Selecting the correct transformer size is one of those decisions that looks simple on paper — add up the load, pick a kVA rating, done — but in practice it's one of the most consistently mis-calculated steps in electrical design. A transformer that's undersized runs hot, trips on overload, sags in voltage under starting loads, and shortens its own insulation life year after year. A transformer that's grossly oversized wastes capital, occupies more space and civil work than necessary, and runs inefficiently at light load for most of its service life. Between those two extremes sits the correct answer, and getting there means avoiding a specific, repeatable set of mistakes that show up again and again across industrial, commercial, and residential installations.
This guide walks through the most common transformer sizing errors engineers, technicians, and facility owners make, why each one causes real-world problems, and what to check instead. It closes with a practical step-by-step sizing approach and a set of frequently asked questions covering the details that don't always fit neatly into a single mistake.
Why Transformer Sizing Mistakes Are So Costly
A transformer isn't a component you resize easily after installation. Civil work, cable sizing, protection coordination, and switchgear ratings are all designed around the transformer's kVA rating, so a sizing error discovered after commissioning usually means either accepting reduced performance and life expectancy, or a costly retrofit. Undersizing shows up as nuisance tripping, overheating, and premature insulation breakdown; oversizing shows up as wasted capital and, less obviously, as poor efficiency — transformers are least efficient at very light loads relative to their rating, so an oversized unit can actually cost more to run over its lifetime despite the extra margin. Both failure modes trace back to the same handful of avoidable mistakes.
1. Ignoring Power Factor
One of the most common mistakes is sizing a transformer based only on real power (kW) without converting to apparent power (kVA) using the load's power factor. Transformers are rated in kVA, not kW, because a transformer has to supply both the real power that does useful work and the reactive power that magnetic loads like motors and inductive equipment demand. Comparing a kW figure directly against a kVA nameplate rating understates the actual apparent power the transformer must deliver.
For example, a 400 kW load operating at a power factor of 0.8 actually requires 500 kVA of transformer capacity — not 400 kVA. Skip the conversion and you'll select a transformer that looks adequate on paper but is genuinely undersized in the field. This mistake is especially common on sites with a large proportion of motor load, where power factor commonly sits between 0.75 and 0.85 without power factor correction equipment in place. Always confirm whether the load figure you've been given is already expressed in kVA or still needs the power-factor conversion applied.
The confusion is compounded when a facility already has power factor correction equipment installed, because the correction changes the apparent power the transformer sees at its own terminals, not just the apparent power the utility bills for. A capacitor bank installed downstream of the transformer, close to the correction-needing loads, reduces the reactive current the transformer itself has to supply — which genuinely reduces the required kVA rating. A capacitor bank installed only at the main incomer, upstream of a sub-transformer feeding a specific area, doesn't help that sub-transformer's own loading at all. Where the correction equipment sits in the single-line diagram relative to the transformer being sized changes whether its effect can be credited toward a smaller kVA rating.
2. Using Connected Load Instead of Demand Load
Many designers simply add up the nameplate ratings of every piece of connected equipment and size the transformer against that total. In reality, not everything runs at the same time or at full nameplate rating simultaneously — a facility's actual demand is almost always lower than its total connected load. Demand factors (the ratio of maximum demand to connected load for one type of equipment) and diversity factors (accounting for the fact that different loads peak at different times) both need to be applied to arrive at a realistic figure.
Skipping this step usually leads to unnecessary oversizing, which drives up transformer and switchgear cost without a corresponding benefit. It can occasionally lead in the other direction too — if a facility genuinely does run most of its connected load simultaneously (some continuous-process plants do), assuming a generic diversity factor that doesn't apply to that specific site can undersize the system. The correct demand load figure comes from actual load studies, historical demand data, or, where neither is available, published demand factor tables for the specific equipment category and industry type — not from a blanket percentage applied without justification.
3. Not Allowing for Future Expansion
Industrial and commercial facilities rarely stay static. Production lines get added, new HVAC equipment goes in, EV charging infrastructure gets retrofitted, and floor space gets repurposed — all of which add electrical load after the transformer has already been installed and the civil work poured around it. A transformer sized tightly for today's measured load can become the bottleneck that blocks or delays a future expansion, or worse, gets quietly overloaded because nobody wants to pay for a transformer replacement mid-project.
A spare capacity margin of roughly 20% to 30% above the calculated demand load is sometimes used as a preliminary planning allowance, but it is not a universal requirement — the right figure depends on how likely and how large future expansion is for that specific facility, the utility's own requirements, and the applicable design standards. A leased commercial building with a fixed tenant mix typically needs less margin than an owner-occupied manufacturing plant with an active expansion roadmap. Whatever margin is chosen, it should be a deliberate decision documented in the sizing calculation, not an afterthought applied inconsistently from project to project.
4. Ignoring Motor Starting Current
Large induction motors typically draw five to eight times their full-load running current for the first few seconds during direct-on-line starting. If a sizing calculation only accounts for steady-state running current, the transformer can appear correctly sized on paper while still causing severe voltage dips every time a large motor starts. That voltage dip doesn't just affect the motor being started — it propagates across the entire bus, potentially dimming lighting, tripping sensitive electronic equipment, or causing contactors on other circuits to drop out momentarily.
The severity of the starting-current problem depends on the ratio between the motor's starting kVA and the transformer's kVA rating, as well as the transformer's percentage impedance — a higher-impedance transformer will show a larger voltage dip for the same starting event. For facilities with large motors relative to transformer size, soft starters, variable frequency drives, or star-delta starting can reduce the starting current spike substantially, but the transformer sizing calculation still needs to account for whichever starting method is actually specified, not assume the mildest case by default.
5. Overlooking Harmonic Loads
Modern facilities run a growing share of nonlinear load: variable frequency drives, UPS systems, LED lighting drivers, server power supplies, and other switch-mode electronics. These loads draw current in non-sinusoidal pulses rather than smooth sine waves, injecting harmonic currents back into the supply. Harmonics cause additional heating in a transformer beyond what the fundamental-frequency current alone would produce — eddy current losses and stray losses both increase disproportionately with harmonic content, and standard-design transformers aren't rated to dissipate that extra heat without a capacity de-rating.
Where nonlinear loads such as VFDs, UPS systems, rectifiers, and switch-mode power supplies are significant, the transformer should be evaluated for harmonic heating rather than relying only on the total kVA load. Depending on the harmonic spectrum and loading profile, the design may require a K-rated transformer, harmonic-mitigation measures, or an appropriate de-rating below nameplate kVA — the required K-factor should be determined from the expected harmonic current profile and applicable design standards, not assumed from a fixed percentage of total load. Ignoring this factor doesn't cause an immediate failure the way a badly undersized transformer might; instead it shows up gradually as reduced insulation life and unexplained overheating that's hard to trace back to its root cause months or years later.
6. Choosing the Exact Calculated Size
Another common mistake is selecting a transformer with a rating exactly equal to the calculated demand load, with zero margin at all. Real electrical systems rarely operate under the exact, static conditions used in a sizing calculation — load varies through the day and across seasons, ambient temperature swings, equipment gets added, and metering and estimation always carry some error. Selecting the next standard transformer rating above the calculated requirement, rather than the exact figure, absorbs this normal variability without immediately pushing the transformer into an overloaded condition.
Transformers are generally available in standardized preferred kVA ratings, although the exact rating series depends on the applicable standard, manufacturer, and market (illustrative examples for distribution transformers include ratings around 5, 10, 15, 25, 50, 63, 75, 100, 160, 200, 250, 315, 400, 500, 630, 800, 1000, 1250, 1600, 2000, and 2500 kVA — the same reference series used by the Transformer Size Calculator). In practice, this mistake usually means rounding down to the nearest available size below the calculated load to save cost, rather than rounding up to the next suitable standard size above it. That decision might save money on the transformer itself, but it removes any operating margin and turns ordinary load variability into an overload condition.
7. Ignoring Ambient Temperature
Transformer nameplate ratings are based on a standard reference ambient temperature (commonly 40°C for oil-filled distribution transformers, or a stated design ambient for dry-type units). In hot climates or poorly ventilated plant rooms, actual ambient temperature routinely exceeds that reference value, and the transformer's effective cooling capacity — and therefore its safe loading — drops accordingly. A transformer that's correctly rated for the calculated demand load at 40°C ambient can be genuinely overloaded at the same demand if it's actually operating in a 50°C plant room with poor airflow.
High-ambient installations may need either a transformer de-rated below nameplate for the specific site conditions, or a larger unit selected up front to absorb the temperature penalty, along with attention to ventilation and enclosure design. This is a particularly common oversight in enclosed substations and rooftop plant rooms in tropical and subtropical climates, where solar heat gain on an enclosure roof can push internal ambient well above outdoor air temperature.
8. Overlooking Cooling Type and Loading Curves
Oil-filled transformers are commonly available in multiple cooling configurations — ONAN (oil natural, air natural) as the base rating, with ONAF (oil natural, air forced, using fans) offering a higher rating on the same core and tank by improving heat dissipation. Treating every transformer as a fixed, single kVA number without checking which cooling stage that number corresponds to can lead to a mismatch between the assumed capacity and what's actually available under normal (fan-off) operating conditions. A transformer's ONAF rating typically isn't available continuously without the cooling fans running and controlled correctly, so a sizing calculation that assumes the higher ONAF figure as the baseline capacity is building in an assumption that depends on auxiliary equipment functioning as designed.
This distinction matters most for a facility's contingency planning: if a design relies on the ONAF rating to meet peak demand, a fan failure or a controller fault during that peak reduces the transformer's safe capacity back down to its ONAN rating, precisely when the load is highest. The applicable continuous rating should follow the manufacturer's specified cooling-stage ratings and the required operating duty — for critical loads where a fan-cooling failure shouldn't translate directly into an overload condition, sizing against the base ONAN rating and treating ONAF uplift as genuine spare margin is the more conservative choice, but ONAF capacity can be credited where the cooling system is included in the design and its availability and control requirements are satisfied.
9. Neglecting Voltage Drop and Fault Level
Transformer sizing doesn't end at kVA capacity. The transformer's percentage impedance affects both the voltage drop under load and the fault current available on the secondary side — two figures that feed directly into downstream cable sizing, protective device coordination, and switchgear fault ratings. A transformer selected purely on kVA capacity without checking its impedance can produce a system where voltage drop under normal load is outside acceptable limits, or where downstream breakers aren't rated to interrupt the available fault current safely. Both of these are sizing-adjacent decisions that get skipped when "sizing" is treated as a single kVA number rather than a full set of electrical parameters.
10. Overlooking Impedance and Vector Group Compatibility for Parallel or Replacement Units
Two situations where transformer "sizing" quietly turns into a compatibility problem, not just a capacity problem, are parallel operation and like-for-like replacement. When two transformers run in parallel to share a common load, their percentage impedances need to be close enough that load divides between them roughly in proportion to their kVA ratings — a mismatch in impedance causes the lower-impedance unit to pick up a disproportionate share of the load, potentially overloading it while the higher-impedance unit runs comfortably under-loaded. Vector group also has to match (or be compatible) for parallel operation, since paralleling transformers with incompatible vector groups can create damaging circulating currents even with well-matched impedance and correctly matched voltage ratios.
Replacement sizing has a related trap: swapping in a "bigger, better" replacement transformer for a failed or aging unit without checking its impedance against the original can unexpectedly change the fault current available on the downstream system, sometimes pushing it above the interrupting rating of existing breakers that were coordinated against the old transformer's impedance. A straightforward kVA-for-kVA (or larger) replacement can therefore still require re-checking protection coordination and fault levels, not just confirming the new unit fits in the same footprint.
How to Size a Transformer Correctly — Step by Step
Bringing the points above together, a sound transformer sizing process generally follows this sequence:
- Establish the actual demand load in kW, using demand and diversity factors rather than simply summing connected nameplate ratings.
- Convert to kVA using the site's actual or design power factor, not an assumed round number.
- Add a future-expansion margin where appropriate — often in the 20% to 30% range as a starting point — based on how likely and how large future growth actually is for that specific facility, and the applicable design requirements.
- Check motor starting requirements against the transformer's impedance to confirm voltage dip stays within acceptable limits for the largest motor(s) on site.
- Assess harmonic content from VFDs, UPS systems, and other nonlinear loads, and apply a K-rating or de-rating where the nonlinear load share is significant.
- Round up to the next standard kVA rating rather than down, to preserve operating margin.
- Correct for ambient temperature and cooling type specific to the installation location and enclosure design.
- Verify voltage drop and fault level against downstream cable and switchgear ratings before finalizing the selection.
- Check impedance and vector group compatibility where the transformer will run in parallel with another unit, or is replacing an existing one whose downstream protection was coordinated against its impedance.
Conclusion
Proper transformer sizing requires more than adding up equipment nameplate ratings and picking the nearest kVA figure. Power factor, demand versus connected load, future expansion, motor starting current, harmonic content, ambient temperature, cooling configuration, downstream voltage drop and fault level, and impedance/vector-group compatibility for parallel or replacement units all interact, and a sizing calculation that treats them individually — rather than as separate boxes to check — is what actually holds up under real operating conditions. Avoiding the mistakes covered here improves system reliability, keeps operating costs predictable, and gives the transformer a realistic chance of running its full rated service life while still leaving room for the facility to grow. None of this replaces a proper engineering study for a specific site — it's a checklist for making sure that study starts from the right questions.