Resolving SINAMICS S120 Speed Fluctuation in U/f Control Mode

1. Problem Statement: Speed Oscillation on a SINAMICS S120 U/f-Driven Stirrer

A reactor stirrer driven by a 4 kW / 1500 rpm induction motor through a SINAMICS S120 drive system (CU310 DP + PM340) was reported to run cleanly at very low speeds but develop severe oscillation once the setpoint rose above approximately 25 rpm. The user-visible behavior was a speed fluctuation swinging between approximately +150 rpm and -150 rpm of setpoint, with the mechanical agitator visibly bouncing in the positive (drive) direction only while the setpoint itself remained constant. The fault was diagnosed as residing somewhere downstream of r1119 but upstream of r1150, persisted after the PM340 Power Module was replaced, and was ultimately eliminated by upgrading the firmware stored on the Control Unit's flash card to V2.6. This article reconstructs that fault, maps the diagnostic chain parameter by parameter, and documents the corrective procedure plus the long-term control-mode recommendation that emerged from the field.

2. System Configuration and Hardware Identification

The reported installation is a classic low-power SINAMICS S120 AC/AC drive built from a separate Control Unit and a Blocksize power module. The catalog numbers reported in the source post and their meaning are summarized below.

Refer to the SINAMICS S120 with CU310-2 DP Operating Instructions and the SINAMICS S120/S150 List Manual for the canonical parameter descriptions used throughout this article.

3. Symptom Description and Operational Boundaries

The user reported a very clean operating boundary:

• Setpoint ≤ 25 rpm: speed is steady, torque stable, no audible or visible oscillation.
• Setpoint > 25 rpm: actual speed oscillates with an amplitude of approximately ±150 rpm around setpoint; the mechanical shaft is visibly reciprocating in the positive direction only.
• Directionality: oscillation is biased positive — the drive overshoots in the drive direction and undershoots in the regenerative direction, ruling out a pure mechanical unbalance.
• Setpoint tracking: the value visible at r1119 tracks the commanded setpoint exactly; the value visible at r1150 shows the oscillation. The fault is therefore introduced between those two diagnostic parameters.
• Ramp-function generator bypass: bypassing the RFG (e.g., toggling p1115 or forcing the setpoint directly into the V/f modulator) did not clear the oscillation, ruling out the RFG itself.
• Hardware swaps: replacing the PM340 made no change, ruling out power-stage failure. Motor insulation, winding resistance, gearbox, and bearings were checked by the user.

4. SINAMICS S120 Setpoint Chain — Where the Fault Lived

The setpoint path inside a CU310-configured S120 running U/f control is a fixed sequence of function blocks. Understanding which parameters the user compared is essential to narrowing the fault domain.

The signal therefore becomes unstable in the slice between r1119 (clean) and r1150 (oscillating). This is the I_max / frequency-controller pre-control region, governed by:

• p1340 — I_max frequency-controller proportional gain
• p1341 — I_{max frequency-controller integral time}
• p1349 — I_{max frequency-controller speed-setpoint damping}
• p1082 — maximum speed
• p1138 / p1139 — ramp-time scaling

Above 25 rpm, the drive entered the active range of the I_max controller (typical with a load that produces a variable reaction torque as the agitator blades bite into the medium). With the firmware shipped on a 2007-vintage CU310, this controller's pre-control path was known to oscillate under certain motor-data / load-inertia combinations — a defect that was corrected in firmware V2.6 and forward.

5. Root Cause: Outdated Firmware on the CU310 CompactFlash Card

The drive had been commissioned in 2007. The CompactFlash card inside the CU310 still carried the firmware that was current at that time (most likely a V2.4 or V2.5 line of the SINAMICS S120 firmware). When the firmware was upgraded to V2.6 (using STARTER / Startdrive with the new firmware image copied to the CF card and a CU warm restart), the oscillation disappeared without any parameter change. The PM340, motor, gearbox, and mechanical train were retained.

This is consistent with documented SINAMICS S120 firmware-2.6 release notes, which explicitly address closed-loop behavior of the V/f characteristic and I_{max pre-control on small-block PM340 modules. See the SINAMICS S120 firmware release notes index for the canonical changelog. The root cause is therefore:}

1. Time-domain instability in the I_max / frequency-controller pre-control region of the firmware version loaded on the CF card.
2. Symptom amplification by the variable reaction torque of the stirrer blades (load-dependent torque demand acting on an open-loop V/f drive without torque pre-control).
3. No hardware fault — the PM340 swap correctly exonerated the power stage.

6. Resolution: Flash Card Firmware Upgrade to V2.6

The corrective action is a standard CU310 firmware update. Procedure summary, with the assumption that STARTER or Startdrive is already installed on the engineering station.

6.1 Prerequisites

• STARTER ≥ V5.x or Startdrive in TIA Portal, with current SINAMICS S120 Support Package.
• Firmware image file for V2.6 (e.g. SINAMICS_S120_CU310_DP_V2.6.exe or the equivalent .zip with *.awl/*.bin contents).
• The exact saved project backup (so parameters can be restored after the upgrade).
• Empty CompactFlash card (Siemens 6SL3054-0AA00-1BA0, ≥ 1 GB) or the existing card if only the firmware is being replaced in-place.

6.2 Step-by-Step Procedure

1. Back up the project. Connect the CU310 via PROFIBUS or Ethernet (service interface X127) to the engineering PC. In STARTER / Startdrive select Project → Save project to PG and additionally Drive → Upload → From CF card to PG. Confirm the offline backup contains parameters p0010=0, r0021=0, and a complete parameter dump.
2. Power down the drive per LOTO; wait for the DC-bus discharge time on the PM340 (typically 5 min, until the READY LED extinguishes).
3. Remove the CF card. The slot is on the front of the CU310 under the small access cover. Eject gently; do not force.
4. Re-write the card with the V2.6 image using one of:
   • STARTER → Target system → Firmware update (writes directly to the inserted CF card via a USB card reader), or
   • Manual: extract the firmware ZIP to the card's root directory; the card must contain at minimum fwdata.bin, cmdfile.txt, and the licence folder.
5. Re-insert the CF card into the CU310, restore power, and observe the boot LEDs. The CU will auto-detect the new firmware on the next power-on.
6. First commissioning step after FW update. STARTER will prompt for a factory reset. Accept; then Download the saved parameter set back to the drive. Confirm p0010 returns to 0 (Ready) and r0019.0 shows Operation.
7. Verify motor identification if any motor data field shows [0] or [?]: run Motor identification via p1910 = 1 (still) or p1960 = 1 (rotating, requires the drive to be safe to spin unloaded). Re-enter nameplate values into p0304, p0305, p0307, p0310, p0311.
8. Re-acceptance test. Run the stirrer at the previously problematic setpoints (50, 100, 200 rpm). Observe r0021, r0023, r1119, and r1150 on the trace — the three must coincide and remain stable.

7. Alternative Tuning Steps if Firmware Cannot Be Upgraded

If a firmware upgrade is operationally blocked (e.g. obsolete plant approval cycle, no licence available), the oscillation can be damped by parameter manipulation alone. None of these are as clean as the firmware fix, but they are useful in an emergency.

8. Why U/f Control Is Suboptimal for Reactor Stirrers

U/f (open-loop V/Hz) control was originally intended for fans, pumps, and conveyors — applications where the load torque varies slowly with speed and where a small steady-state speed error is acceptable. A reactor stirrer is the opposite of that profile:

• Highly variable reaction torque. The impeller pushes against a viscous medium. As the medium aerates, heats, or reacts chemically, viscosity drops and load torque can swing by a factor of 3–10 within seconds.
• No torque pre-control. The V/f curve is a static voltage-to-frequency mapping; the drive has no knowledge of actual shaft load. Speed droops under load and recovers only through the slow I_max controller.
• Stall behaviour. At low frequency (25 rpm is ≈ 1.7 % of nominal), the V/f curve has very little voltage headroom and is dominated by stator-resistance drop and the configured boost. Any load step easily pushes the drive into the I_max loop, which is exactly the region that oscillated in this incident.
• Asymmetric load reaction. When an impeller clears a viscous pocket, the reaction torque releases asymmetrically, producing the very "positive-only" oscillation that was reported.

The cure is to move the drive to sensorless vector control (SLVC), which observes rotor flux via the motor model and actively regulates torque. With SLVC enabled via p1300 = 20 (or 21 for SLVC with FCC), the drive handles the variable torque demand in the torque-control inner loop rather than reacting through the slow outer V/f path. The required steps are documented in the SINAMICS S120 Commissioning Manual and summarized below.

9. Migration Path: U/f → Sensorless Vector Control (SLVC)

1. Capture the existing parameter set (Step 6.1 above).
2. Verify motor nameplate data in p0301–p0311; resolve to a Siemens 1PH/1LE/1LA code or enter equivalent values.
3. Run Motor data identification with p1910 = 1 (still) and p1960 = 1 (rotating) to populate p0350–p0369 automatically. The rotating test requires a safe spin condition; lock the impeller mechanically if needed.
4. Switch control mode: p1300 = 20 (SLVC without encoder) or = 21 (SLVC with FCC).
5. Tune speed controller: p1460 (Kp), p1462 (Tn). Run the SINAMICS auto-tuning routine (Drive → Controller → Speed controller auto-tuning) which sets these values based on the measured inertia.
6. Adjust torque limits: p1520 (upper torque limit), p1521 (lower torque limit) — typically ±150 % of rated motor torque for a stirrer.
7. Enable torque pre-control: p1496 = 1 for acceleration torque pre-control.
8. Verify under load. Trace r0021, r0023, r0080, r0084, and r1119/r1150. Speed deviation between setpoint and actual should stay inside ±1 % under full-load steps.

SLVC requires no encoder on this 4 kW motor — the drive estimates rotor position from the stator current model. If the application later demands <0.5 % speed accuracy at very low speeds, fit an incremental encoder and use p1300 = 22 (vector control with encoder).

10. Verification Procedure After the Fix

After a firmware update or a control-mode migration, a structured re-acceptance proves the fault is gone.

11. Field-Proven Caveats and Lessons Learned

• Old firmware on long-life assets is a known failure mode. Drives commissioned in 2007 with V2.4 / V2.5 firmware will reach end-of-corrective-support for many bug classes. Plant owners should schedule a firmware-upgrade audit on a 5-year cycle, even if the drive appears to run correctly — once support lapses, the firmware becomes a single point of failure.
• PM340 swap did not fix it, but that was a correct diagnostic step. Whenever a drive misbehaves in a way that looks like a power-stage problem, swap the power module first (5-minute job) to eliminate that domain, then move to the Control Unit / firmware.
• Directionality of the oscillation is a powerful clue. Pure mechanical unbalance produces oscillation in both directions; a control-loop oscillation driven by the I_max pre-control biased toward the drive (positive) direction is the signature of a V/f instability under variable torque load.
• The "delay of the ramp generator" hypothesis floated in the discussion is incorrect on its own — the user reported the oscillation persists with the RFG bypassed, and the divergence between r1119 (clean) and r1150 (oscillating) places the fault downstream of the RFG.
• The "module too small" hypothesis raised at the end of the discussion deserves attention: the PM340 catalog number reported (6SL3210-1SE21-0AA0) is the 1.0 kW variant. If that part number is correct, the drive has been operating in continuous I_max limitation, which both accelerates the failure mode of the firmware instability and indicates a sizing error that must be addressed separately. Cross-check the part number against the PM340 Blocksize Datasheet.
• Switch to SLVC as a follow-on improvement. U/f is structurally inadequate for mixers; even with the firmware fix, the drive will continue to droop under sudden load changes because it has no torque pre-control.

12. Frequently Asked Questions