A converter that looks efficient on a data sheet can still underperform once it is switching at full load, running hot, or operating across a wide input range. That is why understanding how to calculate power conversion efficiency matters well beyond a headline percentage. For OEMs and power design engineers, the calculation is straightforward. The engineering judgement behind it is where real system value is won or lost.
At its simplest, power conversion efficiency is the ratio of output power to input power, expressed as a percentage. The formula is:
Efficiency (%) = (Pout / Pin) x 100
Where output power is the useful electrical power delivered to the load, and input power is the total electrical power drawn from the source. If a DC-DC converter takes 1,000 W at the input and delivers 960 W at the output, its efficiency is 96%.
That sounds simple because it is. The complication comes from measuring the right quantities under the right conditions, and then interpreting the result in a way that reflects actual operating performance.
How to calculate power conversion efficiency correctly
To calculate efficiency, you first determine input power and output power using measured voltage and current. In a DC system, the equations are direct:
Pin = Vin x Iin
Pout = Vout x Iout
You then divide output power by input power and multiply by 100. For example, if your converter input is 400 V at 2.5 A, input power is 1,000 W. If the output is 48 V at 19.2 A, output power is 921.6 W. Efficiency is therefore:
(921.6 / 1000) x 100 = 92.16%
In AC systems, the calculation can be more involved because real power must account for power factor. In that case, input power is not simply RMS voltage multiplied by RMS current unless power factor is unity. For single-phase AC input:
Pin = Vin x Iin x power factor
For three-phase systems, use the appropriate three-phase real power equation. This distinction matters in front-end power supplies, motor drives and inverter stages where waveform distortion and phase angle can materially affect the result.
The key point is that efficiency must always be based on real input power and real output power, not apparent power.
What the efficiency figure is actually telling you
Efficiency is a compact way of expressing total loss. If a converter is 97% efficient, then 3% of the input power is being lost, mostly as heat. In a 10 kW system, that 3% represents 300 W of loss. At 94%, losses rise to 600 W. The percentage shift looks small. The thermal and system-level consequence is not.
This is why high-efficiency devices are not just a specification exercise. Lower loss can reduce heatsink size, fan requirements, enclosure complexity and long-term thermal stress on surrounding components. In high-voltage switching applications, semiconductor selection has a direct influence on that result. SiC devices, particularly in fast-switching topologies, are often chosen because they reduce switching losses while supporting strong thermal performance at elevated voltages and frequencies.
Measuring efficiency in the lab
If you want a number that means something, test conditions must be tightly controlled. Input voltage, output load, switching frequency, ambient temperature and cooling method all influence measured efficiency. So does instrumentation accuracy.
A sound measurement setup typically uses calibrated power analysers or precision voltage and current measurement at both input and output. Four-wire sensing may be required to avoid cable drop errors, especially at high current on low-voltage outputs. If your output power is calculated from a voltage measured at the converter terminals but the actual load sees lower voltage due to interconnect losses, the result can be misleading.
Temperature is another major variable. Semiconductor conduction losses and switching behaviour change with junction temperature. Magnetic components and capacitors also shift with heat. A converter tested after two minutes on the bench may show a different efficiency from the same unit at thermal equilibrium after 30 minutes.
That is why efficiency curves published by serious power semiconductor suppliers are usually tied to defined operating conditions, not presented as isolated peak values.
Peak efficiency versus weighted efficiency
One of the most common mistakes is treating peak efficiency as the only meaningful number. In practice, many systems spend little time at their peak-efficiency operating point. Telecom power supplies, EV charging subsystems, solar inverters and UPS platforms often operate across broad load ranges.
A converter might achieve 98.5% at one specific load and voltage combination, but average materially less across the mission profile. For real design work, it is often more useful to evaluate efficiency at 10%, 25%, 50%, 75% and 100% load, then apply weighting based on expected use.
This is where device characteristics become commercially relevant. Lower reverse recovery behaviour, reduced switching loss and stronger high-temperature performance can improve efficiency across a broader operating window, not just at one preferred point. That can be more valuable than chasing a marginally higher peak number.
Where power losses come from
If the efficiency result is disappointing, the next question is obvious: where is the power going?
Losses in power converters are usually grouped into conduction losses, switching losses, magnetic losses, gate drive losses, control losses and parasitic losses. Conduction losses arise when current flows through semiconductors, inductors, transformers and interconnects with finite resistance. Switching losses occur during turn-on and turn-off transitions, where voltage and current overlap for a finite time. At higher frequencies, these losses can dominate.
Magnetic losses include core loss and copper loss in inductors and transformers. Control circuitry, bias supplies and gate drivers also consume power, which can become significant at light load. Parasitics in layout, package inductance and capacitance can create additional switching stress and energy dissipation.
When you calculate efficiency, you are seeing the sum of all of these effects. When you optimise efficiency, you are reducing them one by one.
How to calculate power conversion efficiency for different topologies
The core equation does not change with topology, but the test method and interpretation often do.
For isolated DC-DC converters, output power is straightforward to calculate, but transformer losses and snubber dissipation can materially affect performance. In PFC stages, AC real power measurement is essential because input current is shaped rather than purely resistive. In inverter systems, output waveform quality matters, and whether you are measuring DC-AC stage efficiency alone or total system efficiency must be clearly defined.
In multi-stage architectures, calculate efficiency stage by stage as well as end to end. If a front-end stage is 97% efficient and a downstream stage is 95% efficient, total efficiency is not 192%. It is the product of the two efficiencies:
Total efficiency = 0.97 x 0.95 = 0.9215 or 92.15%
This catches teams out more often than it should. Multi-stage systems accumulate losses quickly, which is one reason advanced semiconductor choices can make a disproportionate difference in high-power designs.
Common mistakes that distort the result
The most frequent error is mixing instantaneous readings with unstable operating conditions. If voltage or current is fluctuating, a basic handheld measurement may not represent true average power. Another issue is ignoring auxiliary loads such as cooling fans, control boards or standby supplies. If those are part of the converter subsystem in real operation, excluding them overstates efficiency.
There is also the question of boundary definition. Are you measuring semiconductor stage efficiency, converter module efficiency, or full system efficiency at the wall? Each can be valid, but only if clearly stated. Procurement teams and design engineers need comparable numbers, not selectively framed ones.
Finally, be cautious with very light-load conditions. At low output power, fixed overhead losses can make efficiency drop sharply. That does not necessarily indicate a poor design. It may simply reflect the converter’s intended operating envelope.
Why semiconductor selection changes the efficiency equation
Efficiency is not determined by the formula. It is determined by the hardware behind the formula. Device choice influences conduction loss, switching loss, thermal margin and achievable switching frequency, all of which shape measured conversion efficiency.
In high-voltage applications, SiC Schottky diodes and SiC MOSFETs can materially improve efficiency by reducing reverse recovery effects and switching energy compared with conventional silicon alternatives. The benefit is especially clear in demanding switching environments where higher frequency, lower thermal burden and compact power density are all commercial priorities. For design teams trying to improve both electrical performance and total system cost, that trade-off is often decisive.
This is where manufacturers such as Quest Semiconductor position advanced SiC devices – not as theoretical upgrades, but as practical tools for cutting loss, easing thermal management and improving conversion performance in real industrial systems.
The result only matters in context
A good efficiency number is useful. A well-understood efficiency number is far more valuable. The same 97% can be excellent in one topology and underwhelming in another, depending on voltage class, switching frequency, thermal constraints and load profile.
So if you are working out how to calculate power conversion efficiency, start with the formula, but do not stop there. Measure real power accurately, define the test boundary, check performance across the operating range, and relate every percentage point back to heat, reliability and system cost. That is where efficiency stops being a lab metric and starts becoming an engineering advantage.