SINAMICS PM230 EMC/EMI: P1800 Limits, Filters, Cable Screening

1. Problem Description

A SINAMICS PM230 unfiltered output power module, part number 6SL3210-1NE31-5UL0, rated for 75 kW, has been installed as part of the propulsion or auxiliary-drive system of a vessel. From the moment the inverter is enabled, low-voltage thermocouple signals feeding the ship's automation network drift and produce erroneous temperature readings. The fault is independent of motor load, and the disturbance disappears within a fraction of a second after the inverter is stopped. The motor feeder has already been wired with a screened, low-capacitance cable, but the screening alone has not eliminated the coupling. The three open engineering questions are:

1. Why does parameter P1800 reject the value 2 kHz even though P1801 appears to indicate a 2 kHz minimum?
2. What is the practical difference between a filtered and an unfiltered PM230, and what is the most economical output filter for a 75 kW load?
3. Can the unfiltered unit be returned to Siemens Italy to be retrofitted to a filtered variant?

2. Affected Hardware and Configuration

The unit at the centre of this case is identified from the MLFB as follows:

The UL0 suffix means no integrated line-side EMC filter is fitted. The unit is intended for cabinet installations where an external filter is mounted on the line side, or for installations where conducted-emission limits do not apply. On a vessel, EMC requirements are dictated by the classification society (DNV, Lloyd's Register, ABS, BV, RINA) and by IACS UR E22. Shifting the EMC compliance burden to the system integrator means the integrator must add filtering, screening, and bonding discipline to bring emissions under control.

For full device ratings, refer to the SINAMICS G120 PM230 operating instructions on Siemens Industry Online Support and to the SINAMICS G120 parameter manual for the firmware build currently loaded on the unit.

3. PWM Ringing and EMI Coupling Physics

The dominant interference mechanism on the load side of any IGBT inverter is not the supply-side harmonics but the high-frequency common-mode (CM) voltage that the inverter's pulse-width modulation creates between the DC bus negative rail and the chassis (PE). The PM230 switches the IGBTs at a programmed pulse frequency that is typically 4 kHz by default. Each switch transition has a rise time in the 50–200 ns range, which produces a voltage step with a slope du/dt easily exceeding 5 kV/μs.

This step is impressed across the stray capacitance of the motor cable (typically 0.2–0.4 nF/m between conductor and screen for a standard 4-core motor cable). The CM current that flows as a result is:

For a 20 m cable at 0.3 nF/m and 5 kV/μs:

The total CM current summed across all three phases and re-circulated through the screen is of the order of tens of milliamps with sub-microsecond edge rates. The spectral content extends well into the VHF band (30–300 MHz), which is why the disturbance is referred to as "MHz ringing" in the field report. The wavelength of a 100 MHz signal is 3 m in free space, so a 20 m motor cable behaves as a multi-wavelength transmission line for this interference, with the cable screen acting as the return path.

The interference enters thermocouple circuits because:

1. Thermocouple voltages are in the 0–50 mV range, and even a few millivolts of coupled noise represents a temperature error of several tens of degrees for a type K thermocouple (41 μV/°C).
2. Thermocouple extension wire typically has no screen, and the input stage of a marine automation analogue card has high input impedance and limited CM rejection at VHF.
3. The PM230 chassis, the motor frame, and the automation cabinet are coupled through the ship's hull and bonding network, providing a low-impedance return path that the disturbance will exploit if the dedicated return path (the motor cable screen) is broken or high-impedance.

4. Root Cause Analysis

From the discussion thread, four interacting root causes are present in this installation:

4.1 PWM MHz common-mode voltage on the load side

The pulse-frequency default of 4 kHz produces edge rates high enough to drive substantial CM current through the motor cable capacitance. The PWM fundamental itself is not the problem — the resonant overshoot and ringing created by the IGBT commutation loop, the cable inductance, and the motor winding capacitance are the problem. This ringing typically settles within 200–500 ns and produces a damped oscillation in the 20–80 MHz range.

4.2 Screen bonding not adequately low-impedance at VHF

Pigtails, earthing posts, and short jumper wires all look low-impedance at 50 Hz but become inductive at VHF. At 50 MHz the impedance of a 50 mm pigtail is already 16 Ω, and at 100 MHz it exceeds 30 Ω. The screen must be clamped to the chassis with a 360° bond (EMC backshell or saddle clamp) at both the inverter end and the motor end, and that bond must be continuous through any isolator fitted in the cable.

4.3 Multiple parallel return paths through the ship's bonding network

If the dedicated screen return path is not the lowest impedance available, the disturbance will return to the inverter through whatever path is lower — typically the ship's hull, the automation cabinet earth bus, and the thermocouple cable shields. This is why a 10 m thermocouple run behaves as if it is sitting directly on the inverter CM noise source.

4.4 No output filtering to limit du/dt and peak voltage

An unfiltered PM230 has no dv/dt reactor and no sine filter on its output. The IGBT edge rate reaches the motor terminals unchanged. At cable lengths above the manufacturer's shielded-cable limit, peak voltages at the motor terminals can exceed 1.5 kV, accelerating winding insulation ageing and broadening the conducted and radiated emission spectrum.

5. Diagnostic Procedure

Run the following procedure before changing any hardware or parameters. Each step isolates one root cause.

1. Read the active fault and alarm buffer from the PM230 (r0947 / r2122) and note any F30002 (DC-link overvoltage), F30004 (converter overtemperature), F07900 (motor blocked) or F07801 (motor overtemperature) entries.
2. Capture P1800, r0027 (actual output current), r0037 (converter temperature), and r0206 (rated power unit power) via STARTER, Startdrive, or the IOP/BOP.
3. With the drive enabled and the motor unloaded, use a battery-powered oscilloscope (not a mains-powered scope) and a high-voltage differential probe to view the line-to-line voltage at the inverter output terminals. Look for ringing with a peak above 1.4 · V_DC-link and a frequency above 20 MHz.
4. With the same scope, view the CM voltage between any output phase and the cabinet chassis earth bar. A reading above 30 V peak at the PWM fundamental indicates that the cabinet earth bar is carrying substantial CM current.
5. Move the scope probe to the screen of the motor cable at the inverter end. Measure the screen-to-chassis voltage. Anything above a few hundred millivolts peak indicates that the screen bond at that point is inadequate.
6. Disconnect the motor cable screen from the motor junction box and re-measure with a low-resistance bond (≤ 5 mΩ DC, verified with a four-wire milliohm meter) at the screen gland.
7. Repeat the thermocouple reading test with the drive stopped vs the drive enabled at zero speed reference. If the noise disappears when the drive is stopped, the coupling path is from the inverter, not from the ship's electrical system.

Document each measurement. They form the baseline against which every change is verified.

6. Resolving P1800 Pulse-Frequency Limits

The first question is why P1800 will not accept 2 kHz. Three mechanisms can cause this. Work through them in order.

6.1 Confirm what P1801 actually represents

In the standard SINAMICS G120 parameter list for the PM230 firmware in question, there is no universal parameter named P1801 that exposes a "minimum pulse frequency" setting. P1800[0...n] is the only pulse-frequency write parameter, and it is indexed (one entry per Drive Data Set in most G120 configurations). The figure shown on the BOP/IOP that the operator has read as "P1801" may in fact be:

• A read-only display parameter such as r0113 (pulse frequency recommendation) or r0024 (output frequency).
• One index of P1800 shown without its index, in which case only P1800[0] has been edited while another index (e.g. P1800[1]) still holds the higher factory default.
• A customer-specific parameter introduced by a project-specific parameter set, or a misread parameter number.

Cross-check against the parameter list of the firmware build installed on the CU. The exact list is part of the SINAMICS G120 parameter manual matching the firmware version of the Control Unit.

6.2 Drive operating state and access level

P1800 can only be written when the drive is in commissioning access level and not in Run. The relevant conditions are:

• Drive in state S1 (Commissioning) or S2 (Ready for operation), not S3 (Run) or S4 (Fault active).
• BOP access level set to "Expert" (parameter P0003 = 3) or higher; some firmware variants gate P1800 behind "Expert" access.
• No active safety function preventing pulse-frequency reduction. STO does not lock P1800, but SS1 with a time-to-ramp shorter than 200 ms can override it.
• No active bypass contactor or motor changeover (if a motor changeover is configured, P1800 is locked at the higher value used for the larger motor).

6.3 Power-module-specific floor

For PM230 power modules, the minimum pulse frequency rises with the rated current of the unit. For 75 kW PM230 units (frame FSE), the firmware typically enforces a minimum pulse frequency of 4 kHz in some firmware builds and 2 kHz in others. If the installed firmware enforces 4 kHz, the rejection of 2 kHz is correct operation, not a configuration error.

To resolve:

1. Read r0190 (Power unit identification) and confirm the MLFB structure against the part number on the rating plate.
2. Read r0200 (actual power unit code number) and confirm it matches the firmware-expected code number.
3. Read the firmware version of the CU (read from r0018 or from the firmware download log) and consult the parameter list for that exact version.
4. If the floor is firmware-enforced, the 4 kHz default cannot be reduced to 2 kHz by parameter editing alone. The remaining option is to reduce the disturbance at the source by adding an output filter (Section 8) and tightening the screen bonding (Section 9), not by reducing the switching frequency.

7. Filtered vs Unfiltered PM230 Comparison

The ship installation described in the source already shows the symptom of a class-A filter being unsuitable: nuisance tripping of the generator earth-fault detection caused by filter discharge current. The discussion thread explicitly warns that supply-side filters can encourage such trips. The filtered PM230 has the same problem because the filter capacitance is internal to the unit and cannot be bypassed. This is why the most economic answer is often not to add a line filter at all, but to combine:

1. Output filter (dv/dt reactor or sine filter) on the load side.
2. Disciplined screen bonding on the motor cable.
3. Physical separation and routing of the thermocouple cable away from the motor feeder.

8. Output Filter Selection for 75 kW

Three output filter topologies are viable for a 75 kW PM230. Their effects on the noise mechanism in this case differ.

For thermocouple-noise mitigation on a 75 kW drive, a sine filter is the most effective because it converts the PWM output to a near-sinusoidal waveform and removes the high-frequency harmonics responsible for CM coupling. A dv/dt reactor is cheaper but does not suppress the CM noise enough to fully resolve a thermocouple pickup problem. The Voltage Clamping Limiter (VPL) is the most cost-effective compromise where motor insulation protection is also a concern.

8.1 Filter sizing

Use the following approach to size the filter for a 75 kW PM230 on a 400 V line:

1. Calculate the base impedance from the rated three-phase apparent power:

For a 75 kW unit at 0.85 power factor:

2. Select filter inductance at 2–3 % of base impedance for a dv/dt reactor, or at 5–8 % for a sine filter:

3. Verify the filter continuous current rating is ≥ the drive rated output current. At 75 kW on 400 V, this is approximately 145 A (verify against the unit's r0207).
4. Verify the filter switching-frequency compatibility — the filter must be rated for the actual switching frequency that P1800 is set to, with margin for transient excursions.

9. Cable Screening and Bonding Procedure

Before adding any filter, apply the following screening discipline. The field report thread is unambiguous that screen bonding is the single highest-impact intervention.

9.1 Required materials

• EMC backshells (or saddle clamps) sized for the motor cable outer diameter, at both ends.
• Bonding braid or short strap (≤ 50 mm length) from the EMC backshell to the cabinet chassis at the inverter end.
• EMC cable gland at the motor junction box, providing 360° contact with the screen.

9.2 Procedure at the inverter end

1. Strip the motor cable outer sheath back far enough to expose 25–40 mm of the braided screen, while leaving the screen intact and continuous.
2. Fit the EMC backshell so the saddle clamps the exposed screen with full circumferential contact. Verify the clamp compresses the screen onto a bare-metal saddle.
3. Bond the backshell body to the cabinet chassis via a short braid (≤ 50 mm) to a chassis-bonding stud. Do not pigtail the screen to an earth post.

9.3 Procedure at the motor end

1. Use an EMC cable gland rated for the cable diameter. The gland's contact ring must bear directly on the screen.
2. Verify the motor junction box body is bonded to the motor frame with a low-impedance bond (≤ 5 mΩ).
3. Confirm the motor frame is bonded to the ship's hull at the motor mounting feet. On a steel-hulled vessel this is automatic; on aluminium or GRP hulls a dedicated bonding strap is mandatory.

9.4 Procedure through any isolator

If a switch-disconnector or isolator is fitted in the motor feeder:

1. Do not interrupt the screen inside the isolator. The screen must remain continuous.
2. Clamp the entering and leaving screens together inside the isolator with a saddle clamp, so the screen is bonded across the isolator body.
3. Bond the isolator body to the same chassis earth bar as the inverter.

9.5 Screen-to-screen contact verification

After assembly, measure the DC resistance from the screen at the inverter end to the screen at the motor end. The reading should be below 100 mΩ for a 20 m cable. Above 1 Ω indicates a broken screen or a pigtail that needs to be replaced.

10. Isolator and Thermocouple Handling

The thermocouple interference observed in the field report is a direct consequence of the noise currents finding a return path through the ship's bonding network and through the thermocouple wiring. Two interventions are needed.

10.1 Route and shield the thermocouple cable

1. Replace the existing thermocouple extension wire with a screened, twisted-pair type. Screen must be a foil-plus-braid or a metallised-foil with a drain wire.
2. Terminate the screen at the automation cabinet end only — do not bond the screen at the thermocouple head end. This prevents ground-loop current from flowing in the screen.
3. Use a bonding clamp at the cabinet gland plate, with a short bond (≤ 50 mm) to the analogue-earth bar.

10.2 Increase physical separation

1. Maintain at least 200 mm separation between the motor feeder and any analogue signal cable throughout the cable run. Cross them at 90° if they must cross.
2. Where the motor feeder runs through a trunk or tray, route the thermocouple cable in a separate tray with a metal divider.
3. For new installations, use a screened, twisted-pair cable with a separate integral drain wire and bond the screen at one end only.

10.3 Verify against the source warning

The field report explicitly warns that the screen end at the inverter "must be clamped to the chassis, rather than trying to terminate the screen to an earth post." This is because the chassis bond is short and wide (low impedance at VHF), whereas an earth post is typically connected via a pigtail or a short wire that has substantial inductance above 10 MHz. Apply this principle consistently at every screen termination in the system.

11. Retrofitting the PM230 to a Filtered Variant

The third source question — whether the unfiltered PM230 can be sent to Siemens Italy to be upgraded — is not a standard service path. The filter PCB in a PM230 is mounted inside the unit during manufacture and the unit's rating plate carries the filter designation (e.g. UL0 vs AL0). Field retrofits to a different filter designation are not part of the standard service offer, because:

• The MLFB is printed on the rating plate and the unit's certificate refers to that MLFB.
• The internal filter is matched to the IGBT module's switching characteristics.
• EMC compliance testing on a vessel refers to the as-installed configuration.

The correct path is to:

1. Order a replacement PM230 with the filtered designation (the AL0 or equivalent MLFB matching the unfiltered part's electrical ratings). Confirm the exact MLFB with the Siemens regional office.
2. Return the unfiltered unit under the standard RMA process via the local Siemens service contact for repair, not retrofit.
3. If EMC compliance for the vessel still cannot be met with the filtered unit (because of generator earth-leakage interaction), use an external output reactor or sine filter and keep the unfiltered unit.

Contact the regional Siemens service desk through the official Siemens Industrial Automation contact page for the precise RMA and replacement process for the vessel's flag state and classification society.

12. Verification Checklist

Run this checklist after each intervention. A change is successful only if every item below is satisfied.

1. Drive operates at rated speed and rated load without raising F30002, F30004, F07801 or any new fault.
2. r0027 (output current) is within the unit's continuous rating.
3. Drive temperature r0037 stabilises within the rated envelope after 2 hours of full load.
4. Line-to-line voltage at the motor terminals shows no ringing above 1.4 × V_DC-link on the oscilloscope.
5. CM voltage at the inverter cabinet earth bar is below 5 V peak with the drive running at rated load.
6. Screen-to-screen DC resistance is below 100 mΩ for a 20 m cable.
7. Thermocouple reading at the automation HMI matches a reference instrument (handheld calibrator or a second thermocouple with a separate routing) within the documented accuracy class.
8. Vessel classification society EMC verification (IACS UR E22 or equivalent) passes.
9. Generator earth-leakage monitor does not record nuisance trips during drive start, run, or stop.

13. Troubleshooting Matrix

14. Safety and Classification Notes

For the absolute limits on pulse frequency, derating curves, motor cable length, and output filter selection, always defer to the current revision of the SINAMICS G120 PM230 operating instructions and the SINAMICS G120 parameter manual matching the firmware build installed on the Control Unit. The parameter values, fault codes, and MLFB structures referenced in this article are based on the published Siemens documentation available at the time of writing and may have been revised in subsequent firmware releases.