Accuracy of measurements

    I was faced with a fact that surprised me and most likely will surprise you. It turns out that measuring the voltage in the network with an accuracy of at least one volt is almost an impossible task.



    The six instruments in this photo show different values, and the maximum differs from the minimum by more than 6 volts.


    In the process of preparing an article on power meters, I conducted an experiment with simultaneous measurement of mains voltage with several instruments and, having obtained such different results, I began to understand with accuracy.

    Typically, for digital devices, manufacturers indicate precision in the form of ± (0.8% + 10). This record means plus or minus 0.8% plus 10 lower order bits. For example, if the device measures voltage and shows integer and tenth values, then at a voltage of 230 volts its accuracy will be ± (230/100 * 0.8 + 10 * 0.1), that is ± 2.84 V (ten low-order units in this case are 1 volt ).

    Sometimes accuracy is specified as ± (0.5FS + 0.01). FS is Full Scale. Such a record means that the device may have deviations of readings up to 0.5% from the limit of the measuring range plus 0.01 volts (if it is a voltmeter). For example, if the range is 750V and ± (0.5FS + 0.01) is indicated, the deviation can be up to ± (750/100 * 0.5 + 0.01), i.e. ± 3.76 V regardless of which voltage is measured.

    There are two unpleasant nuances.

    Often, in the characteristics of the device, manufacturers indicate the general accuracy values ​​for the type of measurement, and on certain ranges everything can be even worse. So, for my UNI-T UT61E multimeter, which I always thought was very accurate, for measuring AC voltage everywhere, including the manufacturer’s website, the accuracy is specified ± (0.8% + 10), but if you read the instructions carefully , you can find on page 48 Here is a sign:



    In the 750 V range at the network frequency, the measurement accuracy actually amounts to ± (1.2% + 10), that is, ± 3.76 V at 230 V. The

    second nuance is that the accuracy of the record depends on how many decimal places the instrument indicates. ± (1% + 20) may be more accurate than ± (1% + 3) if the first device shows two decimal places, and the second one. In the characteristics of the devices, the number of decimal places on each range is rarely indicated; therefore, one can only guess about the real accuracy.

    From the sign above, I found out the amazing. It turns out that my UNI-T UT61E at a voltage of up to 220 volts shows two decimal places, and therefore it has an accuracy of ± 1.86 V at a voltage of 220 V, because in this case the recording is ± (0.8% + 10) 10 is just 0.1 V , but at a voltage of more than 220 volts, it begins to show one decimal place and the accuracy decreases more than twice.

    I do not suck your head yet? :)

    With my second multimeter, the Mastech MY65 is still more interesting. Its box shows the accuracy of measuring AC voltage for a range of 750V ± (0.15% + 3). The device in this range has one decimal place, it means accuracy is ± 0.645 V at a voltage of 230 V.

    But it was not there! The box contains instructions, it already has ± (1% + 15) on the same 750 V range, and this is ± 3.8 V at a voltage of 230 V.

    But this is not all. We look at the official site . And there already ± (1.2% + 15), that is ± 4.26 V to 230 V. Accuracy unexpectedly decreased almost seven times!

    This MY65 is generally weird. Under this name are sold two different multimeters. For example, on the same site, green MY65 and yellow MY65 with different capabilities, different designs and different parameters.

    In Chinese online stores often there is such a thing for $ 3.5, which is stuck in the outlet and shows the voltage.



    Do you know what her accuracy is? ± (1.5% + 2). Now you know how to decipher it. A piece shows whole volts, so at a voltage of 230 volts its accuracy is ± (230/100 * 1.5 + 2), that is ± 5.45 V. As a joke, plus or minus a tram stop.

    So how to measure the voltage in the network with guaranteed accuracy at least up to a volt in living conditions? No way!
    The most accurate multimeter that I managed to find on the network - UNI-T UT71C costs $ 136 and, when measuring AC voltage in the 750 V range, shows two decimal places and has an accuracy of ± (0.4% + 30), that is, at a voltage of 230 volts ± 1.22 AT.

    In fact, everything is not so bad. Many instruments have real accuracy 10 times higher than stated. But this accuracy is not guaranteed by the manufacturer. Maybe it will be much more accurate than promised, but maybe not.

    ps Thanks to Oleg Artamonov for his advice during the preparation of the article.

    2016, Alexey Nadyozhin

    Also popular now: