The title pretty much sums it up. Aside from DC, why are the properties of components often listed at 1 kHz? I keep hearing that it is a "standard", but why 1 kHz? This seems to be pretty common in the datasheets I've come across (primarily for ceramic disk capacitors), but have never seen the reason why 1 kHz was chosen. Also, I have a couple of hand-held capacitance meters that use a 1 kHz pulse to measure capacitance/ESR on components.

  • 2
    \$\begingroup\$ What sorts of components? Post links to datasheets of example components, please. \$\endgroup\$ Commented Jan 20, 2016 at 22:58

1 Answer 1


Typically characteristics of components are given at some frequency that makes sense for the application and/or puts the performance in a good light.

You are incorrect in giving a single frequency. Characteristics may be given at 120Hz, 1kHz, 10kHz, 100kHz, 1MHz, 1GHz or whatever. Electrolytic capacitors are often specified at 120Hz (full-wave rectified mains frequency) or 100kHz (in the same range as many switching power supplies).

1kHz is a common frequency for specifying certain components (such as data converters) used in audio applications, probably because distortion at 1kHz will result in audible harmonics, whereas no human can discern an audible difference between a 10kHz sine wave and a 10kHz square wave. On the darker side, 1kHz is also high enough that certain types of distortion are less apparent- those caused by thermal effects, for example, though that's more of an issue with amplifiers than with data converters.

Edit: In the case of ceramic disk capacitors, you've basically answered the question yourself - they state the test conditions including the frequency that their test instrument uses (1kHz), like your cap meter. Bias voltage is probably more important, and temperature for non-NP0 types. Note that loss is usually spec'd at a much higher frequency such as 1MHz.

  • 2
    \$\begingroup\$ And the reason that it's 1 kHz and not 900 Hz or 2.4 kHz is of course the same reason we use 100 nF for decoupling and 10 kohm for pull-ups: We like nice and round decimal numbers. \$\endgroup\$
    – pipe
    Commented Jan 20, 2016 at 21:38
  • \$\begingroup\$ @SpehroPefhany Thanks for the info, I've updated my question, as I am interested in, for instance, ceramic disk capacitors. If you have any additional info to add to your answer, please do. \$\endgroup\$
    – iwantmyphd
    Commented Jan 24, 2016 at 2:32
  • \$\begingroup\$ @pipe that argument works for using cheap/easy-to-find component values, but for a frequency, it seems less relevant. When the frequency is not that important for the target application, then I think the choice of 1kHz as a "standard" is probably because it's high enough for a stable reading to be given quickly and to be able to measure small capacitors, but low enough for the ESL of a big capacitor to be negligible. Once one manufacturer arbitrarily picks some value in the suitable range, others will likely follow, just so that all devices can be sure of measuring the same quantity. \$\endgroup\$ Commented Jan 24, 2016 at 16:54
  • \$\begingroup\$ @pipe by the way, although pullup value is not very important, one really oughtn't to choose decoupling capacitors with the only criterion being that they are of some common value. Nowadays a 100nF MLCC is available in a small enough package that it ought to work reasonably well in many applications, so at least that value can be most of the time be considered as a possibility. But it certainly isn't the case that any old 100nF capacitor will do. \$\endgroup\$ Commented Jan 24, 2016 at 17:09
  • \$\begingroup\$ @OleksandrR. Even 1uF is small enough and good in many cases. Of course a small 100nF X5R 6.3V MLCC cap might only be 35 or 40nF at 5V bias and 75°C, so the nice even '100nF' is a bit deceptive. \$\endgroup\$ Commented Jan 24, 2016 at 17:36

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.