Calibration: meeting the challenges of high-frequency power measurement

October 13, 2015 //By Clive Davis and Erik Kroon
Calibration: meeting the challenges of high-frequency power measurement
As more and more innovation focuses on energy efficiency and the use of renewable energy resources, engineers are increasingly demanding accuracy and precision from their power measurements. At the same time, new standards such as IEC62301 Ed2.0 and EN50564:2011, covering standby power consumption, and the SPEC guidelines, covering power consumption in data centres, demand more precise and accurate testing to ensure compliance.

To meet these challenges, R&D teams are coming to terms with the need for new levels of precision in power measurement, but these levels of precision can only be achieved if the measuring instruments are properly calibrated with reference to national and international standards.

Regular calibration by a laboratory, which can provide very low measurement uncertainties at the specific measurement points applicable to individual users, should enable instrument makers and their customers to have confidence in their test results.

However, with power measurements in particular, the situation is not so clear cut – to the extent that the accuracy figures quoted in manufacturers' specifications – and indeed some of the parameters listed in calibration certificates issued by well-established test houses – can be meaningless when taken out of context.

One key area which is often neglected in traditional specifications is that of power measurements at high frequencies. Traditionally, AC power meters are calibrated at frequencies of 50-60 440 Hz. Nowadays, however, there is a demand for power measurement at high frequencies on devices such as switch-mode power supplies, electronic lighting ballasts, soft starters in motor controls and frequency converters in traction applications.

Calibration of high-frequency power has lagged behind the development of power meters to address these applications, and few national laboratories can provide traceability up to 100 kHz: the frequency at which instruments have to be calibrated to provide accurate results in these application sectors.

There are a number of other parameters involved in power measurements that determine the performance of an instrument in a particular application. It is no longer sufficient merely to list voltage and current specifications: today’s power environment needs to address variables such as phase shift, power factor and the effects of distorted waveforms.

It is also important to calibrate the instrument under the right conditions. Many test houses still use pure sine waves at only 50 Hz to calibrate power meters, which renders the results virtually useless for users carrying out tests under “real world” conditions.

It is therefore important for users of power measuring instruments to look at the actual “calibrated” performance of different manufacturers’ products rather than just comparing specifications. This is the key thinking behind Yokogawa’s policy of having its own European Standards Laboratory with minimal uncertainties and capabilities which are almost second to none: as confirmed by the fact that it has become the world’s first non-governmental facility to receive full ISO17025 accreditation for power measurements at up to 100 kHz.

What is calibration?

“Calibration” can be defined as the comparison of an instrument’s performance with a standard of known accuracy.

No measurement is ever correct. There is always an unknown, finite, non-zero difference between a measured value and the corresponding “true” value. In other words, a user can never be 100% sure that an instrument is operating within its specified tolerance limits.

However, there are steps that can be taken to minimise the possibility of a measurement falling outside specified tolerance or uncertainty bands. Regular traceable calibration is a method for gaining quantifiable confidence in a measurement system by comparing the instrument’s performance to a standard of known accuracy.

However, all laboratory standards and even national standards have uncertainties of measurement; hence it is difficult to be 100% confident that an instrument is operating within its stated tolerance limits.

It is important to understand the difference between “calibration” and “adjustment”. Calibration is the comparison of a measuring instrument (an unknown) against an equal or better standard. A standard in a measurement is therefore the reference.

Instruments are adjusted initially at the factory to indicate a value that is as close as possible to the reference value. The uncertainties of the reference standard used in the adjustment process will also dictate the confidence that the indicated value is “correct”.

As the instrument ages, the indicated value may drift due to environmental factors (temperature, humidity, oxidation, loading etc.) which will also be dependent on the quality of its design and manufacture. To ensure that the instrument continues to operate within the manufacturer’s tolerances, the instrument should be compared to the reference value on a regular basis (usually annually). If necessary, the instrument can then be re-adjusted.

If there is no appreciable change in the calibrated results, this means that the instrument’s design is inherently highly stable. In this case, there is no need to adjust it, and the user can also rely on the fact that the unit will exhibit the same stability on a day-by-day basis.

It goes without saying that all instruments should be calibrated on a regular basis. Not calibrating carries a number of costs and risks:

In production/acceptance testing, users may encounter false passes or (equally undesirable) false failures.

In an engineering laboratory, inaccurate measurements can distort the findings.

Contractual requirements may inadvertently fail to be met

Quality issues may result in customer dissatisfaction or even product recalls and rework.

Regular calibration by a laboratory, which can provide very low measurement uncertainties at the specific measurement points applicable to individual users, should enable instrument makers and their customers to have confidence in their test results.

Yokogawa’s calibration capabilities

As indicated above, Yokogawa’s European Calibration Laboratory has become the world’s first non-governmental facility to receive full ISO17025 Accreditation for power measurements at up to 100 kHz. This is in addition to its established capability for providing high-accuracy calibration at 50 Hz, especially at very low power factors (down to 0.0001) and at high currents.

At the heart of the laboratory is a special calibration system with the capability to calibrate power up to 100 kHz.

Housed in a climate-controlled environment (23.0 ± 1.0)°C, the system is able to calibrate voltage, current, DC power, AC power, frequency and motor functions, all under fully automatic control.

The system consists of two parts, a signal generator section and a reference measuring unit. Those two parts are separated by a metal shield in order to prevent the heat generated by the signal sources from affecting the reference meters. Different sources are used to generate the calibration signal because a single source is not sufficient to generate all the signals required. Instead, multiplexers are used to select the sources needed for any particular measurement. Similarly, different reference power meters are selected via a multiplexer.

This allows the selection of the best reference power meters for high-current, low-current or low-voltage calibrations. The reference power meters are from Yokogawa and are special models modified with aged components and firmware to enhance the resolution. They have an excellent stability and performance.

The power meter under calibration is connected via a multiplexer on the calibration system. Each element of the power meter under calibration is calibrated separately. A power meter with mixed inputs is easily calibrated. Extra instruments are added to the calibration system to calibrate any additional options of the power meter such as the motor function and analogue output.

The system is designed to minimise effects such as capacitive leakage and crosstalk, with special attention given to the wiring harness and multiplexers. For voltage, twisted and screened coaxial cables are used, while current uses coaxial cabling. The multiplexers use special relays to avoid leakage and crosstalk. The influence of the wiring harness is now kept to a minimum. With a worst-case measurements using 100 V at 100 kHz, the crosstalk suppression to the current channels is better than -93 dB. For mixed-input units, every element is calibrated separately to minimise the effects of loading.


Vous êtes certain ?

Si vous désactivez les cookies, vous ne pouvez plus naviguer sur le site.

Vous allez être rediriger vers Google.