Key takeaways
- Transformer losses split into no-load (core) loss, present continuously, and load (copper) loss, which rises with the square of load current.
- IEC 60076-20 defines efficiency classes (A0-E0 for no-load, A0-C0 for load) used to compare designs across manufacturers.
- Total cost of ownership adds capitalised loss cost to purchase price: TOC = Price + A x P0 + B x Pk.
- A cheaper transformer with higher losses can cost more over 20-25 years once energy prices and loading are capitalised.
- Amorphous-core designs cut no-load loss substantially, which matters most on lightly loaded feeders that run continuously.
Transformer losses fall into two categories: no-load (core) loss, which is drawn continuously whenever the unit is energised, and load (copper) loss, which varies with the square of the load current. Because a distribution transformer typically stays in service for 20-30 years, the lifetime cost of these losses can exceed the purchase price several times over, which is why serious buyers evaluate total cost of ownership (TOC) rather than tender price alone.
No-load loss (core loss)
No-load loss, often written P0, is generated in the magnetic core by hysteresis and eddy currents as the alternating flux cycles through the steel. It is present the instant the transformer is energised and stays essentially constant regardless of load current, so it is sometimes called the “always-on” loss. For a transformer connected to the grid 8,760 hours a year, no-load loss alone can represent a meaningful share of a utility’s system losses, which is why regulators increasingly set minimum efficiency requirements around it.
No-load loss depends on core material, flux density, and manufacturing quality (joint design, burr-free lamination cutting, annealing). Conventional cores use grain-oriented silicon steel; premium designs use amorphous alloy, which has a disordered atomic structure that dramatically reduces hysteresis loss at the cost of being more brittle to work with and typically raising the purchase price.
Load loss (copper loss)
Load loss, Pk (also written Pcu), occurs in the primary and secondary windings due to I²R heating, plus a smaller stray-loss component from eddy currents in windings, tank walls, and structural steel induced by leakage flux. Because it scales with current squared, load loss is negligible at light load and dominant near or above rated capacity. A transformer specified with rated load loss of, say, 8,000 W at 100% load will show roughly 2,000 W at 50% load and around 720 W at 30% load.
Load loss depends on winding conductor (copper has lower resistivity than aluminium, permitting smaller cross-sections for the same loss, or lower loss for the same size), current density, and winding geometry. Reducing load loss generally means more conductor material, which adds cost, weight, and sometimes size.
Loading factor and why it matters
Because the two loss components respond differently to load, the right transformer for a given site depends on how heavily it will actually run. A unit destined for a lightly loaded rural feeder (perhaps 20-30% average loading) should be evaluated with no-load loss weighted heavily, since copper loss will rarely be significant. A unit feeding a continuously loaded industrial process near 70-80% of rated capacity should weight load loss more heavily. Applying a single generic loading assumption across very different applications is a common tendering mistake.
Capitalising losses: the TOC formula
The standard way to compare bids fairly is to convert lifetime loss costs into a present-value figure and add it to the purchase price:
TOC = Price + A x P0 + B x Pk
- Price is the ex-works or delivered purchase price.
- P0 is declared no-load loss in kW.
- Pk is declared load loss in kW at rated current.
- A is the capitalisation value per kW of no-load loss (currency/kW), reflecting energy cost, discount rate, and expected service life, applied at full weight since P0 runs continuously.
- B is the capitalisation value per kW of load loss (currency/kW), reduced from A by the square of the expected average loading factor, since Pk only appears when current flows.
Utilities and large industrial buyers typically state A and B directly in tender documents, derived from their energy tariff, cost of capital, and loss evaluation period (commonly 15-25 years). A simplified way to estimate B from A is B = A x (loading factor)², reflecting the squared relationship between current and copper loss.
Worked example: comparing two 1,000 kVA transformers
Consider two oil-immersed 1,000 kVA, 11/0.4 kV distribution transformers quoted for a site with an average loading factor of 50% and a utility-stated capitalisation of A = USD 6,000/kW (no-load) and B = USD 1,500/kW (load loss, already adjusted for the 50% loading factor).
| Parameter | Transformer A (standard core) | Transformer B (amorphous core) |
|---|---|---|
| Purchase price | USD 18,000 | USD 21,500 |
| No-load loss (P0) | 1.70 kW | 0.55 kW |
| Load loss (Pk) | 10.5 kW | 11.0 kW |
| Capitalised no-load cost (A x P0) | USD 10,200 | USD 3,300 |
| Capitalised load cost (B x Pk) | USD 15,750 | USD 16,500 |
| Total cost of ownership | USD 43,950 | USD 41,300 |
Despite costing USD 3,500 more to buy, Transformer B shows a lower TOC by roughly USD 2,650 once losses are capitalised, because its amorphous core cuts no-load loss by nearly 68%. If the site’s average loading factor were higher (say 70-80%, weighting load loss more heavily and no-load loss less), the gap would narrow or could reverse, since Transformer B’s slightly higher load loss would count for more relative to its no-load saving. This is precisely why the loading assumption must reflect the real application rather than a default figure.
IEC 60076-20 efficiency classes and Ecodesign awareness
IEC 60076-20 establishes efficiency (loss) classes for liquid-immersed and dry-type distribution and power transformers, giving buyers a common scale instead of comparing raw kW figures across manufacturers with different rated outputs. No-load loss classes typically run A0 (lowest loss) through E0, and load loss classes A0 through C0, with each letter step corresponding to a defined percentage change in maximum permitted loss at a given rating.
In the EU, Ecodesign Regulation (EU) 548/2014 as amended sets minimum efficiency Tier requirements for transformers placed on the market, with a Tier 2 threshold that tightened no-load and load loss limits for most rating bands. Buyers specifying for EU markets, or markets that reference similar frameworks, should be aware which tier a quoted design meets, since this affects both compliance and the realistic loss figures achievable at a given price point. Outside the EU, similar minimum-efficiency regimes exist in other jurisdictions; where no mandatory tier applies, TOC evaluation remains the most rigorous way to select between compliant designs.
Stray loss and auxiliary losses
Beyond the two headline components, real installations also see stray losses from eddy currents induced in tank walls, clamping structures, and any nearby steelwork by leakage flux, and, on forced-cooled designs (ONAF/ODAF), the auxiliary power drawn by fans and pumps. Stray loss is usually reported as part of the declared load loss figure rather than separately, but on large power transformers it can be significant enough that manufacturers optimise tank shielding and clamping material specifically to control it. Auxiliary cooling power is not part of P0 or Pk in the strict IEC 60076-20 sense, but should be included in any full running-cost comparison between a naturally cooled (ONAN) unit and a forced-cooled (ONAF) unit of the same rating, since the forced-cooled unit trades a smaller, cheaper core and windings for continuous fan or pump power whenever cooling stage two is active.
Temperature effects on losses
Load loss is temperature-dependent because winding resistance rises with conductor temperature, so a figure measured at 20°C ambient during a factory test is corrected to the reference temperature (75°C or 85°C depending on insulation class) specified in the standard, to make figures comparable between tests performed on different days. No-load loss, by contrast, is essentially independent of winding temperature but is sensitive to supply voltage and waveform distortion, since core loss rises faster than linearly with flux density; a transformer operated at sustained overvoltage, or on a supply with significant harmonic content, can show measurably higher no-load loss in service than its factory-tested value at rated sinusoidal voltage. This is one reason site measurements occasionally diverge slightly from factory test report figures, and it is worth confirming supply voltage and quality assumptions when a discrepancy is investigated.
Reading a test report loss figure
Declared no-load and load loss values are verified during routine factory testing (see our FAT checklist for the full routine test sequence) and are typically permitted a positive tolerance under IEC 60076-1, commonly around 15% on individual no-load loss and 6% on individual load loss, with tighter combined tolerances — buyers should check the applicable edition and specify tolerances explicitly in the purchase order rather than assuming a default. When comparing bids, always confirm losses are stated at the same reference temperature (commonly 75°C or 85°C depending on insulation class) since figures at different reference temperatures are not directly comparable.
Summary
Selecting a transformer on purchase price alone ignores the larger cost usually hidden in decades of no-load and load losses. A structured TOC calculation using the buyer’s own capitalisation values and a realistic loading factor, cross-checked against IEC 60076-20 efficiency classes and applicable Ecodesign tier requirements, gives a defensible basis for comparing bids. Our transformer sizing calculator can help estimate the loading factor to use in this analysis for a given site.
How MARS can help
MARS manufactures oil-immersed and cast-resin dry-type distribution and power transformers from 5 kVA to 31,500 kVA, designed to IEC 60076, with copper, aluminium, or amorphous-core winding options to suit different loading profiles and loss targets. Every unit is routine-tested for no-load and load loss before dispatch, with reports available to support TOC evaluation. To discuss loss classes and capitalisation values for a specific project, visit our products page or request a quote.
Frequently asked questions
What is the difference between no-load loss and load loss in a transformer?
No-load (core) loss occurs whenever the transformer is energised, from hysteresis and eddy currents in the core, and is roughly constant regardless of load. Load (copper) loss occurs in the windings, rises with current squared, and is zero at no load and maximum at rated load.
How do I calculate the total cost of ownership of a transformer?
Add the purchase price to capitalised losses: TOC = Price + A x P0 + B x Pk, where P0 is no-load loss in kW, Pk is load loss in kW, and A and B are capitalisation values in currency per kW set by the buyer's energy cost and expected loading, typically quoted by utilities in tender documents.
What are IEC 60076-20 efficiency classes?
IEC 60076-20 defines loss classes for liquid-immersed and dry-type transformers: A0 to E0 for no-load loss (A0 lowest loss) and A0 to C0 for load loss, letting buyers compare declared losses across manufacturers on a common scale rather than reading raw kW figures alone.
Do amorphous-core transformers really save money?
Amorphous cores typically cut no-load loss by 60-80% compared with grain-oriented silicon steel, at a purchase price premium. They pay back fastest on continuously energised, lightly loaded transformers such as distribution feeders, and slower on heavily loaded, short-duty units where load loss dominates.
What loading factor should I assume when comparing transformer bids?
Use the actual expected average load as a fraction of rated capacity, often 30-60% for distribution transformers. A higher assumed loading factor weights load loss more heavily in the TOC comparison; an unrealistic assumption favours the wrong design.
Is a transformer with lower losses always the right choice?
Not always. Lower-loss designs usually cost more to buy and may be physically larger or heavier. The right choice depends on capitalised loss values, expected loading, duty cycle, and installation constraints, which is why a TOC calculation rather than price alone should decide the award.
Tags: transformer losses total cost of ownership no-load loss load loss efficiency
