We are supported by readers, when you click & purchase through links on our site we earn affiliate commission. Learn more.

Eight energy cost meters for household use

Energy prices will explode in winter 2021/2022. A good time to optimize the energy requirements in your own four walls and thus reduce costs. Our series of articles with the focus on saving energy costs shows how and with what this works. This article originally appeared at the end of 2019 as a contribution to heise + – due to the occasion, we are releasing it again today for free. Some devices may no longer be available or may be available with a different interior.

If the last electricity bill made an unexpected leap, the question arises: Which new device or which changed behavior drove up the energy costs?

This can sometimes be clarified by the laying on of hands: For example, if the lamp base is warm, even if the lamp is not giving any light, then the power goes through. Perhaps in the lamp base, as in some halogen lights from the last millennium, a thick transformer hangs directly on the power supply and makes it cozy.

(Image:

c’t

)

Energy prices will explode in winter 2021/2022. A good time to optimize the energy requirements in your own four walls and thus reduce costs. Our series of articles with the focus on saving energy costs shows how and with what this works.

If you want to know more, you can take an energy cost meter, or EKM for short, and place it between the socket and the suspect consumer. Such devices can be borrowed from energy suppliers, but they are also available for little money from electronics dealers.

We brought eight representative models to the c’t laboratory and checked the plugs for them with measurements: These include five pocket money class EKMs (10 to 20 euros), two with particular convenience thanks to the smart home switching function and remote reading, and one something more expensive specimen, from which we expected particular precision.

If you want to know how exactly such devices measure energy, a brief excursion into physics is inevitable. Anyone who measures the current in amperes alone is far from knowing what it costs to operate an electrical consumer. Because the mains voltage can fluctuate between 207 and 253 volts and it does so in practice. Since it is not a constant, a current measurement alone is not enough. In any case, with complex loads (phase shift between voltage and current), this would only measure the apparent power in volt-amperes (VA). However, this is not relevant for the electricity bill in households.

Usable energy cost measuring devices to be plugged in between them therefore record voltage and current simultaneously, several thousand times per second, because the current with today’s consumers often has anything but the textbook sinusoidal shape (non-linear load). The product of both variables integrated over the measurement period – typically a half-wave of the mains voltage, at 50 Hertz 10 milliseconds – results in the currently used active power in watts (W). The EKM then adds up this instantaneous value over the operating time to the “consumed” energy (watt seconds, Ws, converted to kWh).

Since the time in the candidates can be measured with crystal oscillators at least one order of magnitude more precisely than voltage and current, the active power measurement determines the accuracy of the energy display and is therefore the feature of interest in this test.

To judge the accuracy of a measuring device, you need one that is significantly more accurate itself. We used one of our LMG95 precision power measuring devices, which is specified with a basic inaccuracy of 0.03 percent and is regularly calibrated by the manufacturer.

The app for the Voltcraft SEM6000 switch socket is convenient, but shows an unrealistic number of digits: The device is certainly not accurate to a thousandth of a watt.  But you can believe the specified measurement error of ± 1 percent from 10 watts.  The true value here is between 10.0 and 10.2 watts - precise enough for at home.

The app for the Voltcraft SEM6000 switch socket is convenient, but shows an unrealistic number of digits: The device is certainly not accurate to a thousandth of a watt. But you can believe the specified measurement error of ± 1 percent from 10 watts. The true value here is between 10.0 and 10.2 watts – precise enough for at home.

With a non-linear load of 2.8 watts – such as a small LED light – the LMG95 is off by a maximum of 0.04 watts after taking into account the separate errors in voltage and current measurement channels. With such small loads, the displayed active power can be adopted with one decimal place.

In order to create comparable conditions, we fed the test setup with clean 230 volts at 50 Hertz from an adjustable AC voltage source (AC source Agilent AG6813). Three standard household lamps of various types and three power packs had to be used as test loads.

A 60-watt incandescent lamp does not pose any problems: it draws a sinusoidal current that is in phase with the voltage, as described in physics books as the classic case of an ohmic consumer. Today, however, other loads are more common: A 7-watt energy-saving lamp behaves somewhat more critically with a capacitive phase shift, a 2.7-watt LED lamp with its small and above all non-linear current of just under 0.03 amps (rms value, corresponding to 6.6 VA at 230 volts) to the touchstone: The EKM should be able to measure up to 16 amperes, but still work precisely with small currents.

The current shape (light blue) of many household loads today is anything but sinusoidal: power supplies often draw a needle-shaped current.  Some, often those of notebooks and compact PCs, even change the operating mode when the load increases (NUC power supply unit up to approx. 40 watts on the left, with PFC on the right).  Constant changes can throw energy cost measuring devices off track in such a way that they display unrealistic values.

The current shape (light blue) of many household loads today is anything but sinusoidal: power supplies often draw a needle-shaped current. Some, often those of notebooks and compact PCs, even change the operating mode when the load increases (NUC power supply unit up to approx. 40 watts on the left, with PFC on the right). Constant changes can throw energy cost measuring devices off track in such a way that they display unrealistic values.

USB and PC power packs do not draw sinusoidal currents from the power grid under load. It is either nasty needle-shaped or at best approximately sinusoidal when the power factor correction in the power supply unit, which is prescribed for better compatibility with the power grid, “knocks down” the needles. As representatives of this consumer class, we took a USB charger, an ATX PC power supply and the power supply of a compact PC (NUC8i5BEK). The latter has only one output voltage, while the ATX power supply has several, including a separate standby supply rail for the idle state.

More from c't magazine More from c't magazine

The power supplies ran in two states, first with low load (PC / NUC in Suspend-to-RAM, 2 watts primary) or without load (USB charger, 0.1 watts), then with moderate load (PC / NUC idle with 20 watts, 10 watts for the charger). We controlled these load points reproducibly with electronic loads and were then able to simultaneously note the displays of the test item and reference in order to calculate the relative active power error in relation to the reference.

In order to uncover series variances, we always tested two copies of a model. If the results were close together, which was the case with all test items, we adopted the better result.

To home page