Transformer Sizing Mistakes to Avoid
Most sizing mistakes are not arithmetic errors; they are loads that were forgotten, growth that was assumed away, or operating conditions that were never stated. This guide lists the mistakes that show up after installation — when fixing them is expensive — and the checks that catch them while they are still cheap.
Forgetting loads and diversity
The classic mistake is sizing from connected load alone, or from a nameplate list that omits motors starting together, future equipment, or loads that run at the same time. Work from a load list with running load, starting load, and diversity — which loads genuinely run together — and state your assumptions in writing so they can be challenged. A load list nobody reviewed is a guess with a spreadsheet attached.
Harmonics, power factor and voltage drop
Non-linear loads add harmonic heating that a simple kVA total does not capture, and a low power factor raises current for the same useful power. Check both before choosing a rating, and check voltage drop at the farthest load under starting conditions, not only at full load in steady state. A transformer that is fine on paper at unity power factor can still overheat or sag voltage in a real installation with harmonics and poor power factor.
Growth, cooling and protection coordination
Size for the growth you can reasonably foresee, or at least leave a documented margin and a physical path to a larger unit — a transformer that cannot be replaced without rebuilding the room was undersized in planning, not in rating. Check cooling: ambient temperature, ventilation and altitude all change what a rating means in practice. And coordinate protection so the transformer is protected without nuisance tripping on inrush or motor starts; a correctly sized transformer with wrong protection still fails the installation.
A sane sizing sequence
List every load with running and starting demand, apply diversity honestly, correct for harmonics and power factor, check voltage drop at the farthest load, add a stated growth margin, then verify cooling and protection. Document each step so the next engineer can see why the rating was chosen. If any step was skipped, say so plainly — an undocumented margin is not a safety factor, it is an unknown.
If you want a second pair of eyes on your own report or nameplate before you decide, see the remote transformer services — documents only, fixed price, fixed turnaround, no site visit. For the wider testing picture, start with the transformer testing and commissioning guide.