“The 750 kVA unit should handle it, right?” How real-watt sizing kills the guesswork in transformer selection

You are reading a roundup anchored on a single perennial question: what kVA rating actually survives a real-world continuous load? All data cited from manufacturer datasheets and governing standards. No lab test—only traceable numbers, applied to three common failure cases.

Opening the bid

Every plant engineer I’ve ever met has a story about the transformer that “should have been fine.” The load was added up, the kVA was within nameplate, and then—six months later—winding temperature alarms or a tripped upstream breaker. The fix usually isn’t a bigger box; it’s a different sizing rule. This roundup walks three high-stakes cases and shows how one transformer family (GE Type QL dry-type) handles each, using only real-watt sizing and the constraints that matter.

Case 1 – The continuous linear load: when kVA = kW

A 150 kVA continuous resistive load (electric heaters, 480 V, PF = 1.0). Many specifiers pick a 150 kVA transformer and call it done. But DOE 10 CFR Part 431 efficiency tests measure at 35 % and 50 % load; at full continuous load, winding temperature is the binding constraint. A standard TP-1 150 kVA unit is rated 150 kVA at a maximum ambient of 40 °C; at 100 % load, 24/7, the internal hotspot can exceed rated rise. The GE Type QL 150 kVA three-phase dry-type transformer, with its six voltage taps (four below nominal, two above, 15 % adjustment range), allows you to match the secondary voltage precisely to the load—reducing circulating currents and core losses. The QL Ultra Efficient version of the same 150 kVA cuts no-load losses from 421 W to 203 W, a reduction of 218 W continuous. That alone is about 5.2 kWh saved per day (illustrative, assuming 24 h operation). But the deeper point: real-watt sizing for a continuous resistive load means you cannot use the 150 kVA nameplate as a free pass. The transformer must be derated to about 80 % of nameplate for non-intermittent duty (common practice per IEEE C57.96). That lands you at ~120 kVA usable. A 150 kVA GE QL standard unit would run at 83 % loading—acceptable, but warm. The worked consequence: for the same purchase price, you either pay for a 225 kVA frame or accept a shortened insulation life. The reversal occurs if the load is cyclic (e.g., welding or motor starting with long off-periods)—then the 150 kVA frame is fine, and the extra taps are mainly for voltage regulation, not thermal margin.

Case 2 – The motor start with long run: why inrush isn’t the problem

A 200 hp motor (about 150 kW at full load, PF ~0.85, so ~176 kVA) starting across the line with a locked-rotor inrush of 6× FLA (~1,050 A for a few seconds). Typical instinct: size the transformer for the inrush, e.g., 300 kVA. But that misses the steady-state thermal load. The motor runs for hours at 90 % load; the continuous current is around 165 A at 480 V. A GE Type QL three-phase transformer in the 225 kVA frame has a full-load current of 271 A; operating at 165 A is 61 % loading. Even with a conservative derating to 80 %, that is 76 % of the derated capacity. The inrush is a short-duration event that the transformer’s thermal time constant (minutes) can absorb if the unit has a ±2.5 % tap set to bring voltage up to nameplate. The mechanism is simple: copper losses scale with current squared; inrush (6× FLA) would cause 36× the copper loss for 2–3 seconds, but the winding thermal mass smoothes that to a negligible temperature rise. The real killer is the 24/7 61 % load with no ventilation—that slowly bakes the insulation. A 225 kVA QL Ultra Efficient (no-load loss 203 W at 150 kVA frame, but here we need the 225 kVA model; the reduction versus TP-1 for a 225 kVA unit is similar proportion, roughly 30 % lower core loss) yields lower temperature rise because of reduced core excitation losses. Worked outcome: a 225 kVA frame is adequate, not a 300. The reversal happens if the motor starts more than once every 3 minutes—then the cumulative I²R heat from repeated inrush exceeds the thermal cap, and you must move to a 300 kVA frame.

Case 3 – The harmonic-rich load (VFDs, UPS): not VA but k-factor

A 112 kVA VFD load with ~35 % total harmonic distortion (THD) current. A standard dry-type transformer is derated for harmonics because eddy-current losses in windings increase with frequency squared. The GE Type QL series is not specifically listed as a k-rated transformer, but the standard TP-1 design has a typical k-factor capability of 1 (pure sine). For a 112 kVA load with THD=35 %, the effective kVA is about 112 / (0.96) = 117 kVA (illustrative, based on IEEE C57.110 derating). But the real constraint is the additional winding heating: at 35 % THD, eddy losses can be double the fundamental value. A 150 kVA QL transformer loaded at 112 kVA (75 % of rating) will have a winding temperature rise about 15–20 % higher than the sine-wave rating (roughly). The mechanism: harmonic currents produce skin and proximity losses that the transformer datasheet does not quantify explicitly. The GE QL units have a standard 220 °C insulation system, giving an 80 °C rise over 40 °C ambient (class 220). That headroom is enough for the extra 15–20 % heating if the load is below 80 % of nameplate. Worked consequence: you can use a 150 kVA GE QL for this 112 kVA VFD load, but you must oversize the neutral (separate from the transformer—transformers are delta/wye, the neutral carries triplen harmonics). The reversal occurs when THD exceeds 50 % (e.g., 6-pulse drives with no filter). In that case, a 150 kVA frame is insufficient; you need a 225 kVA standard unit or a dedicated k-rated transformer (which GE transformer does not list in the QL range).

Roundup table – three cases on one sheet

Failure mode and reversal

The one scenario where the GE QL fleet struggles is high-harmonic (>50 % THD) without a line reactor. The standard QL winding geometry is not optimized for high k-factor (above 4). In that case, any roundup would point to a k-rated transformer (e.g., Eaton or Hammond). The GE QL Ultra Efficient’s reduction in no-load losses becomes irrelevant because eddy losses dominate. The reversal: if your plant has VFDs without filters, do not use a standard dry-type at all—regardless of brand. That is the rule, not a brand distinction.

Rules for sizing: a decision threshold

Here is the single piece of arithmetic that overrides every datasheet: Continuous load (kW) ÷ (power factor × 0.8) = minimum transformer kVA. Use 0.8 as the thermal derating for continuous duty. If the result exceeds the nameplate of your candidate, step up one frame. For GE Type QL transformers, that rule yields: for a 200 kW continuous resistive load (PF=1), minimum kVA = 200 / (1 × 0.8) = 250 kVA → 300 kVA frame (since 250 is not a standard rating). For a 150 kW motor (PF=0.85), min kVA = 150 / (0.85 × 0.8) = 221 kVA → 225 kVA. That matches the Case 2 table exactly. Do not round down.

Topology/standards per the cited standards; all product ratings are manufacturer-stated values from the cited datasheets, current to 2026-06; derived/illustrative figures are labelled as such. This is not an independent head-to-head test. GE is a brand affiliated with this site; competitor names are used for identification only.