In this guide
Ask a student how to slow down an induction motor and the first answer is usually "use a smaller pulley" or, worse, "use a fan dimmer". Neither controls speed properly, and the dimmer can destroy the motor. The correct answer — the one used in every factory, pump house and elevator you have ever seen — is the Variable Frequency Drive (VFD): a power-electronics unit that varies both the frequency and the voltage supplied to the motor, giving smooth speed control from near-standstill to rated speed and beyond.
This guide covers what a VFD actually does inside, the motor theory you need (synchronous speed, slip, the V/f ratio), how to size and wire a drive for a student demo rig, the parameter setup sequence, and the safety rules that are non-negotiable around 415 V three-phase power. For the panel-building side, the crane control panel project shows a VFD in a real industrial-style build, while the four-quadrant DC drive project is worth reading for contrast — DC drives solve a related problem a completely different way.
Why motor speed follows frequency
A three-phase induction motor's speed is set by the supply frequency and its pole count:
Ns = 120 × f / P
Ns is the synchronous speed in rpm, f the supply frequency in Hz, P the number of poles. A 4-pole motor on 50 Hz mains: Ns = 120 × 50 / 4 = 1500 rpm. The rotor always runs slightly slower — the difference is the slip (typically 2–5%), which is what induces rotor current. So a "1500 rpm" motor's nameplate usually reads ~1440 rpm.
Change the frequency and the synchronous speed changes proportionally: at 25 Hz the same motor's synchronous speed is 750 rpm. That is the entire principle of VFD speed control. But there is a catch: if you lower the frequency while keeping voltage constant, the magnetic flux in the motor rises (flux ∝ V/f), the core saturates, current spikes, and the motor overheats. So the drive must lower the voltage in proportion to the frequency — keeping the V/f ratio constant. 415 V at 50 Hz becomes ~207 V at 25 Hz. Constant V/f means constant flux, which means the motor produces its rated torque at any speed in the range. This single relationship is the heart of every VFD you will ever commission.
What is inside the box
Every VFD, from a 0.5 HP student unit to a 500 HP industrial drive, has the same three stages:
- Rectifier — converts the AC mains to DC. On small drives this is a simple diode bridge; larger drives may use controlled rectifiers.
- DC bus — a bank of capacitors that smooths the rectified DC (typically ~565 V DC on a 415 V drive, ~325 V on a 230 V drive). These capacitors stay charged for minutes after you switch the power off — the single most important safety fact in this guide, covered below.
- Inverter — six IGBTs (or MOSFETs on small drives) switch the DC bus thousands of times per second using PWM to synthesise a variable-frequency, variable-voltage three-phase output. The motor's inductance smooths the PWM into an approximately sinusoidal current.
Control-wise, small drives offer two modes you will meet:
- V/f (scalar) control — the drive maintains the programmed V/f ratio open-loop. Simple, robust, and entirely adequate for pumps, fans, conveyors and student demos.
- Sensorless vector control — the drive estimates rotor flux and torque from current measurements and regulates them independently. Better torque at low speed (below ~10 Hz) and tighter speed holding under load changes. Most modern small drives include it as a parameter option; for a demo rig, V/f is the honest starting point.
Sizing a VFD for a student demo rig
Match the drive to the motor, not to your ambition:
- Power rating: the drive's kW/HP rating should equal or exceed the motor's. A 1 HP (0.75 kW) motor wants a 0.75–1.5 kW drive. Oversizing one step is common and harmless; undersizing trips on overload.
- Voltage — the student trick: most college labs have only single-phase 230 V, but most small motors are three-phase. The solution is a single-phase-input VFD (230 V in, three-phase 230 V out), widely available up to ~2–3 HP. Wire the motor in delta (230 V winding rating) rather than star (415 V) — check the motor terminal box and nameplate: a "230Δ/415Y" motor is exactly what you want. This one detail unlocks three-phase motor projects in any lab.
- Current rating: check the drive's rated output current against the motor's full-load current on the nameplate — current, not just kW, is what trips the drive.
- Overload capacity: small drives typically tolerate 150% of rated current for 60 seconds. Starting a loaded conveyor needs this headroom; a bare-shaft demo motor does not.
Budget realistically: a 1 HP single-phase-input drive from an established brand costs roughly ₹8,000–15,000 in India; the motor itself ₹5,000–9,000; contactors, MCB, enclosure and wiring add a few thousand more. The project cost guide helps you frame this in a full budget. Never buy the drive from the cheapest no-name listing — a VFD with fake or undersized IGBTs fails explosively, and this is not the component to economise on.
Wiring basics
A typical small drive's terminals:
- Input: R/L1, S/L2, T/L3 (three-phase) or L, N (single-phase-input drives). Feed this through an MCB/MCCB and preferably a line contactor. Size the breaker per the drive manual — typically 1.5× the drive's input current rating.
- Output: U/T1, V/T2, W/T3 to the motor terminals. Never put a contactor, switch or isolator between the VFD output and the motor and operate it while running — opening the output under load destroys the inverter stage. If the design needs isolation, interlock it so the drive is stopped first.
- Earth: bond the drive chassis, the motor frame and the panel earth bar to a proper earth. VFDs generate high-frequency leakage; poor earthing shows up as nuisance tripping, erratic control behaviour and shocks off the motor frame.
- Control terminals: digital inputs (run forward, run reverse, fault reset), analog input (0–10 V or 4–20 mA for speed reference — a simple 10 kΩ potentiometer across the drive's 10 V reference supply is the classic student speed knob), relay outputs (run/fault indication lamps).
Wiring rules that prevent the classic failures:
- Keep motor cables as short as practical and use shielded cable for longer runs, earthed at the drive end — the PWM output radiates interference that resets nearby microcontrollers (your ESP32-based sensorless BLDC controller demo on the next bench will thank you).
- Separate power and control wiring physically; never run the speed-reference wires in the same trunking as the motor cable.
- Torque the power terminals to the manual's spec. Loose terminals arc, overheat and melt terminal blocks — the most common physical failure in student panels.
Parameter setup: the commissioning sequence
Every drive has its own parameter numbering (Delta, ABB, Siemens, Schneider all differ), but the setup sequence is universal. Do these in order, with the motor uncoupled from any load for the first run:
- Enter the motor nameplate data: rated voltage, rated current, rated frequency (50 Hz), rated speed (rpm), number of poles, rated power. The drive's protection and V/f calculations all derive from these — wrong nameplate data is the root cause of most "the drive keeps tripping" complaints.
- Set the control mode: V/f for a first build; sensorless vector once the basics work.
- Acceleration and deceleration time: 5–10 seconds for a demo rig. Too fast an acceleration trips on overcurrent (the motor demands starting current the drive cannot supply); too fast a deceleration trips on overvoltage (the motor regenerates into the DC bus). These two parameters fix more "trips" than any others.
- Minimum and maximum frequency: cap the maximum at 50 Hz (or the motor's rating) until you understand overspeed implications — running a 50 Hz motor at 100 Hz doubles bearing and fan stress and is not a party trick.
- V/f curve / torque boost: a small low-frequency voltage boost compensates stator resistance on small motors. The factory default is usually fine; excessive boost overheats the motor at low speed.
- Carrier (switching) frequency: higher values make the motor quieter (less audible whine) but heat the drive more. Default is fine for demos.
- Command source and speed reference: keypad vs terminal control; potentiometer vs preset speeds. Wire the start/stop buttons and E-stop to digital inputs per the manual's example diagrams.
First-run procedure: power up with no run command, verify the display shows no faults, confirm the DC bus charges (the display usually shows bus voltage), then run at 10 Hz with the motor uncoupled and check rotation direction — swap any two motor phases (U/V/W) to reverse direction, never rewire the input. Ramp to 50 Hz, listen for abnormal noise, check motor current against the nameplate, then couple the load and repeat.
Safety: the non-negotiable section
- The DC bus stays charged after power-off. Wait the time the manual specifies (typically 5–10 minutes), then verify zero voltage at the DC bus terminals with a meter rated for 1000 V DC before touching anything inside. Students have been shocked by "switched off" drives.
- 415 V three-phase can kill. If your rig uses a three-phase supply, the panel build, earthing and first energisation should be supervised by lab staff. Single-phase-input demo rigs are the right scope for unsupervised student work.
- Emergency stop: wire a mushroom-head E-stop to a digital input programmed for coast-to-stop or fast ramp stop — and understand it stops the motor, not the mains. A separate mains isolator with lockout is the full answer for maintenance.
- Enclosure discipline: terminal covers stay on when energised. No "just checking the terminals" with the power on.
- Motor thermal protection: at low speeds the motor's shaft-mounted fan moves little air while the windings still carry full current. For extended low-speed running, fit an external cooling fan or a motor thermistor wired to the drive's fault input. A demo that runs at 10 Hz for an hour can cook a motor that is perfectly happy at 50 Hz.
The energy-saving math (why industry loves VFDs)
For centrifugal pumps and fans, the affinity laws apply:
- Flow ∝ speed
- Pressure ∝ speed²
- Power ∝ speed³
Worked example: a pump running at 80% speed delivers 80% of the flow but consumes 0.8³ ≈ 51% of the power. At 50% speed it consumes 12.5%. This cubic relationship is why VFDs pay for themselves in months on pump/fan loads — and it makes a quantified energy-saving demo (measure input power at several speeds, plot against the cubic curve) a strong project in its own right. The Electrical branch hub lists related builds in this space.
VFD vs the alternatives: decision table
| Method | Speed control | Starting current | Cost | Right when |
|---|---|---|---|---|
| DOL starter | None (full speed only) | 6–8× full-load current | Lowest | Small motors, infrequent starts, no speed need |
| Star-delta starter | None (reduced-voltage start only) | ~2–3× in star | Low | Larger motors needing gentler starts, fixed speed |
| Soft starter | None (ramp start/stop) | 3–4×, adjustable | Medium | Smooth starting, fixed speed, pump applications |
| VFD | Full, smooth, reversible | Limited to ~1.5× by the drive | Highest | Any real speed-control need, energy saving on variable loads |
| Pole-changing / Dahlander motor | 2 fixed speeds | High at each speed | Medium | Legacy two-speed applications |
If the requirement is "start a 5 HP motor without dimming the lab lights", a star-delta starter or soft starter is the economical answer. If the requirement includes the words "variable speed", the VFD is the only correct answer — a soft starter cannot do it, whatever the seller claims.
Troubleshooting table
| Symptom | Likely cause | Fix |
|---|---|---|
| Drive powers up, motor does not start on run command | Command source set to keypad while wired to terminals (or vice versa); E-stop input active | Check command-source parameter; verify digital input states on the display |
| Motor runs backwards | Phase sequence | Swap any two of U/V/W at the drive output |
| Overcurrent trip on starting | Acceleration time too short; wrong motor current parameter | Lengthen accel ramp; verify nameplate current entry |
| Overvoltage trip on stopping | Deceleration too fast; regenerated energy has nowhere to go | Lengthen decel ramp; add a braking resistor for high-inertia loads |
| Motor overheats at low speed | Shaft fan ineffective below ~20 Hz at full torque | Limit continuous low-speed torque, add external cooling fan |
| Nearby microcontroller resets when motor runs | EMI from PWM output cable | Shielded motor cable earthed at drive end; separate power and signal routing |
| Drive trips "motor overload" at rated speed | Nameplate current entered too low; actual overload | Verify nameplate data; measure output current; check the mechanical load |
| Audible whine from motor | Low carrier frequency | Raise carrier frequency within the drive's derating guidance |
Talking to the drive from code (Modbus RTU)
Most small drives expose RS-485 Modbus RTU, which lets a microcontroller or PC set frequency and read status. The sketch below uses Python with the minimalmodbus package — register addresses differ by drive brand and model; take them from your drive's communication manual, they are placeholders here:
import minimalmodbus
import time
drive = minimalmodbus.Instrument('/dev/ttyUSB0', 1) # port, Modbus slave address
drive.serial.baudrate = 9600
drive.serial.timeout = 1.0
FREQ_CMD_REGISTER = 0x2001 # placeholder - see your drive manual
RUN_CMD_REGISTER = 0x2000 # placeholder - see your drive manual
STATUS_REGISTER = 0x2100 # placeholder - see your drive manual
drive.write_register(RUN_CMD_REGISTER, 0x0001) # run forward
drive.write_register(FREQ_CMD_REGISTER, 2500) # 25.00 Hz (drive-specific scaling)
time.sleep(5)
status = drive.read_register(STATUS_REGISTER)
print("drive status word:", hex(status))
drive.write_register(RUN_CMD_REGISTER, 0x0000) # stop
Scale notes: many drives represent frequency in 0.01 Hz units (2500 = 25.00 Hz) and pack run/stop/direction into a command word — both are drive-specific. Get the manual's Modbus map before writing a line of code, and test every command with the motor uncoupled first.
Reading the motor nameplate like a commissioning engineer
Every parameter you enter comes from this one metal plate. A typical student-lab motor reads something like:
3~ Motor | 0.75 kW | 230Δ/415Y V | 3.4/2.0 A | 50 Hz | 1440 rpm | cos φ 0.78 | IP55 | Ins. Cl. F
Decode it once and you will never guess parameters again:
- 0.75 kW — rated shaft output power. Size the drive to at least this.
- 230Δ/415Y V — winding voltage: 230 V in delta, 415 V in star. On a 230 V single-phase-input drive, connect delta. On a 415 V three-phase drive, connect star. Getting this wrong is the most common wiring error in student labs — a delta-connected motor on 415 V draws excessive magnetising current and trips or burns.
- 3.4/2.0 A — full-load current in delta/star respectively. Enter the one matching your connection into the drive.
- 1440 rpm at 50 Hz — confirms a 4-pole motor (1500 synchronous minus slip).
- cos φ 0.78 — power factor at full load; useful when estimating input current for breaker sizing.
- IP55 / Insulation class F — ingress protection and thermal class; class F insulation tolerates 155 °C winding temperature, which is part of why the low-speed cooling warning matters.
Photograph the nameplate before the motor goes into the rig — once it is bolted down behind a coupling guard, the plate is often unreadable.
Braking: what happens when you stop a heavy load
When a VFD decelerates a motor, the motor briefly acts as a generator and pumps energy back into the DC bus, raising its voltage. On a light demo load the drive absorbs this in its capacitors and the decel ramp just works. On high-inertia loads (flywheels, large fans) a fast stop can push the bus voltage past the overvoltage trip — the classic "trips every time I press stop" fault.
The fixes, in order:
- Lengthen the deceleration ramp. Give the losses time to absorb the energy. Free and usually sufficient for student rigs.
- DC injection braking (a drive parameter): at near-zero speed the drive injects DC into the windings, holding the rotor. Useful for stopping, not for dissipating large energies.
- Braking resistor: the proper answer for repeated or fast stops of inertial loads. The drive switches the resistor across the DC bus (via a built-in or external braking chopper) and burns the regenerated energy as heat. Size per the drive manual's resistor tables — resistance too low destroys the chopper, too high does nothing. Mount the resistor where its heat cannot cook nearby wiring.
For a first student build, design the demo so the load inertia is modest and option 1 suffices; add the resistor only if the application genuinely needs fast stops.
Derating: the fine print that bites in Indian labs
Drives are rated at 40 °C ambient and sea level. A student panel in a non-air-conditioned lab in May sees 45 °C inside the enclosure — above rating. The honest responses: ventilate the enclosure (filtered vents low, exhaust fan high), keep the drive away from the braking resistor's heat, and if the manual gives derating curves, apply them rather than hoping. Similarly, raising the carrier frequency for quieter operation increases the drive's internal losses — the manual's derating guidance for high carrier frequencies is not decorative. None of this matters for a 10-minute demo; all of it matters for a rig that runs all day during project exhibition week.
Three demo-rig designs that work
If you need a concrete project shape around a VFD:
- Conveyor speed control with preset speeds. Belt conveyor, 3–4 digital inputs selecting preset frequencies (10/25/40/50 Hz), direction reversal, E-stop. Demonstrates parameter setup, multi-speed logic and wiring discipline. Pairs naturally with the crane control panel build for panel-layout ideas.
- Pump energy-saving demonstrator. Small centrifugal pump, two pressure gauges, an energy meter on the drive input. Run at 100/80/60% speed, log power, and plot the measured points against the theoretical cubic curve from the affinity laws. The gap between theory and measurement — and explaining it — is a strong discussion section.
- Multi-motor sequential control. Two motors started in sequence with interlocking (the DOL/star-delta world meets the VFD world), showing when each starting method is appropriate. The sequential starter panel project covers the contactor-logic side.
The short version
Motor speed follows frequency (Ns = 120f/P), and a VFD varies frequency while holding the V/f ratio constant so torque stays available at any speed. Size the drive to the motor's current, use a single-phase-input drive with the motor in delta when the lab has no three-phase, enter the nameplate data before touching anything else, set sane accel/decel ramps, respect the charged DC bus after power-off, and never switch the output while running. Do that, and the drive becomes what it is in industry: the most useful box in the panel.