Simovert Masterdrive VC Torque Fluctuation Extruder

Problem Overview

A Simovert Masterdrive VC running an extruder in field service begins reporting an actual torque that swings between 40 % and 80 % of rated, while actual speed and actual current remain visibly stable on the operator panel (PMU) and via Profibus to the controlling S7-400 CPU. The drive has been in service for years, and no warning or fault code is raised. Because the speed loop reports a closed and steady value, the standard operator reflex — "the load must be surging" — does not fit, and the problem almost always traces to the feedback path or to the speed controller gain. This reference documents a field-proven diagnostic flow, parameter-level checks, and the encoder-vs-sensorless comparison that isolates the root cause in roughly 30 minutes on a running machine.

Drive and System Identification

Confirm the exact drive variant before changing parameters. The Simovert Masterdrive VC family (Siemens 6SE70 series) covers compact and chassis units with a common parameter structure. In an extruder installation of this type the following are typical:

The PPO type used in the cyclic Profibus telegram determines which process data words are available to the PLC. For this case, confirm that the words containing the actual torque value (typically PZD6 in PPO type 4 or 5) and the actual speed value (typically PZD3) are read by the S7 program and not swapped or scaled incorrectly. A Profibus scaling error in the S7 side can mimic a fluctuating torque while the drive's local PMU reads clean — always cross-check the PMU value before trusting the PLC tag.

Symptoms and Field Observations

The reported signature is distinctive and rules out several common faults at a glance:

• Torque setpoint / actual value: 40 % → 80 % swing, irregular period (typically 1 s to several seconds), no alarm.
• Actual speed (n-actual): Visually steady on the PMU; PLC trace flat within ±0.5 %.
• Actual current (I-actual): Steady; tracks the load demand but does not oscillate.
• Drive status: Ready, Run, no fault, no warning. r001 (control word) = 0x047F, r002 (status word) = operating.
• Mechanical: No acoustic change, no screw RPM change, no throughput change at the die.

Root Cause Matrix

Map symptoms to the most likely subsystem before opening any parameter. The matrix below is ordered by the expected frequency in field reports on extruder lines.

Diagnostic Flowchart

Step-by-Step Diagnostic Procedure

Work in the order below. Each step either narrows the field of suspects or applies a fix. Do not skip step 1 — the PLC scaling issue costs nothing to check and is a frequent red herring.

1. Compare drive-local vs Profibus values. From the PMU scroll to r020 (actual torque %); from DriveMonitor or the S7 trace read the corresponding PZD. If the PMU is flat and the PLC swings, the drive is innocent — correct the S7 PPO mapping and scaling block. Most CUVC builds transmit torque as a normalized value (16384 = 100 %); the S7 FC should multiply by 100 / 16384 to convert to percent.
2. Inspect the mechanical encoder coupling. Lock out, hand-turn the motor shaft slowly. Watch the drive's n-actual (r037) on the PMU. A healthy coupling shows smooth count change. Steps, jumps, or slip indicate a worn or loose coupling. Replace with a zero-backlash servo-class bellows; align within 0.05 mm; tighten both grub screws against the flat.
3. Re-check encoder cable and shield. Shield must be bonded at the drive end only, terminated 360° at the cabinet gland. Verify continuity of each twisted pair. For HTL encoders, a 1.5 kΩ pull-up / pull-down termination is sometimes required at the SBP terminals — confirm against the encoder card's Siemens SIMOVERT MASTERDRIVE Compendium for the SBP wiring diagram.
4. Run the drive in sensorless vector control (no encoder). Change P100 from 4 to 3. This switches the control structure from closed-loop speed (with encoder) to sensorless vector control (encoder feedback disabled). If the torque now sits flat at the expected load value, the encoder path is confirmed as the cause and you can plan the encoder repair at the next scheduled stop. If the torque still swings, the encoder path is not at fault and the problem lies inside the speed controller or the load.
5. Verify the speed controller K_p. Parameter P235 = speed controller proportional gain. P236 = speed controller integral action time (T_n). Read the present values. Reduce P235 by 30 % (do not change P236 yet). Observe the torque trace. If the swing amplitude shrinks, the controller is over-tuned. Run the auto-tuning routine (P236 = 0 to start, then follow the prompts) or apply the manufacturer's rule of thumb: K_p × T_n ≈ 0.5 × motor mechanical time constant.
6. Decouple the motor from the gearbox / screw and run no-load. If torque still swings with no load, the source is upstream of the gearbox (encoder, controller, motor). If torque goes flat at no-load, the load itself is the cause — inspect screw flight wear, gearbox backlash, coupling to thrust bearing.
7. Re-run motor identification. With the motor decoupled and cold, set P115 = 2 (full auto-tune, static + dynamic), enable the drive, apply a brief run command, and let the drive measure stator resistance, leakage reactance, and magnetizing curve. Store with P971 = 1 (save to EEPROM).
8. Restore production parameters and re-monitor. Reset P100 to 4 if the encoder path is healthy, write parameters back (P971), and re-engage the line.

Encoder Feedback Diagnostics in Detail

The 6SE70 encoder feedback path runs: encoder → shielded cable → SBP/SES slot → CUVC processor → speed controller. A fault anywhere in that chain can produce exactly the symptom in the report — torque computed correctly from i_q and motor model, but speed input to the controller noisy, so the controller issues a noisy torque correction.

Key parameters and indicators for the encoder path:

The encoder option board (SBP) sits in slot 2 or slot 3 of the electronics box. Open the door of the unit, locate the SBP, and confirm it is fully seated in its connector. A partially inserted SBP, often the result of a maintenance intervention, is a recurring field cause of intermittent encoder problems.

Speed Controller Tuning (P235 / P236)

The Masterdrive VC speed controller is a PI structure. P235 sets the proportional gain, P236 the integral time. Their interaction is the second most common cause of an isolated torque swing:

• High P235, low P236 → stiff, fast, but oscillates if load has compliance or if feedback is noisy.
• Low P235, high P236 → soft, slow, but stable. Torque signal becomes smoother because the controller issues smaller corrections per unit of speed error.
• T_n = 0 → pure P-controller, no integral. Speed will sit with a static error, but the torque command will be very smooth.

For a field check, halve P235, leave P236 unchanged, and run for 5 minutes. If the torque swing visibly shrinks, K_p was the cause. The clean fix is a full auto-tune. DriveES or DriveMonitor can launch P236 in the background, but the field-comfortable procedure is:

1. Set P236 = 0 (start auto-tune) and enable the drive.
2. Apply a small positive speed setpoint (e.g. 10 % of rated).
3. Wait for the routine to complete (drive issues short torque pulses and measures response).
4. Read the new P235 and P236 values written by the routine.
5. Save with P971 = 1.

Sensorless Vector Test (P100 = 3)

This is the fastest discriminator in the toolbox. P100 selects the speed-control structure:

To run the diagnostic test:

1. Note the existing P100 value (4 in this case).
2. Switch to P100 = 3 and acknowledge any warning ("feedback missing").
3. Run the line at the normal operating point for 2 minutes.
4. Record the torque trace. Compare amplitude of swing to the value at P100 = 4.
5. Restore P100 = 4 if the encoder path is to remain in service.

A swing amplitude drop of more than 50 % in sensorless mode is a near-certain confirmation of an encoder-path problem. Replace the coupling first, then the cable, then the SBP board, in that order of cost.

Profibus and PLC-Side Verification

The S7-400 reads torque and speed from the drive as PZD words. The 6SE70 with CBP2 supports PPO types 1 through 5. For an extruder the typical configuration is PPO type 4 (six PZD in each direction) or PPO type 5 (ten PZD). The torque actual value is normally PZD6 and the speed actual value PZD3, but this must be confirmed against the drive parameter assignment (e.g. P554, P555 in older terminology or the connector list in DriveMonitor).

Verification steps on the PLC side:

1. In STEP 7 / TIA Portal, open the Profibus diagnostics for the drive DP slave. Verify the slave is in "Data Exchange" state, no diagnostic interrupts pending.
2. Open the process image and read the PZD words live. Convert to engineering units in the S7 program and write to a VAT for trend.
3. Trigger a cross-check: read the same torque value from the drive's r020 using DriveMonitor or a temporary OP1S, and compare. If r020 is flat and the PLC swings, the problem is in the S7 block (scaling, swap, or double-read from a different PZD).
4. Check the PPO mapping against the drive's connector list. A common mistake is to map the connector for actual current (K0151) into the slot where the PLC expects torque — they look similar in magnitude and units.

Mechanical and Load-Side Checks

On extruders, the screw, gearbox, and thrust bearing are coupled to the motor through a coupling that itself can be a source of fluctuation. Distinguish the two cases:

A 30-minute no-load run with the motor decoupled from the gearbox is one of the most informative field tests available and should be scheduled before any deep electronics work.

Verification and Monitoring After Repair

Once a fix is applied, monitor the following for at least one full production shift before signing the work order off:

• Actual torque (r020) on the PMU — peak-to-peak swing should be less than 5 % of rated for a stable extruder.
• Actual current (r019) — flat within the normal load envelope.
• Actual speed (r037) — flat within the closed-loop tolerance band set by P376 / P377.
• Speed controller output (r049) — should be smooth, no high-frequency component.
• Drive temperature and motor temperature — should track the previous baseline.

For a permanent record, configure DriveMonitor to log r020 and r037 at 100 ms intervals for 10 minutes and save as a CSV. Compare the post-repair trace against the pre-repair trace; a swing reduction of more than 80 % is the acceptance criterion.

Parameter Reference Card

Frequently Asked Questions