Variable Frequency Drives

A variable frequency drive takes the fixed-frequency supply and manufactures a new one, at whatever frequency it chooses, so it can run a motor at whatever speed you want.

That is the whole idea. Everything else — the fault codes, the parameters, the diagnostic traps — comes out of how it does it.

Why they are everywhere

Speed control on a fan or a pump is worth money: flow goes roughly with speed, but power goes roughly with the cube of speed, so running a fan at 80 % speed instead of throttling it at full speed uses about half the energy. Beyond that, drives give controlled acceleration instead of a DOL slam, torque control, soft stopping, and a start that does not draw six times full-load current.

The one thing to hold on to

A drive knows a great deal about itself and almost nothing about your machine. It measures its own output current, its own bus voltage, its own heatsink temperature. It infers what the motor is doing. It has no idea what the load is attached to.

That is why the most valuable diagnostic question in this chapter is not "what does the fault code mean" but "did the drive trip on its own protection, or was it told to stop by something else?" Those are different worlds. The first is a conversation with the drive; the second means the drive did as it was told and something external switched it off — and the answer is somewhere out on the machine.

Three stages, and every fault code lives in one of them

The rectifier turns incoming AC into DC. The DC bus holds it on a bank of capacitors at roughly 1.35 times the incoming AC voltage — around 540 V DC on a 400 V supply. The inverter switches that DC on and off through IGBTs thousands of times a second, varying the width of each pulse so the average follows the sine wave it is imitating.

Placing a fault code on that picture is most of the diagnosis:

| Code | Stage | What it usually means | | --- | --- | --- | | Undervoltage | Rectifier / bus | Supply dip, lost input phase, loose input terminal | | Overvoltage | DC bus | Decelerating too fast — the motor is generating into the bus | | Overcurrent | Inverter output | Load, short, or an accel ramp the load cannot follow | | Earth / ground fault | Output side | Motor or motor cable insulation | | Overtemperature | Drive | Blocked filters, failed fan, high panel ambient | | Overload (I²t) | Drive's model | Sustained current above rating — usually a real load problem | | External fault | None of them | Something else told it to stop |

Volts per hertz, and why low speed is not free

Torque in an induction motor depends on magnetic flux, and flux depends on the ratio of voltage to frequency. Hold that ratio constant and torque stays available all the way down. That is what a drive does below base speed.

Above base speed it cannot: it only has the incoming supply to work with, so voltage flattens out while frequency keeps rising. V/Hz falls, flux falls, torque falls. This is the field-weakening region, and it is why running a machine faster than nameplate has a ceiling that is about torque rather than about the drive's willingness.

The left-hand end is what catches people out in practice. A standard TEFC motor has a cooling fan on its own shaft. Slow the motor and you slow its cooling, while the load may still be demanding full torque and therefore full current. Sustained low-speed running needs either forced ventilation, a motor rated for it (an inverter-duty motor), or a reduced continuous rating.

Parameters that actually matter

Drives have hundreds of parameters and about a dozen decide whether a machine runs properly.

• Motor nameplate data — voltage, current, frequency, rated speed. The drive's motor model is only as good as these, and a drive that has been swapped without them entered correctly will misbehave in ways that look like mechanical faults.
• Acceleration and deceleration ramps — the commonest cause of both overcurrent and overvoltage trips.
• Current limit — how hard the drive will push before it protects itself.
• Minimum and maximum frequency — and the skip frequencies used to jump over a mechanical resonance.
• Control mode — simple V/Hz for fans and pumps, sensorless vector where torque matters at low speed.
• DC injection braking and brake resistor settings, if fitted.

Regeneration, and why fast stops trip on overvoltage

When a drive decelerates a spinning load faster than the load wants to slow down, the motor becomes a generator. That energy has nowhere to go but back into the DC bus, which pushes the bus voltage up. Push it far enough and the drive trips on overvoltage to protect itself.

The fixes are all about where that energy goes: extend the decel ramp so the load coasts more, fit a brake resistor to burn it as heat, or on larger installations use a regenerative front end that puts it back into the supply. High-inertia loads — big fans, centrifuges, anything with a flywheel — are where this shows up.

Output cabling is part of the drive

The PWM output has extremely fast voltage transitions, and a long motor cable behaves as a transmission line. Reflections at the motor end can nearly double the voltage seen by the winding insulation, so manufacturers specify maximum cable lengths and often recommend output filters beyond them.

The same fast edges push small currents through the motor bearings, which over time etch the races — the classic fluting failure on a drive-fed motor. Shaft grounding rings and insulated bearings exist for this. Both problems are invisible until a motor fails early, which makes them worth knowing about rather than discovering.

1. Read the fault code and the fault history, not just the current fault

   The active fault tells you what stopped it this time. The history tells you whether this is the fourteenth occurrence at the same point in the cycle, which is a completely different problem. Note timestamps and, where the drive records it, the output frequency and current at the moment of the trip.

2. Establish whether the drive protected itself or was told to stop

   This is the branch point for everything that follows. An internal protection trip — overcurrent, overvoltage, overtemperature — is a conversation with the drive. An external fault input means the drive was doing its job and something else switched it off, so the answer is out on the machine and not in the panel.

3. Note what the drive was doing when it tripped

   Accelerating, decelerating, running at steady speed, or sitting at zero. Overcurrent on acceleration and overvoltage on deceleration are almost different faults with different causes, and the drive usually records which phase of the cycle it was in.

4. Compare the drive's own readouts against the motor nameplate

   Output current against motor FLA, output frequency against expected speed, DC bus voltage against roughly 1.35 times the supply. These are free, they need no instrument, and they are trustworthy — the drive is measuring its own terminals.

5. Check the incoming supply under load

   Safety

   All three phases at the drive's input terminals, measured while the machine is working. Balance matters as much as magnitude, exactly as in Chapter 5. A loose input terminal reads perfectly at rest and sags under load.

6. Check cooling before believing anything thermal

   Drive fan running, heatsink clear, panel filters clean, panel fans working, door seals intact so the airflow goes where it was designed to go. Read the heatsink temperature from the drive rather than estimating it.

7. Separate the drive from the load

   The move that settles most arguments. Uncouple the motor if you practically can, or jog it unloaded, and compare the current with the loaded figure. A drive and motor that behave perfectly uncoupled have told you the fault is mechanical — which is exactly the Conveyor 04 finding.

8. Test the motor and cable with the drive disconnected

   Safety

   Following Chapter 4: isolate, prove dead, wait out the DC bus, disconnect at the drive's output terminals, then insulation test the motor and cable as a unit. Never test through the drive — it will destroy the output stage, often with a delayed failure that nobody attributes to the test.

9. Check the parameters against the record, then back them up

   Particularly after any replacement: nameplate data, ramps, current limit, control mode. When the machine is running properly again, save the configuration. The best moment to make that backup is the moment you have just proved the settings are correct.

• Replacing the drive because the drive is displaying the fault

  Why

  The drive is the only component on the machine with a screen, so it is the only one that ever appears to be complaining. But most of what it reports is observation of things outside itself — an overcurrent is usually the load, an earth fault is usually the cable, an external fault is by definition something else. Swapping it is expensive, slow, and frequently changes nothing.

• Reading only the active fault and not the history

  Why

  The active fault is on the screen and the history is three menus away, so under pressure the screen wins. The history is where the pattern lives — same fault, same point in the cycle, fourteen times in two days is a diagnosis, and a single occurrence is barely a data point.

• Lengthening the accel ramp until the trips stop

  Why

  It genuinely fixes the trip, and sometimes a ramp that was always too aggressive is the real answer. The trap is using it when the load has changed: the drive was reporting that this machine now needs more current to start than it used to, and stretching the ramp silences the report while the bearing or the build-up that caused it carries on.

• Insulation testing without disconnecting the drive

  Why

  The machine is isolated, the test is the thorough thing to do, and the cable runs from the drive to the motor — so testing "the motor circuit" from the panel feels natural. It puts hundreds of volts DC across the output semiconductors. Chapter 4 covers why the resulting failure is often delayed and therefore rarely attributed.

• Running a standard motor slowly for long periods

  Why

  The drive will happily hold 10 Hz all day and nothing complains immediately. But the motor's cooling fan is on its own shaft, so it is now producing possibly full torque with a small fraction of its airflow. The failure arrives months later as a burnt winding and gets recorded as a motor fault.

• Not backing up parameters until after a drive has failed

  Why

  The moment the backup is worth having is the moment it no longer exists. Nothing prompts you to do it, the machine runs fine without it, and the cost lands entirely on whoever is standing in front of a dead drive at three in the morning with a replacement and no settings.

Five questions. Each is answerable by reasoning about what the drive can and cannot see, rather than by recalling a fault-code table.