VFD Commissioning Guide

---

Before you close the disconnect, you walk the panel one more time.

This is the last chance to catch installation errors before they become fault codes. A loose power terminal that could have been caught with a torque wrench becomes an arcing fault. A misrouted cable becomes a noise problem that takes days to diagnose. A shield strand touching the T2 conductor inside a motor junction box becomes a ground fault on first power-up, a day of lost production, and an unpaid return trip for the contractor who skipped the checklist.

Twenty minutes with a multimeter, a torque wrench, and a megger prevents twenty thousand dollars in surprises. Every time.

A comprehensive pre-energization check moves through seven phases. First, verify environmental conditions: ambient temperature at the drive, altitude and derating, humidity, and contamination level. Second, confirm mechanical installation: mounting orientation, drive spacing, fastener torque, and feedback device attachment. Third, work through the electrical checks covering AC supply verification (voltage, imbalance below 3%, phase rotation), grounding and bonding (paint scraped at every connection, less than 0.1 ohms at the PE terminal, direct home-run conductors), cable and shield verification (input on L1/L2/L3, motor on T1/T2/T3, shield terminations at both ends for motor cable), and surge suppression on every coil.

Fourth, megger the motor and cable. With the drive disconnected from the motor, test T1 to ground, T2 to ground, and T3 to ground with 1,000 VDC for 480V systems. All three readings should be above 1 megohm. Anything lower means either a cable problem or a motor insulation problem, and you find it now, not after you apply power and the drive trips on ground fault.

Fifth, prepare the motor. Confirm the coupling state, because that determines which type of motor identification run you will perform. Sixth, walk the complete installation one last time. Confirm the panel is clean, doors close without pinching cables, and stakeholders have signed off. Seventh, sign and date the checklist.

This connects directly to the installation practices in the VFD Installation Guide. If your installation was done right, the pre-energization check confirms good work. If corners were cut, this is your last clean catch point.

The drive has never seen power before. The DC bus capacitors are completely discharged. This is not a normal start.

Close the disconnect. Most drives will display a startup screen on their HMI, followed by a status screen showing the ready state. Inside the drive, the soft-charge circuit is doing its job: a current-limiting resistor restricts inrush current while the DC bus capacitors charge, and after a few seconds, a relay closes to bypass the resistor for normal operation. You will hear a distinct click when that relay closes. That click is the DC bus reaching its operating voltage, and it means the drive is transitioning to ready.

The whole process takes a few seconds to perhaps thirty seconds depending on drive size. During this window, the drive may show a "charging" status. Do not attempt to operate the drive until the charging sequence is complete. This is also the reason drives are limited to approximately two starts per minute. The soft-charge resistors need time to cool between power cycles.

If the drive does not power up, or if it immediately displays a fault code, do not repeatedly cycle power. Read the fault, consult the manual, and identify the cause before re-energizing. The three faults you will see most often at first power-up are input phase loss (one phase of the three-phase supply is missing or miswired), ground fault (a wiring error in the output or an insulation problem the megger test missed), and overvoltage or undervoltage (input voltage outside the drive's acceptable range).

If the drive powers up with no faults, verify the input. Measure the three line-to-line voltages at the drive's input terminals with a true-RMS multimeter. L1 to L2, L2 to L3, L3 to L1. Record all three. Most drives accept plus or minus 10% of nominal, which means 432 to 528 VAC for a 480V drive.

Check voltage balance. Sum the three readings, divide by three for the average, and find the maximum deviation from that average. Maximum deviation divided by average, expressed as a percentage, is voltage imbalance. Anything above 2% to 3% is a power system problem that should be investigated before proceeding. Voltage imbalance causes current imbalance in the input rectifier, which causes heating, which causes nuisance trips and capacitor aging that you pay for years later.

Verify the DC bus voltage on the drive's display. For a 480 VAC input, expect approximately 650 VDC (1.35 times line-to-line). For a 240 VAC input, expect approximately 325 VDC. Significant deviation means something is wrong at the input or the rectifier.

The drive is not running the motor yet. You have just confirmed the drive itself is healthy.

Every VFD needs to know three things before it can control a motor properly: what the motor is, how you want it controlled, and how fast you want it to change speed. That is the heart of parameter configuration, and it applies across every major manufacturer. The specific parameter numbers come from the product manual. The reasoning behind what to set and why is what this guide is about.

The single most consequential entry is the motor nameplate data: horsepower (or kW), rated voltage, full-load amps at the actual voltage configuration, rated speed in RPM, frequency, and insulation class. Every value must be entered correctly. Errors in nameplate data degrade performance and protection for the life of the installation. If the motor is dual-rated at 230/460 VAC and you wire it for 460, you enter the 460V FLA, not the 230V FLA. The conversion factor between HP and kW is 0.75, so a 20 HP motor is entered as 15 kW.

Rated speed tells the drive two things at once. It tells you the pole count (a nameplate reading 1770 RPM is a 4-pole motor with 1800 RPM synchronous speed) and it tells you the slip at rated load.

After nameplate data is entered, run the motor ID Run. The drive measures the actual electrical characteristics of your specific motor and builds an internal model used for control. Nameplate data tells the drive what the motor should look like based on design. The ID Run tells it what the motor actually looks like based on measurements.

The difference matters. Manufacturing tolerances, aging, rewinding, and the specific cable between drive and motor all affect electrical characteristics. Two motors with identical nameplates can have measurably different stator resistance and magnetizing current. For sensorless vector control, an inaccurate model produces poor speed regulation, torque ripple, and reduced efficiency.

A Normal ID Run (rotational autotune) requires the motor decoupled from the load, takes 60 to 90 seconds, and produces the most accurate model. A Reduced ID Run (stationary autotune) measures stator parameters with the motor still connected and is used when decoupling is impractical. Any ID Run is better than none.

Three control modes cover roughly 95% of VFD applications. V/Hz control maintains a constant ratio between voltage and frequency and is the right choice for fans, pumps, and other variable-torque loads where plus or minus 2% to 3% speed accuracy is acceptable. V/Hz is also the only option when one drive controls multiple motors simultaneously.

Sensorless vector control uses the motor model from the ID Run to estimate rotor position and flux without an encoder. It gives you plus or minus 0.5% to 1% speed accuracy, faster dynamic response, and better low-speed torque than V/Hz. It is the right choice for conveyors, mixers, extruders, and single-motor pump applications where tighter speed control matters.

Closed-loop vector control adds an encoder or resolver for direct shaft feedback. It delivers plus or minus 0.01% speed accuracy, full torque at zero speed, and the fastest dynamic response. Reserve it for winders, hoists, positioning applications, and servo-type performance requirements.

When in doubt, start with V/Hz. It is simple, forgiving, and robust. Upgrade to sensorless vector only if V/Hz does not meet the application requirements. Complexity without need is a liability.

The default ramp times in most drives are 5 to 15 seconds, and for many applications those defaults are far too aggressive. Large fans (100 HP and above) should start at 90 seconds or more of acceleration time. The rotational inertia of a big fan wheel demands more current than the drive can sustain if you try to ramp it up in 10 seconds, and you will trip on overcurrent during acceleration every time until someone extends the ramp.

Configure the motor thermal protection to match the actual motor, not the drive's maximum output. If a 30 HP drive controls a 20 HP motor (the right sizing, per our Installation Guide), the electronic thermal overload must be set to the motor's FLA of 25 amps at 460 VAC, not the drive's rated output current of 40 amps. The drive's thermal model uses an I-squared-t calculation to estimate the motor's thermal state, and it needs the motor's actual FLA and thermal class to do that correctly.

The first time the motor turns is a controlled experiment, not a production run.

If you can decouple the motor from the load, do it. The uncoupled test lets you verify rotation direction without risking the driven equipment and gives you a clean no-load baseline for current. Set the speed reference to 5 Hz and command the drive to run from the local keypad. On a 4-pole motor, 5 Hz produces about 150 RPM, slow enough to see rotation direction and stop quickly if something is wrong.

Is the shaft turning the right way? Most centrifugal pumps and fans have a rotation arrow on the housing. If the motor is turning the wrong direction, stop immediately. During commissioning you can use the drive's direction parameter to electronically reverse the phase sequence, but for the permanent installation the correct method is to physically swap two of the three motor leads at the drive's output terminals. The reason is operational safety. If the drive is ever reset to factory defaults during a firmware update, a parameter reset, or a drive replacement, the direction parameter reverts to its default value and the motor suddenly runs the wrong way without warning. Swapping the physical wires makes correct rotation survive any future configuration change.

With rotation direction confirmed, listen to the motor. An uncoupled motor at low speed should be quiet. A high-pitched whine is normal (caused by the PWM switching frequency exciting the motor laminations). A grinding noise is not. Check the current on the drive's display. An uncoupled motor at minimum speed should draw only its magnetizing current, typically 25% to 40% of FLA. Significantly higher current means either a winding problem or an error in nameplate data.

Gradually increase the speed to 30 Hz, then 60 Hz. Watch for abnormal vibration or noise at any intermediate speed. Mechanical resonance points are real, they are application-specific, and you find them by ramping through the speed range once and noting where things get louder or rougher than expected. Record the no-load current at full speed. Stop the drive. If the motor coasts instead of decelerating under drive control, check the stop mode parameter. If the drive trips on DC bus overvoltage during deceleration, the ramp is too short for the load's inertia and you need either a longer deceleration time or a braking resistor.

Now couple the motor to the load. Lock out the drive disconnect before anyone puts hands near the coupling. Verify alignment with a laser alignment tool where possible, or with dial indicators as a minimum. Reinstall the coupling guard. Remove the LOTO.

Run the coupled system at 5 Hz and observe. For pumps, incorrect rotation will be obvious from the low motor current and lack of flow. For conveyors, you will see it in the belt direction. For fans, look at airflow at a register or inlet. Ramp through the speed range again under load, watching current, voltage, and bearing temperatures at intermediate points.

At full speed under normal load, record your commissioning measurements: output voltages (T1 to T2, T2 to T3, T3 to T1), output currents on each phase, DC bus voltage, drive heatsink temperature, and motor speed. Compare current to motor FLA. If the motor is running at 85% of FLA, you have 15% thermal margin. If it is running at 98%, you have almost none, and any rise in load, ambient, or voltage imbalance will trip the thermal overload.

Document every adjustment, every observation, every anomaly and how it was resolved. The load testing record is part of the permanent commissioning package.

A 200 HP cooling tower fan was spinning at full speed one Monday morning when a maintenance technician opened the fan housing access door to inspect the belt tension. The door had a safety interlock switch wired to the drive's Safe Torque Off input. When the door opened, STO was supposed to remove the drive's output and the fan was supposed to coast to a stop.

The fan kept spinning.

The STO circuit had an open wire at a splice junction inside the conduit. The interlock switch signal never reached the drive. The wire was landed at both ends during installation, the box was checked on the install sheet, but nobody had ever opened the door while the fan was running to verify the function actually worked. The technician was experienced enough to notice the fan was still running before reaching into the housing. Nobody was hurt.

The gap between "the wire is connected" and "the safety function actually works" is exactly the gap this step of commissioning exists to close. A visual inspection can verify that a wire is landed on the correct terminal. It cannot verify that the switch closes when the door opens, that the wire has continuity from end to end, that the drive's safety logic responds correctly to the signal, or that the motor actually stops when the function activates. The only way to verify a safety function is to activate it under controlled conditions and observe the result.

Test Safe Torque Off first. Run the motor at roughly 30 Hz under load. Activate STO by the method that will be used in normal operation: open the interlock door, press the E-stop, or trigger the safety relay. The drive should immediately stop producing output and the motor should begin to coast. The display should show STO status. Try to restart while STO is active. The drive must refuse to start. Release STO and confirm the drive returns to ready state.

Test every E-stop button, not just one. IEC 60204-1 defines three stop categories: Category 0 removes power immediately and the motor coasts; Category 1 decelerates the motor along a ramp and then removes power; Category 2 decelerates to zero and maintains holding torque. Know which category your system is designed for before you test.

Test every interlock. Access door interlocks, flow interlocks on pumps, pressure interlocks, temperature interlocks, level interlocks. Pay particular attention to the logic direction. Some interlocks are normally closed during normal operation and open to create the condition; others are normally open and close to create it. The wrong logic configuration produces a safety function that works backward, where the drive runs when the door is open and stops when the door is closed. This has caused injuries.

Test fault response behavior. Verify the motor thermal overload trips at the correct current. Confirm the drive responds to simulated fault conditions the way the application requires.

Document every safety test. Record the activation method, the drive response, the motor behavior, whether the drive rejected a restart attempt while the safety function was active, and the normal recovery sequence. In a facility audit or an incident investigation, this documentation is the difference between defensible practice and a problem.

Commissioning is finished. The drive is running, the motor is turning, the process is doing what it is supposed to do. You have one more step, and if you skip it you will regret it later.

Let the system run under normal production conditions for 24 to 48 hours. Thermal equilibrium takes time. Vibration signatures stabilize as bearings seat. Production patterns reveal themselves. After that window, record the operational baseline.

The baseline captures seven categories of measurements at a known steady-state operating point. Electrical measurements (output currents on each phase, output voltages, DC bus voltage, output frequency, output power, current utilization as a percentage of drive rating) establish the electrical fingerprint. Thermal measurements (inverter temperature, ambient at the drive, motor frame temperature, drive-end and non-drive-end bearing temperatures) establish the thermal fingerprint. A thermal image of the drive panel interior, the motor, the junction box, and the input power connections gives you a visual reference that makes future hot-spot problems immediately obvious.

Vibration measurements at both motor bearings in three axes (horizontal, vertical, axial) establish the mechanical baseline. For a new, properly aligned motor, overall vibration should be below 0.15 inches per second per ISO 10816 guidelines. Process performance measurements capture what the system actually delivers. Drive internal diagnostics (run hours, thermal utilization, DC bus stability) complete the picture. A full parameter backup, saved to at least two locations, makes the configuration recoverable if anything goes wrong later.

Field data consistently shows that temperature and current dominate all degradation signatures. A partially closed discharge valve increases motor current. A dusty heatsink raises inverter temperature. A loose input power connection creates a hot spot that shows up in a thermal image long before it becomes a fault. Without a baseline, a technician looking at a 62-degree inverter temperature has no idea whether that is normal. With a baseline that says the inverter was 44 degrees at commissioning, the same technician immediately identifies an 18-degree rise, checks current, finds the closed valve, and resolves the problem in twenty minutes instead of twenty hours.

Store the baseline record in three places: inside the drive panel door for immediate access during troubleshooting, in the facility CMMS for long-term archival, and with the commissioning package. We cover how to use these measurements for ongoing condition assessment in depth in our Maintenance and Reliability Guide.

Commissioning done right takes time. You will spend one to three hours on load testing alone, another hour on safety verification, and another two to four hours stabilizing the baseline. Call it a full day for a single drive, and more for a multi-drive panel.

That day prevents the weeks of troubleshooting, resets, and 2 a.m. phone calls that follow a rushed startup. Every measurement becomes the reference for every future maintenance decision. Every parameter you document saves the next technician from having to reverse-engineer why something was changed. Every safety function you test is one less surprise when an operator opens a guard door.

The commissioning record, the load testing record, the safety verification record, and the operational baseline together become the installation's birth certificate. If a drive fails two years from now and needs replacement, those documents let the next technician bring the system back to its known-good state in hours instead of days.

Before the First Fault: A Field Guide to VFD Installation and Reliability covers the complete commissioning sequence across Chapters 13 through 17, including the full checklists, worked examples, and diagnostic techniques practitioners rely on during startup. The companion online training program walks through each step with video demonstrations and downloadable reference forms.

Get the book: Before the First Fault

Enroll in the course: Before the First Fault training program

Contact the author: Dr. Carl Lee Tolbert, PhD, CMRP Wayward Leaders LLC www.waywardleaders.com

The drive does not know whether you commissioned it correctly. But six months from now, when the process is running smoothly or the phone is ringing at 2 a.m., the drive will have already told you.