What an Inverter-Duty Efficiency Figure Actually Measures

An inverter-duty datasheet carries an efficiency figure, and it looks like the figure on a standard datasheet. It was produced by a different measurement, against a different supply, and it answers a narrower question.

Direct answer

Converter-fed efficiency has its own test standard because it is a different quantity. A converter output is not a sine wave, and the harmonics produce additional high-frequency losses in the machine that a sinusoidal supply never causes.

To make results comparable, the testing is done against a comparable converter — a standardised reference, so that two motors tested by two laboratories are being asked the same question.

That is the strength of the method and also its limit. The standard says so directly:

What “comparable converter” buys, and what it costs

The reference converter exists so that comparison is possible at all. Without it, every efficiency figure would be a figure for one particular motor-plus-converter pair, and no two datasheets would be comparable.

The cost is that the answer is about the reference, not about your drive. Determining the real losses for a specific installation means testing the whole power drive system with the converter that will actually be used.

Two situations widen the gap:

Multi-level and current-source converters. The baseline for the comparable converter is a two-level voltage-source converter. With multi-level voltage-source or current-source converters, the additional high-frequency losses vary much more with speed and load — so the deviation from the standardised figure is larger, not smaller.

Anything unusual in the pulse pattern. Losses can also be determined by calculation, but that requires the converter’s pulse pattern as an input — and the standard is explicit that such calculation procedures are outside its own scope.

The practical reading: the further your converter is from an ordinary two-level drive, the less the standardised figure tells you.

Low voltage only — and the gap that leaves

This is the sentence most likely to affect a specification:

The standard limits the comparable-converter route to low-voltage machines, and says plainly that it does not reach medium voltage. Whatever the figure is worth for comparing designs, that worth stops at the low-voltage boundary.

Put that beside the line-fed classification, which assigns IE classes to high-voltage cage motors inside a narrow window but excludes machines designed specifically for converter supply, and a gap appears:

• A converter-designed medium-voltage machine has no IE class from the line-fed standard, because converter-specific designs are excluded.
• It also has no comparable-converter efficiency figure, because that approach does not reach medium voltage.

That is not a loophole anyone is exploiting. It is a genuine boundary, and the consequence is that medium-voltage converter drives are compared on measured performance and declared values rather than on a class or a standardised figure. Knowing that in advance saves an argument about a class that does not exist.

Four methods, and which one matters

The standard sets out several methods, and the one used changes what the figure means:

Direct measurement and summation of losses are not interchangeable in what they prove; calculation is different again. This is the same discipline as any other declared value — a number without its method is not comparable with anyone else’s number, which is the point made in type test versus routine test.

So the useful request is not “what is the efficiency” but “what is the efficiency, determined by which method”.

The part that is genuinely useful: seven load points

Alongside the test methods, the standard specifies an interpolation procedure. Losses can be determined at any load point — any combination of torque and speed — within the constant-flux range, the field-weakening range and the overload range, derived from measurements at seven standardised load points.

That matters because a variable-speed drive is bought precisely so the machine does not sit at its rated point. A fan or pump on a drive spends most of its life at part speed, where power falls as the cube of speed. The efficiency at rated load is the least representative number you could pick.

The scope is broad here too: the procedure applies to any variable-speed AC motor, induction or synchronous, rated for variable frequency and variable voltage supply.

What to ask for

1. Whether the figure is line-fed or converter-fed — they are different measurements and should not be compared with each other.
2. Which method produced it.
3. Whether it is a comparable-converter figure, and therefore a design comparison rather than an installation prediction.
4. Losses at your operating point, not only at rated — this is available, via the interpolation procedure.
5. For medium voltage, what is being offered instead, since neither route above applies.
6. If the installation matters commercially, whether a drive-system test with the actual converter is possible.

LEADGO’s variable-frequency and inverter-duty series are specified against the converter rather than in isolation, for the reasons above and in VFD reactors and filters. When an efficiency figure is requested for a converter-fed application, the reply that is worth anything names the method and says whether it is a comparable-converter figure — the number alone does not carry that, and two numbers produced differently look identical on a datasheet.

Going deeper

• Do high-voltage motors have IE classes? — the line-fed window, and the converter exclusion that creates the gap above
• VFD reactors and filters — the derating curve, and why it needs the converter named
• Drive system efficiency, IE and IES — classifying the system rather than the motor
• Type test versus routine test — where a declared value comes from, and what a report proves