot fully reject.

6.1 Power Cable Resistance Test

With the motor leads disconnected from the drive (always confirm lockout), measure resistance phase-to-phase at the drive-end of the cable:

Test Acceptance Action if Failed
U−V resistance 0.3–2.0 Ω for 1FK6 class (depends on motor); ±5% balance required Inspect connector pins for oxidation or back-out
V−W resistance Same as U−V within ±5% Replace cable if balance > 10%
W−U resistance Same as above Replace cable
U, V, W to PE > 10 MΩ at 500 V Replace cable; check for coolant ingress
Insulation resistance (phase-PE) > 100 MΩ preferred Replace cable

6.2 Winding Resistance

Measured at the motor junction box (cable disconnected from motor):

  • Phase-to-phase 0.3–1.5 Ω typical for 1FK6 family; balance ±3%
  • Phase-to-ground > 100 MΩ
  • If balance exceeds 5%, suspect a turn-to-turn short; replace motor

6.3 Encoder Cable

Common failure mode on rotaries that have been moved. Inspect for pinch points at the slip ring or rotary union. Shield continuity must be one-piece from encoder to drive end; a shield break produces velocity-dependent noise that masquerades as a hum.

7. Drive Parameter Root Causes (MD32200, MD1407)

If mechanical and electrical are both clean, the hum is in the drive control loops. Two MDs are responsible for almost every audibility issue on a healthy motor-cable pair.

7.1 MD32200 $MA_POSCTRL_GAIN (Kv)

The Kv factor is the proportional gain of the position controller, in units of (feed/min) per mm (or (rev/min) per mm for a rotary). The default of 2.7 is typical for a small rotary; it is not a universal value. Increasing Kv does not always make the hum louder. A Kv that is too high for a high-inertia rotary (Nikken CNB-160 weighs 60 kg) will produce a low-frequency oscillation that the drive tries to cancel, resulting in a noticeable hum. Lowering Kv below the system natural frequency can silence it but increases following error at corners.

Effective Kv range for a 60 kg rotary: 0.5 to 4.0. Start with 1.0, raise in 0.5 increments while monitoring following error.

7.2 MD1407 $MD_SPEEDCTRL_GAIN_V (Speed Controller P-Gain)

On 611D, this is the P-gain of the speed controller. Units are Nms/rad (Nm·s/rad) on the torque/velocity plane. On SINAMICS S120 the equivalent is p1460[0]. The 611D default is determined by the motor type at commissioning; on a 1FK6, values between 0.5 and 3.0 are typical.

MD1407 that is too high produces the whistle; MD1407 that is too low produces a low-frequency moan at low speeds. The audible signature of the original machine was a whistle, which is the more common 611D failure mode and a strong indicator that MD1407 needs to come down.

7.3 Speed Controller Integral Time MD1408

The integral component of the speed controller (MD1408) should be roughly 10–20× the value of MD1407 in ms for a healthy motor. If MD1407 is reduced without checking MD1408, the integral term can become dominant and produce a low-frequency drift hum. The relationship is:

MD1408 [ms] ≈ 10 × MD1407 [Nm·s/rad] (rough field starting point; refine by following-error plot)

8. Position Loop Gain (MD32200) Tuning Procedure

  1. Power on the machine, enter user area, navigate to StartUp → Machine Data → Axis MD. Filter to MD32200 for the affected axis.
  2. Note the current value (typical 2.7 on a Nikken small rotary).
  3. Lower the value by 30% (e.g., 2.7 → 1.9). The HMI prompts whether the change is effective immediately or after POWER ON; on 611D, MD32200 is generally effective immediately on the affected axis once the drive is re-initialized, but the standard commissioning practice is to apply at next POWER ON.
  4. Power on the machine, jog the axis at 50% rapid, monitor the HMI Service → Axis → Following error trace. A healthy Kv produces a following error proportional to velocity (the slope equals 1/Kv). If the trace shows oscillation at a fixed period, Kv is too high.
  5. Iterate in 0.3 increments. The target is no audible noise, following error below the value set in MD36400 $MA_CONTOUR_TOL, and no fault within 30 minutes of part-program run.
  6. Save the machine data to the CF card (StartUp → Series Commissioning → Save).
Effective-immediately behavior: MDs flagged with the attribute effective immediately take effect on the next Reset or on a soft Reset. On 840D, this is a stop-and-restart of the axis, not a full NCK restart. Drive-side MDs (MD1407 and others) require a drive warm restart, which the HMI performs automatically when the data is saved with effective immediately set, but the axis must be in disabled state.

9. Speed Controller (MD1407) Tuning Procedure

The 1.850 → 1.4 reduction documented in the source fault is in the right direction but overshoots the safe range. Use the following stepped procedure to avoid the "lower MD1407 too far and start faulting" trap:

  1. Open the drive MD view: StartUp → Machine Data → Drive MD. Filter to MD1407 for the affected drive (drive 8 in the alarm log, but for the rotary axis you are working on, the drive address is the axis's drive address; do not assume the drive number is the same as the axis number).
  2. Record the current value, MD1407 = 1.850.
  3. Reduce by 10% only: MD1407 = 1.665. Save. Wait for the drive to re-initialize. Reset the controller enable.
  4. Run the rotary axis in JOG at 50% rapid, 30° swing. Listen for the whistle. Check for any 300608 alarm in the alarm log.
  5. If still whistling, drop another 10%: 1.50. Repeat.
  6. Continue in 0.15 steps until the whistle disappears or the first 300608 alarm appears. The sweet spot is the highest MD1407 value that does not produce audible noise.
  7. Once MD1407 is fixed, adjust MD1408 (TN) to maintain the 10–20× ratio described in Section 7.3.
Why overshooting to 1.4 caused faults: At MD1407 = 1.4 on a Nikken CNB-160, the speed controller cannot reject the load disturbance from the worm gear. The integral term ramps up to saturate the current command, which triggers 300608 (speed control at limit) when the measured speed cannot reach setpoint. The fix is to keep MD1407 high enough to reject the disturbance and rely on the position loop (MD32200) to control the audibility. In practice, MD1407 should be set first, then MD32200.

9.1 Verification by Following-Error Trace

Use the HMI integrated trace to record following error during a JOG move. Set the trigger threshold to 2× the steady-state following error. A healthy Kv/MD1407 combination shows:

  • Steady-state following error: Fss = V / Kv (linear axis); or Fss = V / Kv (rotary, where V is in deg/s and F is in deg).
  • Acceleration following error: Facc = J / Kv where J is axis jerk, dominated by mechanical inertia reflected back through the gear.
  • Oscillation period in the trace should be < 50 ms. A 200–400 ms period oscillation indicates mechanical resonance not yet addressed.

10. Auto-Optimization on Limited-Travel Axes

The rotary in question has ~60° of total swing. The standard 840D auto-optimization routine (StartUp → Optimize Axis) requires a configured travel range large enough to reach a stable speed. For a 60° swing, the routine will refuse to start or will produce garbage if the axis is commanded beyond its software limit. Two practical workarounds:

10.1 Manual Circularity Test

The HMI circularity test (under Commissioning → Optimization → Circularity) requires two linear axes to be moved in a circle. Use the rotary as the index axis and a single linear as the radial. With the backlash and Kv dialed in, the circle should round-trip within the configured position window (MD32250 $MA_RATED_AMP).

10.2 Step-and-Listen Optimization

  1. Set MD1407 to the current value, write a 5-line part program that moves the axis +20°, settles 2 s, −20°, settles 2 s, M30.
  2. Run in single-block, listen for the whistle in the dwell period.
  3. Change MD1407, repeat. This is essentially what the 840D auto-optimization does internally, but with much finer control over each step.

10.3 What the Auto-Optimization Will Travel

The 840D optimization sequence uses the following logic for travel bounds:

Ttravel = min(SW_limit_high − SW_limit_low − safety, configured_optimize_range)

Where configured_optimize_range is from the HMI prompt. For a 60° swing, enter the value 50° in the prompt; the drive will run multiple moves within that 50° span. This is sufficient to estimate the moment of inertia and tune the speed controller.

11. Motor and Cable Field Testing

Use a digital multimeter with a low-Ohms scale (resolution 0.01 Ω) and a 500 V insulation tester (megger). De-energize the drive, lock the main disconnect, and verify absence of voltage at the drive output terminals.

Measurement Tool Expected Failure Mode Indicated
U−V, V−W, W−U at drive end (cable + motor) DMM, Ω mode 0.3–2.0 Ω (motor dependent) ±5% balance Imbalance > 10% = cable damage or turn short
U, V, W to PE (drive end) Megger at 500 V > 10 MΩ (preferably > 100 MΩ) Coolant ingress, insulation damage
Encoder +5 V to GND DMM 4.75–5.25 V with drive enabled Encoder supply cable short
Encoder clock/data at drive connector (oscilloscope) Scope, 200 MHz RS485 differential, 1 MHz baud, clean edges Ringing, missing pulses
Holding brake resistance (24 V brake) DMM 20–50 Ω typical (motor dependent) Open = brake coil open; low = shorted turns

12. Parameter Backup, Restore, and Change Activation

Before any optimization, take a series-commissioning backup to a CF card. The 840D stores machine data in three logical areas: NCK MD, PLC, and drives. A full backup covers all three.

StartUp → Series Commissioning → Create Series Start-up File → select all three areas → save to CF card

To restore a single MD after a failed optimization:

  1. Edit MD on the HMI: StartUp → Machine Data, find the MD, set to known good value.
  2. Press Save; the HMI prompts whether to write to CF card. Confirm.
  3. Reset the axis. Verify no alarm.

To restore a full backup:

  1. Boot the NCU in service mode (CNC-STARTUP switch on the front of the NCU box, position 1).
  2. Insert the CF card with the series start-up file.
  3. Select StartUp → Series Commissioning → Load and follow the prompts.
  4. Normal boot. The drive may re-initialize twice.
Important: Drive MD changes on a 611D are effective only after a drive restart. The 840D HMI will report the change is active, but the drive continues using the old value until either: (a) a soft axis reset is performed by toggling DB38.DBX2.1, or (b) a full drive warm restart is triggered. NCK MDs are effective at the next Reset.

13. Safety and High-Energy Considerations

Drive MD tuning is not free of risk. The current at standstill on a stalled 1FK6 can exceed rated within 200 ms if the controller fights a mis-wired encoder.

  1. Confirm the emergency-stop circuit is functional before applying the first command.
  2. Keep one hand on the E-stop button during the first move after any change.
  3. Reduce the rapid override (MD32010 $MA_JOG_VELO_RAPID or feedrate override 0%) for the first verification move.
  4. Set MD36610 $MA_AX_EMERGENCY_STOP_TIME to a small value (e.g., 200 ms) to ensure the drive de-energizes quickly on a fault.
  5. Monitor drive fault LED (red on 611D modules) throughout. Any trip means back off the previous change by 30%.

14. Field-Proven Sequence (End-to-End)

The following 12-step sequence is the recommended field workflow for a 2006-class 840D that developed an audible hum after backlash adjustment. Total time: 60 to 90 minutes including verification.

  1. Capture the alarm log. Note 21612, 25201, 300608 status and timestamps.
  2. Verify the mechanical change. Re-read the backlash value; if < 0.02 mm on a Nikken, restore to 0.03 mm.
  3. Decouple the motor from the rotary. Run motor unloaded, listen. If hum persists, go to step 5; if not, mechanical is binding, fix and re-test.
  4. Re-couple. Verify coupling concentricity < 0.05 mm with a dial indicator.
  5. Megger the power cable and the motor windings. Replace cable if insulation < 10 MΩ or resistance imbalance > 5%.
  6. Inspect the encoder cable at the rotary's slip ring. Replace if any shield break is visible.
  7. Confirm the holding brake is releasing. Read DB38.DBX62.5 in the PLC. If 0, the brake is dragging; service the brake.
  8. Set MD1407 to 1.65 (10% below the original 1.85). Save. Reset axis. Jog in 30° moves. Listen for whistle.
  9. If still whistling, reduce MD1407 by another 0.15 (to 1.50). Repeat. Stop at the first sign of 300608 alarm.
  10. Adjust MD1408 to maintain ~10× the MD1407 value in ms (e.g., MD1407 = 1.5 → MD1408 ≈ 15 ms).
  11. Reduce MD32200 (Kv) by 30% from original (e.g., 2.7 → 1.9). Verify following error stays below MD36400 setting.
  12. Run a part program for 30 minutes. Confirm no alarms and no audible noise.

15. Speed/Position Control Loop (Block Diagram)

SINUMERIK 840D / 611D Cascaded Control Loop Position Setpoint Pos ControllerMD32200 Kv Speed ControllerMD1407 / MD1408 Current ControllerMD1190 / MD1191 1FK6 / 1FE1 Motor Encoder feedback (motor / direct measuring system) Tuning direction if humming: 1) Lower MD1407 (speed P-gain) in 10% steps to silence whistle 2) Adjust MD1408 to maintain 10× ratio 3) If low-frequency hum persists, lower MD32200 (Kv) in 0.3 steps

16. Verification Checklist

Check Method Pass Criterion
Alarms cleared HMI alarm display No 21612, 25201, 300608 in 30 min
No audible noise at standstill Ear, 1 m from motor < 40 dB ambient
Following error in tolerance HMI trace Fss = V / Kv ±20%
Jog move at 50% rapid Manual jog Smooth, no stall
30-min part program Run a known good program No alarms, no audibility change
Backlash test Standard test Within MD36500 ENC_COMP_TOL
Position window MD32300 MA_LAG_TOL, MD36400 No following-error alarm in trace

17. Reference: Common 611D Drive MDs

MD Name Typical Default Effect When Too High Effect When Too Low
MD1190 $MD_CURRENTCTRL_GAIN_P 0–2.0 (motor specific) Current loop oscillation Sluggish response
MD1191 $MD_CURRENTCTRL_GAIN_I 0–2.0 Integral windup Steady-state current error
MD1407 $MD_SPEEDCTRL_GAIN_V 0.5–3.0 Whistle Disturbance not rejected (hum)
MD1408 $MD_SPEEDCTRL_INTEGRATOR_TIME 5–50 ms Overshoot, oscillation Steady-state speed error
MD1409 $MD_SPEEDCTRL_REF_MODEL_FREQ 0–200 Hz Stiffness loss Reference model disabled
MD1410 $MD_SPEEDCTRL_REF_MODEL_DAMPING 0.5–1.0 Resonance amplification Underdamped response
MD1500 $MD_SPEED_FILTER_TIME 0–10 ms Phase lag No filtering

FAQ

Why did the 840D axis motor start humming immediately after a backlash adjustment?

The backlash reduction pre-loaded the gear train of the rotary (e.g., Nikken CNB/CBN) and/or shifted the motor shaft's axial position, causing either mechanical binding that the speed controller tries to reject, or dragging of the holding brake. The hum typically disappears when the motor is decoupled from the load, which confirms the mechanical path. A safe first step is to restore backlash to the manufacturer value and re-test before changing any drive MD.

Does changing MD32200 POSCTRL_GAIN take effect immediately or do I need to restart?

MD32200 is an NCK axis MD; on the 840D it takes effect on the next axis reset. Drive MDs such as MD1407 take effect on a drive warm restart triggered by saving the value via the HMI with effective immediately set; the axis must be in disabled state during the restart. A full NCK restart is not required but is the safest option if the change is significant.

How do I silence a 840D drive whistle without triggering alarm 300608?

Reduce MD1407 (speed P-gain) in 10% steps (e.g., 1.85 → 1.665 → 1.50) and stop at the highest value that produces no whistle. Then verify MD1408 (speed controller integral time) is roughly 10× the new MD1407 in ms. If you reduce MD1407 below the level required to reject load disturbance, the controller saturates and produces 300608 instead of the whistle. This is the failure mode the original poster hit at 1.4.

Can I run the 840D auto-optimization on a rotary that only swings 60°?

Yes, with a 50° optimization range entered in the HMI prompt. The routine will run multiple moves within that 50° span. The optimization may not converge to a single set of values if the moment of inertia varies significantly across the swing (e.g., with a heavily asymmetric workpiece). In that case, manual step-and-listen tuning (Section 10.2) is more reliable than the auto sequence.

What does alarm 300608 Drive 8 Speed Control at Limit mean, and why does it keep returning?

Alarm 300608 is a drive-specific alarm (slot 8 in the alarm range 300000–300999) indicating the speed controller output is saturated. It returns as long as the speed error (setpoint − actual) cannot be reduced. The most common causes are excessive load, low MD1407, or encoder feedback loss. Clear by addressing the underlying speed-loop saturation, not by resetting the alarm.

How do I test the motor and cable for the hum source?

With the drive de-energized and locked out, measure phase-to-phase resistance (target 0.3–2.0 Ω with ±5% balance) and phase-to-ground insulation at 500 V (target > 10 MΩ). Imbalance above 10% or insulation below 10 MΩ means replace the cable; if the motor at its junction box fails the same tests, the motor itself is suspect. Verify the encoder cable's shield is continuous and the holding brake is releasing (DB38.DBX62.5 reads 1 when active).

Back to blog