Troubleshooting Siemens MM440 Input Current Unbalance on a 110kW Fan Drive
1. Problem Description
A common field report on the Siemens MICROMASTER 440 (MM440) is that the three-phase input current drawn by the drive appears unbalanced on the upstream protection/monitoring device, even though the motor itself is mechanically and electrically healthy. A typical scenario is:
- Drive rating: 110 kW, 3-phase, 400 V class.
- Load: fan motor (quadratic torque, no reverse, no rapid acceleration).
- Direct-on-line (DOL) test: input current balanced, protection relay stays reset.
- VFD run: line currents reported as unbalanced by 10-20 A between phases, with apparent swings of -10 A to +10 A from the average.
- Result: upstream phase-failure relay, motor protection relay, or phase-sequence monitor trips or alarms even though the drive itself is healthy and shows no fault.
This is one of the most reported MM440 measurement artifacts on drives above 30 kW, where the harmonic content of the input current is high enough to confuse inexpensive current transformers and average-responding ammeters.
2. Root Cause: Non-Linear Input Current from the 6-Pulse Rectifier
The MM440 (like nearly every low-voltage general-purpose drive) rectifies the 3-phase AC supply into a DC bus using a 6-pulse diode bridge. The DC bus charges large electrolytic capacitors; the inverter then chops this DC into a variable-frequency PWM output to the motor. The input side of the bridge therefore does not draw sinusoidal current. Each diode pair only begins to conduct when the instantaneous line-to-line voltage exceeds the DC bus voltage plus two diode drops:
V_conduction(t) = V_LL·sin(ωt) ≥ V_DC + 2·V_F
Below that threshold, the diodes are reverse-biased and the line current is zero. Above the threshold, current rushes into the DC bus capacitors in a narrow conduction pulse. The result is the classic "double-hump" or "pulsed" current waveform at the AC input with a high crest factor:
Crest Factor (CF) = I_peak / I_RMS
| Load Type | Typical Crest Factor | Typical THDi |
|---|---|---|
| Direct-on-line motor (sinusoidal) | 1.41 | < 5% |
| 6-pulse VFD, no line reactor | 2.0 - 2.5 | 35 - 80% |
| 6-pulse VFD with 2% line reactor | 1.7 - 1.9 | 25 - 45% |
| 6-pulse VFD with 5% line reactor | 1.5 - 1.7 | 15 - 30% |
| 12-pulse rectifier + phase-shift transformer | 1.5 - 1.6 | 5 - 8% |
| Active Front End (AFE / Low Harmonic drive) | 1.4 - 1.5 | 3 - 5% |
The dominant harmonic orders drawn by a 6-pulse rectifier follow the characteristic equation h = 6k ± 1:
| Order h | Frequency at 50 Hz | Typical Magnitude (no reactor) | Typical Magnitude (2% reactor) |
|---|---|---|---|
| 5 | 250 Hz | 50 - 65% of I₁ | 30 - 40% of I₁ |
| 7 | 350 Hz | 30 - 45% of I₁ | 15 - 25% of I₁ |
| 11 | 550 Hz | 8 - 12% of I₁ | 5 - 9% of I₁ |
| 13 | 650 Hz | 5 - 9% of I₁ | 3 - 7% of I₁ |
| 17, 19, 23, 25 | 850 - 1250 Hz | 2 - 6% of I₁ | 1 - 4% of I₁ |
Each phase of the line current is rich in these harmonics, and the instantaneous peak-to-RMS relationship is highly distorted. This is the physical cause of the apparent imbalance.
3. Why Direct-On-Line Shows Balanced Current
A DOL induction motor presents an essentially linear impedance (L-R series circuit) to the supply. The stator current is the supply voltage divided by that impedance, producing a near-sinusoidal current waveform with crest factor 1.41 (the textbook value for a pure sine wave). Standard CTs, protection relays, and averaging ammeters were designed for this case and read it accurately.
The VFD input current is fundamentally different. The phase currents are not sinusoidal; they are 5th-, 7th-, 11th-, 13th-... harmonic-rich pulsed currents. Any instrument that assumes a sine wave will read the wrong RMS value, and the reading will vary as the harmonic phase relationships shift with motor load, bus voltage, and supply impedance. Hence the apparent -10 A / +10 A swing between phases is a measurement artifact, not a real current imbalance.
4. Why Standard Measurement Instruments Fail
Most low-cost clamp meters and many panel-mount ammeters are average-responding, RMS-calibrated. They rectify the AC signal, average it, scale by 1.11 (the form factor of a sine wave), and display "RMS". On a non-sinusoidal waveform, the form factor is no longer 1.11, so the displayed value is wrong by 10-40%.
| Meter Type | Method | Accuracy on VFD Input |
|---|---|---|
| Analog moving-iron | True RMS by thermal/heating | ±3-5% (good) |
| Cheap digital clamp (average-responding) | Rectify + average + scale 1.11 | ±20-50% (UNRELIABLE) |
| True-RMS digital clamp (≥ 1 kHz bandwidth) | Σ(i²·Δt) integration | ±2-3% (acceptable) |
| Hall-effect current probe + oscilloscope | Direct waveform capture | Reference / best |
| Three-phase power analyzer (e.g., Fluke 435, Hioki PW3198) | Wideband True-RMS + harmonic spectrum | Reference / best |
For a 110 kW MM440, expect a fundamental input current around 200 A at full load (depends on motor efficiency and power factor). The harmonic component can add 30-80% extra peak current, so the peak line current pulses can reach 350-450 A even though the true RMS is only 200-220 A.
5. Diagnostic Procedure
Use this step-by-step procedure to confirm that the unbalance is an instrument artifact rather than a real fault.
- Lock out the motor coupling if practical. Run the MM440 into a no-load or light-load condition so the DC bus reaches steady state.
-
Measure with two different meter types on the same phase:
- A standard clamp meter (the one currently alarming the protection relay).
- A True-RMS clamp meter rated to at least 1 kHz bandwidth (Fluke 87V, Fluke 376 FC, Kyoritsu KEW 6305, Hioki CM4373, etc.).
- Capture the current waveform on at least two phases using a Hall-effect current probe (e.g., Fluke i2000 Flex, Chauvin Arnoux PAC series, Tektronix TCP303) connected to a scope with a current-versus-time trace.
- Compute the crest factor: CF = I_peak / I_RMS. If CF is in the 2.0-2.5 range, the input current is non-sinusoidal as expected for a 6-pulse rectifier without a reactor.
-
Run a three-phase power quality analyzer on the line side of the drive (Fluke 435-II, Hioki PW3198, Dranetz HDPQ, Schneider ION7650, Siemens Sentron PAC4200). Confirm:
- Voltage THD < 5% (utility is clean).
- Current THD between 30% and 80% (typical for 6-pulse, no reactor).
- Fundamental current imbalance < 5% between phases.
- Read the MM440 internal measurements (r0034, r0036, r0069, r0070) - these are computed inside the DSP and represent the actual fundamental component at the line side. They should be balanced within a few amps.
- Verify F0020 is not active in the fault buffer (r0947, r0948, r0949). A real phase-loss fault will set F0020 and latch the drive.
6. Solution 1: Use True-RMS Measurement
The lowest-cost, fastest field fix. Replace the existing protection relay or meter with a True-RMS device whose bandwidth covers at least the 50th harmonic (2.5 kHz at 50 Hz supply). Suitable families:
- Schneider Electric TeSys T motor management relays with True-RMS CTs.
- ABB CM-UFD.MxN or REF615 with harmonic-aware current measurement.
- Siemens SIRIUS 3RR2 current monitoring relays with selectable True-RMS mode.
- Multilin 869, SEL 749M, Woodward MFR300/400 - all with harmonic-tolerant current inputs.
7. Solution 2: Install a Line Reactor
A line reactor (also called line choke or AC reactor) is a three-phase iron-core inductor installed between the upstream breaker and the MM440 line terminals. The standard sizing rule for a 6-pulse drive is:
Z (%) = (X_L / V_phase / I_rated) × 100%
For 110 kW MM440 typical recommendations:
| Reactor Impedance | Effect on THDi | Effect on Crest Factor | Effect on V_DC bus | Use Case |
|---|---|---|---|---|
| 2% (minimum recommended) | Reduce ~25-40% | 1.7 - 1.9 | Drop ~3-4% | General harmonic mitigation |
| 3% (typical industrial) | Reduce ~35-50% | 1.6 - 1.8 | Drop ~4-5% | IEEE 519 borderline compliance |
| 5% (heavy harmonic area) | Reduce ~50-65% | 1.5 - 1.7 | Drop ~6-8% | Weak supply / diesel-gen |
For a 110 kW drive at 400 V / 200 A, a 2% reactor is rated roughly 220 A continuous, ~1.5-2 mH per phase. Manufacturers include:
- Schneider Electric Vl range (VW3-A4...
- ABB line reactors (LR series, e.g., LR-D-200-4).
- Siemens 6SL3000-0CE32-3AA0 (matching SINAMICS line reactor, fits MM440 by rating).
- MTE RL series, TCI HRL, Hammond HX series.
Mounting: vertically, with at least 100 mm clearance top/bottom for cooling, in the same cabinet or in a separate wall-mount enclosure if floor space is limited. Keep the cable between breaker and reactor short (≤ 10 m) to limit radiated heat from high-di/dt switching transients in the reactor.
8. Solution 3: Adjust the Protection Relay Settings
If the upstream protection cannot be replaced and a line reactor is not feasible, the next step is to set the existing phase-failure / phase-imbalance relay so it ignores harmonic-driven measurement jitter. The procedure:
- Capture the steady-state RMS line current on each phase with a True-RMS meter (see Section 5). Record the average and the peak deviation.
- Compute the imbalance percentage:
Imbalance % = (max(I_phase) - min(I_phase)) / avg(I_phase) × 100% - If the imbalance % is between 5% and 15% (typical for 6-pulse VFD on average-responding CTs), raise the relay's imbalance threshold from the default 5-10% to 20-30%, AND increase the trip time delay to at least 5-10 seconds so transient reading noise does not latch the trip.
- Disable the negative-sequence voltage element if the relay has one (it will also see harmonic negative sequence as if it were a real negative-sequence fault).
- Use the relay's "fundamental only" filter if available. Some relays (Schneider Sepam, ABB REF615, Siemens 7UT75) can be configured to extract the 50/60 Hz component from the measured current using a DFT or Goertzel filter.
9. Solution 4: Long-Term Harmonic Mitigation
If the facility has many 6-pulse drives and the utility is enforcing IEEE 519 or a local harmonic standard, consider these longer-term options:
| Option | THDi Reduction | Cost (relative) | Footprint | Notes |
|---|---|---|---|---|
| Line reactor (2-5%) | 25-65% | $ | Small | Lowest cost; first step |
| Passive harmonic filter (tuned to 5th + 7th) | 50-70% | $$ | Medium | Risk of resonance with supply |
| Active harmonic filter (AHF / SineWave) | Up to 95% | $$$ | Medium | Schneider AccuSine, ABB PQF, Siemens 4A) |
| 12-pulse rectifier + phase-shift transformer | 5-8% THDi | $$$$ | Large | Requires dual 6-pulse bridge and Δ-Y / Δ-Δ transformer |
| Active Front End (AFE / Low Harmonic drive) | 3-5% THDi | $$$$ | Same as drive + LCL filter | Best solution; requires IGBT rectifier section |
For a single 110 kW fan drive, the most common retrofit is a passive 5th + 7th harmonic filter or a 2-3% line reactor. For plants with many drives, an active harmonic filter sized to 200-300 A installed on the main bus is usually more cost-effective than per-drive filtering.
10. MM440 Parameter Verification
Verify the following MM440 parameters to ensure the drive is operating normally and the apparent unbalance is not caused by drive-side issues (P parameters are user-adjustable, r parameters are read-only):
| Parameter | Description | Expected Value |
|---|---|---|
| P0003 | Access level | 3 (expert) for full access |
| P0010 | Commissioning parameter | 0 = Ready, 1 = Quick, 30 = Factory |
| P0100 | Europe / North America / kW / hp | 0 = Europe [kW], 50 Hz |
| P0201 | Rated drive power | 110 kW |
| P0205 | Rated input voltage | 400 V (verify nameplate) |
| P0304 | Rated motor voltage | 400 V |
| P0305 | Rated motor current | Nameplate value (≈ 200 A) |
| P0307 | Rated motor power | 110 kW |
| P0308 | Motor power factor (cos φ) | Nameplate (typically 0.85-0.90) |
| P0210 | Supply voltage | 380-480 V (set to actual) |
| P2001 | Reference frequency | 50 Hz |
| P1240 | Vdc controller configuration | 0 (disabled) or 1 (enabled) per application |
| r0034 | DC link voltage actual | 540-620 V (depends on supply) |
| r0021 | Output frequency | Operating value |
| r0024 | Output frequency smoothed | Operating value |
| r0027 | Output current actual | Operating value (True-RMS, balanced) |
| r0032 | Active current actual | Operating value |
| r0069 | Actual phase current (fundamental) | Should show three balanced values (via r0069[0], r0069[1], r0069[2] on S7 mapping) |
| r0947 | Last fault code | 0 if no fault |
| r0948 | Fault time | Check if F0020 occurred |
Note: the r0069-style per-phase current readouts are derived inside the DSP from the inverter's output currents, not from the line-side input. The drive itself does not measure line-side currents as a standard feature. To measure line-side fundamental current you must use an external analyzer.
11. F0020 - Mains Phase Failure
The MM440 monitors its own DC bus ripple and infers mains phase failure from excessive ripple. The fault code is F0020 ("Mains phase failure" / "Line phase missing"). Behavior:
- Trigger condition: DC bus ripple exceeds a threshold derived from P0290 (line supply configuration) and the DC bus voltage r0034.
- Default response: Fault (drive stops, output disabled, fault relay opens).
- Reaction setting: P0290 = 0 (trip on missing phase), 1 (warn only on missing phase - drive continues but flag is set).
- Threshold adjustment: P0290 = 1 lets the drive ride through single-phase loss for a short time, which is sometimes useful on weak supplies but not for normal fan drives.
To read the MM440 fault buffer via the BOP (Basic Operator Panel):
MENU > DIAGNOSE > r0947 (last fault)
r0948 (time stamp)
r0949 (value)
r0951 (fault code list)
To read via the AOP (Advanced Operator Panel) or Starter/STARTER commissioning software, navigate to the Diagnostics → Fault Buffer view.
12. Quantifying Real vs. Apparent Imbalance
Once you have a power quality analyzer capturing line-side data, the fundamental component (50 Hz) gives the real imbalance and the total RMS gives the apparent imbalance. Example for a 110 kW fan drive running at full load:
| Phase | I_RMS total (A) | I₁ fundamental (A) | THDi | Apparent Imbalance | Real Imbalance (I₁) |
|---|---|---|---|---|---|
| L1 | 218 | 195 | 52% | +6% | +1% |
| L2 | 202 | 193 | 50% | -2% | 0% |
| L3 | 195 | 194 | 48% | -5% | 0% |
| Avg | 205 | 194 | 50% | — | — |
In this example the upstream meter reports a 6% / 2% / 5% imbalance, but the real fundamental imbalance is only 1% / 0% / 0%. The 5-6% apparent imbalance is entirely from harmonic content being misread.
13. Effects on the Supply System and Equipment
Although the apparent unbalance is mostly a measurement artifact, the underlying harmonics do have real effects on the supply system:
- Voltage distortion at the PCC: the harmonic currents flowing through the source impedance create voltage harmonics. A stiff utility (low Z_source) sees minimal voltage distortion; a weak supply (diesel generator, long cable) can see voltage THD > 8%, which violates IEEE 519 and can damage other equipment.
- Overheating of supply transformers and cables: 5th and 7th harmonics add constructively in the neutral of a 3-phase system, and increase I²R losses in cables and transformers. The K-factor of a transformer feeding multiple VFDs should be K-13 or higher.
- Resonance with PFC capacitors: utility PFC capacitor banks at 5th harmonic can resonate with supply inductance, amplifying the 5th harmonic voltage. Always check resonance before installing capacitor banks.
- Protection relay misoperation: the original symptom in the field report. Negative-sequence overcurrent elements, especially older static relays, can trip on harmonic content.
- Motor bearing currents: high-frequency common-mode voltage from the inverter can discharge through motor bearings, causing pitting. A shaft grounding ring or insulated bearing on the non-drive end is recommended for 110 kW motors.
None of these effects mean the drive itself is damaged or malfunctioning. The MM440 is designed to draw this harmonic current. The remedy is upstream - on the protection, on the supply impedance, or on the source.
14. Recommended Commissioning Procedure for a New Installation
To avoid the field report scenario at the commissioning stage, the following procedure is recommended for any MM440 (or any 6-pulse VFD) installation above 30 kW:
- Install a 2-3% line reactor at the drive input as a default. Cost is a fraction of the drive; benefit is reduced harmonic current, reduced DC bus ripple, reduced RFI, and reduced stress on the upstream breaker.
- Use True-RMS protection relays with bandwidth ≥ 2.5 kHz on the upstream side. Set the imbalance threshold to 15-20% (not 5%) and the trip delay to 5-10 s.
- Use True-RMS clamp meters (Fluke 87V, Hioki CM4373, etc.) for any field measurement. Do not use average-responding meters on VFD inputs.
- During commissioning, perform a power quality survey on the line side at full load with a three-phase power analyzer. Document the THDi, TDD, and individual harmonic magnitudes. File the report with the installation record.
- Verify F0020 is not active in the fault buffer after 1 hour of full-load run. If it is, investigate supply quality.
- Mark the drive as "non-linear load" on the single-line diagram and on the panel. Future electricians will then know not to interpret the current reading on an average-responding meter as the true RMS.
15. Verification Checklist
| Check | Method | Pass Criteria |
|---|---|---|
| Drive status | MM440 display, r0947 | No fault, F0020 not in buffer |
| DC bus voltage | r0034 on BOP/AOP | Within ±10% of 1.35·V_LL |
| Fundamental current balance | Power analyzer, 50 Hz component | < 5% imbalance |
| THDi | Power analyzer | IEEE 519 compliance (typically < 8% TDD for < 69 kV systems) |
| Crest factor | Oscilloscope + Hall probe | 1.7 - 2.0 with 2% reactor |
| Voltage THD at PCC | Power analyzer | < 5% IEEE 519 |
| Protection relay behavior | Field test, raise threshold if needed | No nuisance trips under normal load |
| Motor temperature | PT100 / RTD if installed | Within insulation class B/F rating |
16. When the Unbalance Is Real
The following symptoms indicate a real current imbalance that is not an instrument artifact:
- MM440 reports F0020 (mains phase failure) or F0021 (ground fault).
- One phase of the line-side current measured with a True-RMS meter is > 10% lower than the other two.
- Voltage on one phase at the drive input terminal is > 3% lower than the other two (measured with True-RMS voltmeter).
- DC bus ripple (r0034) shows > 8% peak-to-peak variation at 100/120 Hz.
- Loose terminal on breaker, line reactor, or drive input; corroded bus bar; blown line-side fuse.
If any of these appear, the drive is correctly flagging a real supply problem. Repair the upstream issue (tighten terminals, replace fuse, contact utility) before continuing operation. Do not raise the protection threshold to mask a real fault.
17. Migration to SINAMICS G120
For new installations, the MM440 has been replaced by the SINAMICS G120 family. The G120 has improved input rectifier topology, optional built-in line filter class A or B, and a Low Harmonic variant (G120P / G120E) with an active front end. Parameter mapping from MM440 to G120 is documented in the Siemens migration guide; the diagnostic approach to input current unbalance is identical.
Is the 10-20 A input current unbalance on my MM440 a real fault?
No, in the vast majority of cases it is a measurement artifact. A 6-pulse rectifier draws non-sinusoidal current with crest factor 2.0-2.5, and average-responding clamp meters or protection relays misread this. Verify with a True-RMS meter or a three-phase power analyzer - the fundamental 50 Hz component should be balanced within 2-5%.
My upstream phase-failure relay trips when the MM440 starts. How do I fix this without disabling the relay?
Three options in order of preference: (1) replace the relay with a True-RMS unit rated to at least 2.5 kHz bandwidth; (2) install a 2-3% line reactor at the drive input to reduce THDi from ~70% to ~30%; (3) raise the relay's imbalance threshold to 15-20% and the trip delay to 5-10 s, and document the change. Option 1 is the engineering best practice; option 3 is the field expedient.
What does the F0020 fault code mean on the MM440?
F0020 is "Mains phase failure" - the drive has detected excessive DC bus ripple, which it interprets as a missing supply phase. It can be configured via P0290 to either trip (P0290=0, default) or warn only (P0290=1). If F0020 trips on a healthy supply, verify terminal tightness, check for a blown line-side fuse, and confirm the supply voltage is balanced within 3%.
Do I need a line reactor on a 110 kW MM440?
Yes, Siemens strongly recommends a 2% line reactor for any MM440 above 5.5 kW to limit input current THD and reduce stress on the DC bus capacitors. For 110 kW, a 2% or 3% reactor rated to ~220 A continuous is the standard recommendation. A reactor is the lowest-cost mitigation; passive filters, active filters, 12-pulse rectifiers, and active front ends are progressively more expensive.
What True-RMS meter should I use to measure the MM440 input current?
Use a True-RMS digital clamp meter with bandwidth of at least 1 kHz (preferably 5 kHz) and crest-factor rating of at least 3.0. Recommended models: Fluke 87V, Fluke 376 FC, Fluke 381, Hioki CM4373, Kyoritsu KEW 6305, Chauvin Arnoux F407. Avoid average-responding meters (most cheap digital clamps). For a full picture, use a three-phase power quality analyzer such as a Fluke 435-II, Hioki PW3198, or Dranetz HDPQ.