VFD Selection and Sizing Guide

---

The single most important concept in VFD sizing is the torque profile of the load. Get this wrong and every other decision downstream is built on sand.

Variable torque loads include centrifugal fans, centrifugal pumps, centrifugal blowers, and centrifugal compressors. They follow the affinity laws: the torque a centrifugal load requires increases with the square of speed, and the power required increases with the cube. At half speed, a centrifugal pump requires only one-eighth the power it uses at full speed. This is the math that makes VFDs so attractive on HVAC and water applications. It is also why VT loads need less overload capacity from the drive: a fan accelerating from rest is pushing against almost no air resistance, so starting current is minimal.

Constant torque loads are everything else. Conveyors, mixers, extruders, hoists, elevators, positive displacement pumps, reciprocating compressors. These loads need approximately the same torque at low speed as at high speed. A conveyor carrying 10 tons of material requires the same force to move that material whether the belt runs at half speed or full speed. Power consumption scales linearly with speed instead of cubically, so the energy savings are smaller. More important for sizing: breakaway torque on a CT load can hit 150% to 200% of running torque, because the drive must overcome static friction, material weight, and rotational inertia just to get the load moving.

The torque profile determines three things that drive the rest of the selection: the required overload rating, the control mode, and the thermal capacity. A VT-rated drive provides roughly 110% overload for 60 seconds. A CT-rated drive provides roughly 150% for 60 seconds plus 180% breakaway torque for 0.5 seconds. Getting the torque profile wrong is the most common sizing mistake and the one with the most downstream consequences.

One practical note: a centrifugal pump is a variable torque load, but a positive displacement pump is constant torque. If the application says "pump" without specifying the type, ask. The wrong assumption here has caused more than one installation to fail.

VT sizing looks easy, and it often is. A centrifugal fan or pump on a drive sized to the motor FLA in VT mode will usually run. But "will run" is a low bar, and a few recurring mistakes still trip up specifiers on this supposedly easy case.

Because centrifugal loads need almost no starting torque, VT-rated drives give you a higher horsepower rating per frame size. A drive that is rated 200 HP at CT duty is typically rated 250 HP at VT duty on the same hardware. This is not a trick. The drive's overload capability is lower in VT mode (110% for 60 seconds instead of 150%), which means the drive can continuously deliver a higher average current without exceeding its thermal limits. For a fan or pump that will never demand more than 110% of nameplate, the VT rating gives you more motor per dollar.

The most common VT sizing mistake is oversizing for startup surge when the application does not actually need it. Centrifugal loads simply do not surge at startup. If someone is pushing to upsize a fan drive "because it needs to handle the startup," the answer is almost always that a correctly selected VT-rated drive will handle the startup fine. The exception is a fan with a very heavy wheel (high inertia) that cannot be accelerated in a reasonable time. That is a separate consideration addressed by ramp time, not drive sizing.

Another common mistake is assuming affinity law savings apply to every pump. They do not. A positive displacement pump (gear pump, lobe pump, progressing cavity pump) is a constant torque load, not a variable torque load. Slowing it down saves roughly the linear amount of power, not the cubic amount. If someone is projecting 87% energy savings at half speed on a PD pump, the math is wrong.

Finally, VT sizing still needs real-world derating. A centrifugal pump drive installed at 2,000 meters altitude, in a 45-degree Celsius pump room, with a 300-foot cable run, and an 8 kHz carrier frequency for noise reduction, will still exceed the capacity of a drive that was sized only against the motor FLA. The affinity laws reduce the drive's current demand at partial load; they do not reduce the derating factors that apply at the top of the speed range.

This is where sizing mistakes happen most often, and where they do the most damage.

Consider a conveyor that refused to start when the drive was set to Variable Torque mode. The drive was sized correctly for the motor FLA. The installation was clean. The ramp time was adjusted. None of it mattered. A loaded conveyor needs 150% to 160% of rated current just to break the belt loose, and VT mode caps the drive at 110%. The drive was never going to move that belt until somebody changed the load type parameter to Constant Torque, which took thirty seconds once the diagnosis was correct. This VT/CT mismatch is one of the most common first-week faults, covered in detail in our VFD Troubleshooting Guide, as a classic case of the drive doing exactly what it was configured to do.

The lesson for selection: you cannot rescue a constant torque application by setting a VT-rated drive to CT mode. The drive must be specified and sized for CT duty from the beginning. A drive nameplate with dual ratings (for example, 200 HP at CT and 250 HP at VT) tells you what the drive can actually deliver in each mode. If you buy the 200 HP VT rating and try to run a 200 HP conveyor, you are pushing the drive to deliver 250 HP of CT performance, and it will not last.

CT applications also demand more acceleration margin. A loaded conveyor, a hoist lifting a heavy pick, a mixer starting in cold product: all of these demand current well above the steady-state running value during the acceleration period. The ORCA-Size methodology (covered in the next section) treats this explicitly through an acceleration demand factor that adds 15% to 20% to the required current rating for high-inertia starts. Traditional sizing ignores this factor entirely, which is why conveyor drives sized by FLA-match frequently trip on OL2 faults in the first week.

One more CT consideration worth flagging: hoists and cranes are constant torque in both directions, and they generate regenerative energy continuously when lowering loads. The drive selection must account for continuous braking duty, not just occasional deceleration events. This shifts the accessory specification toward regenerative (active front end) drives or toward heavy-duty braking resistors rated for the full lowering cycle. A crane drive sized without regenerative duty in mind will trip on overvoltage every time an operator lowers a load.

Traditional sizing answers one question: is the drive's nameplate current bigger than the motor's nameplate current? That question is inadequate because it ignores the conditions where the drive actually lives and works. ORCA-Size answers a different question: is the drive's effective capacity, after accounting for every condition that reduces it, still bigger than what the motor and load actually demand?

ORCA-Size stands for Operating conditions, Required torque profile, Current and voltage ratings, and Application-specific derating factors. It organizes the factors that traditional sizing ignores into two groups: installation factors (cable length, ambient temperature, altitude, carrier frequency) and application factors (load class, acceleration demand, low-speed turndown, overspeed). You look up a factor for each condition that applies, multiply those factors together, and multiply the result by the motor's FLA. The answer tells you the minimum continuous current rating your drive needs.

The power of the approach is that these factors compound multiplicatively, not additively. A drive at 50 degrees Celsius AND 2,000 meters altitude AND 200-foot cables experiences the product of these deratings, not their sum. Traditional sizing treats each factor as independent or ignores most of them. The Denver conveyor from the opening of this guide is the canonical example. Traditional sizing called for a 77-amp drive. ORCA-Size called for a 200-amp frame. The 77-amp drive was running at more than double its effective capacity, which is exactly why it kept tripping.

The same principle catches less obvious mistakes. A crane hoist with its carrier frequency bumped up from 4 kHz to 8 kHz for noise reduction just lost 20% of its thermal headroom. A booster pump running at 75 Hz for process reasons is drawing roughly 33% more current than its 60 Hz nameplate suggests. These decisions are invisible on a traditional sizing spreadsheet and catastrophic to the drive six months later. ORCA-Size makes them visible.

The methodology was developed as part of the broader ORCA (Optimized Reliability and Condition Assessment) framework. It is detailed in a peer-reviewed (pending) paper titled ORCA-Size: A Reliability-Focused Alternative to Traditional Variable Frequency Drive Sizing, published on TechRxiv, including the Monte Carlo validation across 3,600 trials and the alignment with published manufacturer derating curves from ABB, Siemens, Rockwell, and Danfoss.

The full worksheet and practical implementation are in Before the First Fault, Chapter 3, along with five worked examples covering cooling tower fans, inclined conveyors, crane hoists, overspeed pumps, and multi-drive installations. What matters for the selection decision is this: an ORCA-Size drive arrives with genuine thermal margin (15% to 30% after all real-world factors are applied), while a traditionally sized drive often arrives operating at 95% of its thermal capacity with no headroom for anything. Our Maintenance and Reliability Guide covers how that margin translates into years of additional service life.

The drive and the motor are a system, not independent components. A drive sized correctly for an incompatible motor will still produce problems. A drive paired with the right motor can deliver the service life the nameplates promise.

Three factors dominate motor selection for VFD duty: insulation voltage withstand, bearing protection, and cooling at low speed.

PWM output stresses motor insulation in ways that line power never does. IGBT switching transitions happen in 0.1 to 0.5 microseconds, roughly 1,000 to 10,000 times faster than utility voltage changes. Combined with cable impedance, those fast transitions create reflected waves that can double the voltage at the motor terminals. On a 460V system, peak motor terminal voltages of 1,200 to 1,600 volts are common on cable runs between 100 and 300 feet.

NEMA MG1 Part 31 defines inverter-rated motor insulation. The standard requires the insulation to withstand peak voltages of 3.1 times the motor's rated voltage at 0.1 microsecond rise times. For a 460V motor, that is approximately 1,430 volts peak. Standard motors designed only for line-frequency stress may have insulation rated for 1,000 volts peak, which is inadequate for VFD duty on anything but short cable runs.

The cost premium for an inverter-rated motor is typically 10% to 20% over a standard motor of the same HP and frame. On a 30 HP motor, that is $200 to $400. Against the cost of an unplanned motor failure, the emergency labor to replace it, and the lost production while the replacement is installed, it is the cheapest insurance in the building. For new installations, specify inverter-rated motors by default.

PWM output also creates common-mode voltages that can discharge through motor bearings, causing electrical discharge machining (EDM) damage. Field data suggests 50% or more of VFD-driven motor systems experience some bearing damage when no mitigation is installed. The early stages are invisible. By the time bearing noise is audible, the damage is well advanced.

Five mitigation approaches in rough order of effectiveness: properly specified and terminated VFD cable (covered in our VFD Installation Guide) provides the low-impedance common-mode return path that prevents most bearing currents at the source. Insulated bearings (ceramic-coated outer race) break the discharge path on the motor side. Shaft grounding rings provide a low-impedance path from shaft to frame for $30 to $150 per motor. Ceramic rolling element bearings offer complete electrical isolation. Common-mode filters on the drive output attenuate the source voltage.

For motors above 50 HP, specify insulated non-drive-end bearings as a baseline. For critical applications or long cable runs, add a shaft grounding ring.

A TEFC motor cools itself with a fan on its own shaft. When the VFD slows the motor down, the cooling fan slows too. At one-sixth of base speed, you have one-sixth of the cooling air, but the motor still generates heat from magnetizing current (roughly 30% of FLA) and load current. Sustained constant-torque operation below 20 to 30 Hz without supplemental cooling will overheat a TEFC motor.

The fix is straightforward. For continuous low-speed operation on CT loads, specify a TEBC (blower-cooled) motor whose cooling fan runs at full speed independent of shaft speed. Or oversize the motor frame so thermal mass and surface area compensate. Or specify an inverter-rated motor with a high speed range rating (100:1 or 1000:1), which is designed for the low-speed thermal challenge.

For variable torque loads, this is less of an issue: the load itself drops with speed, so motor heat generation decreases proportionally with cooling capacity.

Cable and accessories are selection decisions, not installation afterthoughts. What you specify before the drive arrives determines what the installer has to work with, and specification shortcuts here show up as intermittent field problems years later.

Use VFD cable. The cost difference over THHN is small. The performance difference is significant. VFD cable has thicker insulation rated for PWM voltage stress, symmetrical ground conductors for low-impedance common-mode return, and integral shielding for EMI containment. ANSI/ICEA S-138-738-2024 establishes the minimum construction standard for VFD cable in the United States.

If budget will not support VFD cable, specify XHHW-2 as the minimum acceptable alternative, not THHN. THHN's PVC insulation has poor corona resistance, high capacitance (up to 10 times the capacitance of VFD cable per foot), and hygroscopic moisture absorption that degrades voltage withstand over time. ABB frames it plainly: THHN is tolerable, XHHW-2 is better, VFD cable is best. Do not call THHN "good."

Motor terminal voltage climbs with cable length. A simple decision rule based on one-way cable length: no filter required under 50 feet on inverter-rated motors; output reactor for 50 to 250 feet; dV/dt filter for 250 to 500 feet; sine wave filter above 500 feet. These are conservative field practice thresholds; newer drives with improved IGBT gate control can sometimes extend them modestly. For multi-motor installations where one drive feeds several motors, the cable length for filter selection is the total additive length of all motor leads, not the distance to the nearest motor.

A 3% to 5% input line reactor is the highest-value accessory in the VFD specification. At $50 to $200 for most drive sizes, it delivers four distinct benefits: harmonic current reduction (from 80% to 100% THDi down to 25% to 45%), surge and transient protection, phase balance improvement, and extended DC bus capacitor life through reduced ripple current. The trade-off is a proportional voltage drop at full load (roughly 3% or 5%), which matters only at the top of the motor's speed range.

For the overwhelming majority of installations, specify the line reactor. The argument for omitting it is almost always a budget decision, and the budget is almost always wrong.

Every VFD needs a disconnect means by code. The question is what kind.

VFD input fusing has two jobs: protecting the branch circuit wiring from overcurrent, and protecting the drive's input components from short circuit fault current. The fuse type and rating must follow the drive manufacturer's recommendation, which is typically semiconductor-rated fusing (Class J, Class CC, or Class RK) sized for the drive's SCCR (Short Circuit Current Rating). Generic motor circuit breakers without supplemental fusing may not clear fast enough to protect the drive's rectifier diodes under short circuit conditions.

Before you specify the protective device, calculate the available fault current at the installation point. A 1,500 kVA transformer with 2% impedance can deliver roughly 90,000 amps of fault current at its secondary terminals. A 15 kVA isolation transformer with 5.5% impedance limits fault current to 327 amps at the same voltage. The difference determines whether you need high-interrupt-capacity Class J fuses or whether a standard circuit breaker is adequate.

The panel's overall SCCR is determined by the weakest component in the system. A 65 kA breaker protecting a drive with 10 kA SCCR gives you a 10 kA panel, not a 65 kA panel. NEC Article 409.110 requires the SCCR to be marked on industrial control panels, and underestimating it can have catastrophic consequences. This is code compliance and safety, not an optional specification detail.

Selection is the foundation. Every chapter of the VFD lifecycle that follows (installation, commissioning, troubleshooting, maintenance) is easier when the drive was sized correctly for the actual application, the motor was matched to the drive, and the cable and accessories were specified to support both. Once the correctly sized drive is installed, the commissioning process takes the installation from energized hardware to verified operation. When the baseline is recorded against a drive that has real thermal margin, troubleshooting becomes diagnosis instead of guesswork.

The Denver conveyor from the opening is still running today, once it was replaced with the 200-amp frame that ORCA-Size identified. The motor survived. The plant stopped losing production to resets. The lesson was expensive but clean: size the drive for the conditions where it will actually live, not the conditions on the datasheet.

Before the First Fault: A Field Guide to VFD Installation and Reliability covers the complete ORCA-Size methodology in Chapter 3 (with worksheet, five worked examples, and Monte Carlo validation), motor selection for VFD duty in Chapter 4, and cable and accessory specification in Chapter 5. The companion online training program walks through sizing decisions with video demonstrations and downloadable worksheets.

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 you select determines the installation you get. Specify for the conditions where it will actually live, and the next five to fifteen years take care of themselves.