Troubleshooting MM440 Drive EMI on 3RG6013 Ultrasonic Sensor

Problem Overview

A MICROMASTER 440 (MM440) 2.2 kW inverter (Siemens part number 6SE6440-2AD22-2BA1) is energised on the same machine as a 3RG6013-3AD00 SONAPROBE ultrasonic sensor wired into a SIMATIC S7-1200 DC digital input. The motor cable is approximately 50 m long. With the drive stopped, the sensor behaves correctly and the S7-1200 input LED follows the target. The moment the drive is released, with any speed reference, the S7-1200 input LED begins to chatter on and off while the sensor's own status LED stays off. The PLC therefore counts pulses continuously even though the sensor's output transistor is not being commanded by the transducer electronics. From the S7-1200's perspective the input is reading a clean 24 V signal most of the time, but the digital input is being momentarily pulled below the LOW threshold and back above the HIGH threshold at a high repetition rate.

This is a classic common-mode EMI symptom: the IGBT bridge in the MM440 is switching the DC bus at 2–16 kHz (parameter P1800), and the resulting dv/dt on the motor terminals capacitively couples high-frequency current from the motor windings into the motor frame, the motor cable, and ultimately back to the drive's DC bus. If the motor cable shield does not present a low-impedance 360° path back to the drive chassis, that common-mode current finds the next available return path - which in a typical cabinet is the sensor cable shield, the sensor body, or the 24 V control wiring. Once the current flows through the sensor cable, it lifts the sensor's 24 V supply or pulls the PNP output below the S7-1200 HIGH threshold long enough to register as a logic transition.

System Topology and Components

The 6SE6440-2AD22-2BA1 is the factory variant with the integrated Class A line filter populated. The filter addresses conducted emissions back onto the line side; it does not address motor-side common-mode voltage. The motor-side dv/dt problem must be solved with cable shielding, output reactors, sinusoidal filters, or a combination. Treat the "it has a filter" claim as addressing only the supply side and not the field-wiring problem described here.

Root Cause: PWM Common-Mode Current and the Motor Cable

Modern IGBT inverters do not produce a sine wave on the motor terminals. They produce a pulse-width modulated waveform that switches the DC bus at P1800 (default 4 kHz on the MM440) with edge rise times in the 100 ns to 1 µs range. The spectrum of those edges extends well into the MHz region. The motor windings, the motor frame, the cable conductors, and the cable screen form a distributed capacitive network. The phase-to-ground capacitance of a typical 4-pole induction motor at 50 Hz is on the order of 4–10 nF; at MHz it looks like a low impedance to ground.

Every switching edge dumps a small common-mode current pulse (typically 0.5–5 A peak) into that capacitive network. The return path must be a low-inductance shield that is bonded 360° at both ends. If the screen is interrupted - by a piggy termination, a cable break, an isolator without bonding straps, or a non-metallic junction box - the common-mode current must find another route. The most common alternative routes, in order of likelihood, are:

1. The control cable shield (sensor cable, encoder cable, brake cable).
2. Capacitive coupling from the motor frame to the structural steel / ground grid, then back to the drive chassis through the PE conductor.
3. Conduction through shared 24 V supply or signal commons.

When the common-mode current flows on the sensor cable, it raises the local ground reference of the sensor. Because the S7-1200 input is referenced to the cabinet 0 V rather than the sensor body, the voltage difference appears as a noise spike on the input line itself. If the spike crosses below 5 V on a 24 V PNP output, the input de-asserts. If the spike recovers above 15 V before the input filter time constant, the input re-asserts. The result is exactly the on/off chatter observed on the S7-1200 LED.

Parameter Investigation: P1800 vs P1080

Field discussions sometimes confuse parameter numbers. On the MM440 the relevant parameters are:

Verify the live value on the drive with the BOP or via USS/Modbus to parameter r0018 (read-only) which echoes the actual switching frequency. Lowering P1800 from 4 kHz to 2 kHz reduces the number of switching events per second by 50 %, which directly reduces the RMS common-mode current. The trade-off is increased motor audible noise and a small increase in current ripple / torque ripple. On a 2.2 kW MM440, 2 kHz is normally acceptable for fans, pumps, conveyors, and similar quadratic or constant-torque loads, but is generally not suitable for spindle or low-noise applications.

Required Cable and Shielding Practice

For a 50 m motor run the cable must be a symmetric 3- or 4-core screened cable with a dedicated PE conductor inside the screen. Steel Wire Armour (SWA) is acceptable as the screen if it is terminated with the appropriate EMC gland and bonded 360° to the drive chassis and the motor terminal box. PVC-jacketed CY or YY cable with a foil + braid screen is acceptable for shorter runs but should be replaced with a robust SY or SWA cable for installations above 25 m.

1. Strip back the outer jacket at the drive end by the minimum needed to reach the terminals.
2. Expose the braid or armour full-circumference; do not twist it into a pigtail.
3. Use an EMC cable gland (e.g. Lapp SKINTOP MS-SC, Weidmüller VGM, or equivalent) that clamps the screen 360° against the gland body. The gland body must be in metal-to-metal contact with the gland plate, which must be in metal-to-metal contact with the drive chassis, which must be bonded to the cabinet PE bus.
4. Repeat the same procedure at the motor terminal box. The motor frame has a dedicated grounding stud; the screen must be terminated to that stud using a 360° clamp, not a ring terminal on a single screw.
5. If a motor isolator is fitted, the screen must be continuous through the isolator. Use a copper braid jumper (typically 16 mm² or larger, length as short as practicable) bonded to the screen at both sides of the isolator. The standard "screen continues past the isolator" approach without a jumper will not work because the isolator contacts open the screen path.
6. Pigtails are unacceptable. A pigtail adds series inductance that is transparent at 50 Hz but acts as an open circuit at MHz. The same applies to "drain wire" terminations where the drain wire is wrapped around the cable once and landed on a single terminal screw.

Drive-Side Output Filtering Options

If the screen integrity is confirmed but the chatter persists, add a motor-side filter to slow the edge rate and reduce the common-mode current at the source. The MM440-compatible Siemens accessories are documented in the Siemens Industry Online Support catalogue for the MICROMASTER 440 product family.

For sensor chatter specifically, a common-mode choke on the motor output is often the most cost-effective choice. It does not affect motor performance, and it directly attacks the coupling path that the sensor cable is picking up. The choke should be installed at the drive end, as close to the MM440 output terminals as possible, with all three phase conductors passed through the same core in the same direction.

Sensor-Side and Signal-Side Mitigations

Sensor-side changes should be applied in parallel with cable and drive-side fixes, never as a substitute for them.

1. Use a screened sensor cable. A 3-core screened cable (e.g. LiYCY 3 x 0.34 mm²) bonded 360° at the cabinet gland plate and terminated with a drain wire or pigtail no longer than 25 mm at the sensor body. The sensor body itself, if metallic, should make contact with the cable screen via the cable gland at that end as well.
2. Add a ferrite on the sensor cable. A clip-on ferrite (e.g. Lapp FER-13, Würth WE-FLEX, or equivalent) at the cabinet entry side. For MHz-range common-mode current, a MnZn ferrite with 200–500 Ω impedance at 25 MHz is appropriate. One or two turns through the core.
3. Add a 24 V decoupling network. Place an RC snubber (e.g. 100 Ω + 100 nF X2-class) across the sensor's 24 V and 0 V terminals at the cabinet entry. This shunts RF current to local chassis ground before it can reach the S7-1200 backplane. The capacitor must be the safety / X / Y class approved for connection across mains-derived 24 V.
4. Configure the S7-1200 input filter. In TIA Portal, on the CPU properties → Digital inputs, increase the input filter time constant. Options typically include 0.1 ms, 0.5 ms, 3 ms, 10 ms, and 20 ms. Moving from 0.1 ms to 3 ms or 10 ms will suppress brief noise spikes while still capturing legitimate sensor transitions. For ultrasonic sensors with cycle times in the tens of milliseconds, 3 ms is usually invisible to the application.
5. Add a software debounce. In the S7-1200 user program, implement a debounce counter or timer on the input tag. A 5–10 ms "on" delay plus a 5–10 ms "off" delay will reject the noise bursts while preserving the true sensor edges. This is the last line of defence and must not be the only line.

Cabinet Layout and Cable Segregation

Even with perfect shielding, parallel runs of motor and signal cables inside the cabinet will couple energy. Apply the following segregation rules:

• Maintain at least 200 mm between unshielded motor conductors and 24 V signal cables. If both are in cable trays, use separate trays with a metal divider bonded to chassis.
• If a crossing is unavoidable, cross at 90°.
• Do not run the sensor cable in the same conduit as the motor cable, even if both are screened.
• Bond both ends of the divider / tray to the cabinet PE bus. A floating divider is no divider.

The 24 V power supply feeding the sensor should be located in the same cabinet as the S7-1200, not remote. Remote 24 V supplies create long return paths that couple noise easily. If a remote supply is unavoidable, run the 0 V return next to the +24 V and bond both ends to chassis at the source and at the load.

Verification Procedure

After applying the cable and drive-side fixes, run the following verification sequence. Document baseline and post-fix values.

1. Power-down inspection. With the drive isolated and locked out, measure the screen continuity from the motor frame to the drive chassis ground stud with a 4-wire milliohm meter. Reading should be < 1 Ω. Inspect the gland clamps visually for full 360° contact. Inspect the motor terminal box for the same.
2. Static check. With drive powered and stopped, verify the S7-1200 input reflects the actual sensor target. Move a target in and out of the sensor beam and confirm the input follows with the configured filter time. No chatter is expected.
3. Dynamic check - low speed. Run the motor at 5 Hz for 30 s, then 10 Hz for 30 s, then 25 Hz for 30 s, then 50 Hz. Monitor the S7-1200 input tag in TIA Portal with online watch. The input should remain stable at the last valid state throughout, regardless of the sensor's actual output (because the target is fixed). Any toggling indicates residual coupling.
4. Oscilloscope check. Place a 10:1 passive probe on the sensor output at the S7-1200 terminal, with the probe ground clip on the S7-1200 0 V terminal. Set the scope to 1 V/div, 10 µs/div, AC-coupled. With the drive running at 25 Hz, look for negative-going spikes that cross below ~5 V. The amplitude of the spikes is the figure to record. Acceptable: < 1 V peak. Borderline: 1–3 V peak. Unacceptable: > 3 V peak.
5. Common-mode current check. Clamp a current probe around the sensor cable (all three conductors). With the drive running, the RMS current should be < 5 mA. If it is above 50 mA, the cable shield is not doing its job or the drive-side filter is missing.
6. Frequency sweep. If the chatter still occurs at one speed but not another, run a frequency sweep from 5 Hz to 50 Hz in 5 Hz steps. Note the speeds at which chatter occurs. If chatter is speed-dependent, suspect a resonance in the cable / sensor capacitance; P1800 adjustment or a common-mode choke is the correct response.

Troubleshooting Matrix

Cross-Reference to a Sibling Machine

The reported case includes a second, similar machine where the wiring and cabling are very close to the failing machine, the only difference being an S7-200 in place of the S7-1200. Use the working machine as a benchmark: clamp a current probe around the working machine's sensor cable and around the failing machine's sensor cable while both are running. The working machine should show a much lower common-mode current. If it does not, then the working machine is a latent failure waiting to happen, and the same fixes should be applied to both.

Acceptance Criteria for the Fix

The fix is considered complete when all of the following are true:

• Sensor output is stable at every speed from 0.1 Hz to 50 Hz, with P1800 set to 4 kHz.
• S7-1200 input tag does not toggle while the sensor target is fixed and the drive is running.
• Oscilloscope measurement on the sensor output shows < 1 V peak noise.
• Common-mode current on the sensor cable is < 5 mA RMS.
• Motor cable screen resistance from motor frame to drive chassis is < 1 Ω.

Document References

Authoritative Siemens documentation for the components in this scenario is available on the Siemens Industry Online Support portal at https://support.industry.siemens.com. For the MICROMASTER 440, search the support portal for "6SE6440" to retrieve the operating instructions, parameter list, and accessory catalogue. For the S7-1200 system manual and the digital input filter configuration, search "S7-1200 System Manual". For the SONAPROBE 3RG6 family, search "3RG6013" for the product datasheet. For cabinet-level EMC installation guidance, refer to the SIMATIC S7-1200 Automation System manual, the section on EMC-compliant installation, and the IEC 61800-3 standard for the EMC requirements applicable to adjustable-speed drive systems.