Resolving Siemens MM440 F002 Overvoltage on Fan Loads

1. Problem Summary

A Siemens MICROMASTER 440 (MM440) inverter rated 75 kW is driving a 55 kW three-phase induction motor coupled to a centrifugal fan load. The drive is controlled over PROFIBUS-DP (cyclic PZD communication, typically PPO type 1 or type 3). Under normal running, the drive delivers the demanded speed setpoint and the motor accelerates/decelerates cleanly. When the PLC issues the stop command, however, the drive faults with F002 (DC link overvoltage) within the deceleration ramp.

Key symptoms recorded from the field:

• Fault code: F002 appears in r0947 shortly after a stop request.
• DC link voltage measured through r0026 climbs to 795 V DC while the threshold P2172 is set to 850 V DC.
• The drive is sized larger than the motor (75 kW inverter / 55 kW motor), so the inverter's continuous current capability is not the limiting factor.
• Mechanical load is essentially friction-only air resistance (fan with no gearbox losses to speak of).
• Initial ramp-down time (~10 s default or a few tens of seconds) is far too short for the rotor inertia.

This is one of the most common field faults on MM440 installations driving high-inertia fans and pumps. The root cause is regenerative energy feedback that exceeds the inverter's ability to absorb it through the DC bus, and the solution tree is well-defined: extend the deceleration ramp, enable the Vdc-max controller, and where the inertia demands it, use a flying restart together with pulse-block coast-down to avoid DC bus pumping entirely.

2. The F002 Fault on MICROMASTER 440

F002 (DC link overvoltage) is generated when the smoothed DC bus voltage in r0026 exceeds the trip threshold defined by parameter P2172. For a 400 V class MM440 (3-phase, 380–480 V AC line input), the nominal DC bus sits around 540–620 V DC, and the factory default for P2172 is 850 V DC. Exceeding this value triggers an immediate pulse inhibit, the drive enters fault state, and the motor coasts to rest.

The MM440 has a brake chopper integrated only on the larger frame sizes (FX/GX) when an external braking resistor is fitted to terminals B1+, B2-, and the inverter is set up in P1230 = 1 (brake chopper enabled) and P1231 (chopper duty). The 75 kW MM440 frame ships with the chopper hardware present but the braking resistor is a field-installed accessory. Without a resistor, all of the regenerative energy must be absorbed by the DC bus capacitance, which is exactly the failure mode being observed.

3. Why Fans Trip F002 — The Physics of Regenerative Braking

When the drive decelerates a motor faster than the load can mechanically absorb the kinetic energy, the motor becomes a generator. The slip reverses, the active power flow into the motor reverses, and energy is pushed back into the DC bus of the inverter. Two paths exist to dispose of this energy:

1. Convert it to heat in a braking resistor (via the chopper).
2. Re-route it to the line through a regenerative line-side converter (active front end). The standard MM440 does not have this; only the Active Line Module topologies (SINAMICS S120, G120P) do.

If neither path is available, the DC bus capacitance charges. The relationship is:

ΔE_bus = ½ × C_dc × (V_trip² − V_nom²) = ½ × J × (ω_start² − ω_end²) − W_loss

where:

• C_dc = total DC link capacitance (µF), typically a few thousand µF for a 75 kW class unit.
• V_trip = F002 trip voltage (850 V DC).
• V_nom = nominal DC bus at full speed (~600 V DC).
• J = total inertia at the motor shaft (kg·m²).
• W_loss = mechanical + windage + switching losses during the ramp (small for a fan).

For a 55 kW centrifugal fan the rotor + impeller combined inertia is on the order of 2 to 8 kg·m². Stored kinetic energy at 1500 rpm (157 rad/s):

E_k = ½ × 5 × 157² ≈ 61.6 kJ

If the operator tries to dissipate that in 10 s, the average braking power is ~6.2 kW; in 5 s it is ~12.3 kW. With a 75 kW inverter, those values seem trivial — and they are, for the inverter. The problem is that the DC bus has no way to sink that power because no chopper/resistor is fitted. The DC link simply climbs to 850 V DC and trips.

4. PROFIBUS Control Word Sequencing on the MM440

The original control sequence being sent by the PLC master is:

Bit mapping (PROFIdrive v3 standard, applies to MM440 with PROFIBUS option module):

The sequence 0x047E → 0x047F → 0x0476 is the textbook PROFIdrive state machine: Ready → Operate → Stop. The Bit 0 transition in 0x0476 initiates a controlled ramp-down to zero using P1121. There is nothing wrong with the sequencing itself — the issue is that the ramp defined by P1121 is shorter than the time constant of the mechanical system.

5. Reading r0026 and Confirming the Root Cause

The trace performed by the user — r0026 climbing to 795 V DC while P2172 = 850 V DC — is the smoking gun. To capture this in the field, use one of three methods:

1. STARTER commissioning tool trace (Siemens). Add r0026 (V DC), r0027 (output current A), r0063 (actual speed rpm) and r0080 (actual torque) to a trace, trigger on r0026 > 750, duration 5 s pre / 30 s post.
2. DriveMonitor (legacy) trace function over RS-485/USS. Sample at 100 ms.
3. External HMI reading parameter r0026 over PROFIBUS non-cyclically (PKW channel) and trending on the HMI. Set the HMI poll at 250 ms.

Look specifically for the signature of regen pumping: r0026 rises after Bit 0 of the control word is cleared, peaks at or near P2172, and the fault occurs on the rising edge, not at constant running. If the trace shows r0026 falling monotonically during decel, the issue is something else (e.g., line overvoltage, weak supply).

6. Solution 1 — Extend the Deceleration Ramp (P1121)

The simplest intervention. P1121 defines the time the ramp generator takes to bring the output from maximum frequency (P1082) down to 0 Hz. Recommended first-pass value for a 55 kW fan with the existing 75 kW MM440:

The discussion confirms that P1121 = 110 s cleared the fault in this installation. That is consistent with the field calc: at 110 s the average regen power drops to roughly 0.6 kW, well within what the DC bus capacitance and internal losses can absorb without exceeding the 850 V trip.

When to use this alone: When the process tolerates a 60–120 s decel. Fans on HVAC systems usually do. Fans on process cooling that must re-start on demand (e.g., a furnace combustion air fan) may not.

7. Solution 2 — Activate the Vdc-max Controller (P1240)

The MM440 has a built-in DC bus voltage regulator that dynamically stretches the deceleration ramp when the bus voltage starts to climb, preventing F002. This is the proper software-only solution and should be enabled in parallel with a conservative P1121.

How it works: the Vdc-max controller monitors the DC bus, and when it exceeds the engagement threshold (typically around 700 V DC for 400 V class), the controller overrides the ramp frequency profile and reduces the deceleration rate in real time, allowing the drive to absorb the regen energy at a rate the bus can handle. Combined with P1121 = 30–60 s, this resolves the vast majority of MM440 fan F002 faults without any external hardware.

8. Solution 3 — Flying Restart (P1200) and Pulse-Block Coast-Stop

For very high-inertia fans — large boiler ID/FD fans, large centrifugal compressors, large exhaust fans — the cleanest control approach is to not brake at all. Instead:

1. Send OFF2 (Bit 1 = 0, control word = 0x047A) on stop. The inverter instantly inhibits pulses, the motor coasts.
2. Set P1200 = 1 to 6 (Flying restart active). With the 6kW and larger MM440 firmware, the drive will scan for an already-rotating motor on the next start command and resume control from whatever speed it finds.

This is the "contactor stop + flying start" pattern. Advantages:

• No regen energy ever flows into the DC bus because the drive never decelerates the motor — it disconnects.
• Elimination of the F002 trip mode by construction.
• Suitable for fans that may already be windmilling before the drive is turned on (e.g., exhaust fans on a windy rooftop).

Disadvantages:

• No controlled decel — the motor coasts to a stop on its own (which for a large fan can take 10–30 minutes).
• Process interlocks must be rethought (e.g., access doors cannot be opened while the fan is still rotating).

The user in the discussion explicitly identifies this as the "very good solution for most fans" — it is, but only when uncontrolled coast-down is acceptable to the process.

9. Solution 4 — Braking Resistor Sizing

If a controlled decel is required and the deceleration time is too long for the process (e.g., a kiln fan that must be re-started within 60 s for a recipe change), install an external braking resistor and enable the chopper. The MM440 75 kW frame (FX or GX) has integrated chopper hardware; the resistor is the only field-installed component.

For a 55 kW fan on a 75 kW drive, design the resistor for the peak regen power. Using the kinetic energy calculation:

P_peak ≈ (J × ω²) / t_brake

For J = 5 kg·m², ω = 157 rad/s, t_brake = 30 s:

P_peak = (5 × 157²) / 30 = 4.1 kW continuous during the 30 s ramp

But the worst case is the start of the ramp when the motor is at full speed and the kinetic energy must be removed quickly:

P_peak = (E_k) / t_brake ≈ 61.6 kJ / 30 s = 2.05 kW average

For 15 s decel the average is ~4.1 kW. Add a 2× safety factor for the resistor continuous rating and a 3× factor for the peak rating. Typical MM440 resistor recommendation for a 75 kW frame with a 30 s decel is in the 6 kW continuous / 25 kW peak range. Siemens braking resistor catalog numbers (verified for the FX/GX frame MM440):

When using a resistor, change the control strategy to use only Bit 0 (ON/OFF1) for start/stop, with all other bits held high. The ramp defined by P1121 is now dissipating into the resistor, so you can shorten the decel to whatever the resistor and process can sustain.

10. Decision Tree — Which Solution to Apply

11. Step-by-Step Commissioning Procedure (Recommended for this case)

1. Connect the MM440 over PROFIBUS with STARTER or SIMATIC TIA Portal. Confirm the PPO type (1 = 4 PZD / 2 PKW, 3 = 2 PZD / 2 PKW) matches the PLC hardware configuration. Verify cyclic PZD is established (green link LED on the PROFIBUS option).
2. Reset the drive to factory defaults (P0010 = 30, P0970 = 1) if you have inherited a non-standard commissioning.
3. Run a quick commissioning (P0010 = 1), entering motor nameplate data: 55 kW, motor voltage, motor current, motor cos φ, motor frequency, motor speed.
4. Set P1120 = 30.0 and P1121 = 60.0 as a starting point.
5. Set P1240[0] = 1 to enable the Vdc-max controller.
6. Set P1200 = 1 (flying restart active after mains on, fault, OFF2).
7. Save to the BOP or upload via PROFIBUS.
8. Test cycle 1: from the PLC, send 0x047E → 0x047F with a 30 Hz setpoint. Let the drive accelerate. Then send 0x0476 (Bit 0 cleared, OFF1 ramp to 0) and observe the trace of r0026, r0027, and r0063. The DC bus must remain below 750 V DC throughout the ramp.
9. Test cycle 2: stop the drive, wait 30 s, restart with 0x047E → 0x047F while the fan is still coasting down. The drive must catch the rotating motor and ramp it back to setpoint without an F0001 (overcurrent) or F002.
10. If test cycle 1 still trips F002, double P1121 to 120 s and re-test. If still trips, raise P1245 to 1.30 and re-test.
11. If process still cannot tolerate a 120 s decel, install the 6 kW braking resistor per Solution 4 and shorten P1121 back to 15–30 s.

12. Verification Checklist

• r0026 peaks at < 750 V DC during the entire decel trace.
• No F002 in the fault buffer (r0947[0..7] all zero) after 10 stop-start cycles.
• No F0001 on restart with the fan still coasting (flying restart success).
• No F003 (overcurrent) on accel — verify P1120 is not too short for the motor's locked-rotor current.
• Actual decel time matches P1121 ±10% with P1240 enabled. If the actual decel is much shorter, Vdc-max is not engaging (check P1247).
• DC bus settles back to ~600 V DC nominal within 2 s of the drive entering "ready" state.

13. Frequently Asked Questions