0
\$\begingroup\$

I faced a problem of measuring capacity in ceramic capacitors.

Measurements did with Smart Tweezers ST5S. It has two output voltage levels that can be selected: 0.45 Vrms and 0.9 Vrms.

The capacitor is 22µF ±20% 6.3V X5R 0603 (1608 Metric).

Tolerance limits are from 17.6 µF to 26.4 µF.

According to specification test methods are:

The capacitance/D.F. should be measured at a frequency of 1±0.2kHz and a voltage of AC1±0.2V(r.m.s.). (Test Methods).

When I set up output voltage level on SmartTweezers:

  • 0.45 Vrms -- capacitance 17.14µF (out of tolerance limits)
  • when 0.9 Vrms -- capacitance 18.17µF

Why we have so difference? Maybe there is some formula which represents depending on: test measurement voltage VS capacitance?

\$\endgroup\$
3
  • 1
    \$\begingroup\$ EEVblog #626 - Ceramic Capacitor Voltage Dependency \$\endgroup\$
    – Jeroen3
    Commented Mar 1, 2018 at 8:03
  • 1
    \$\begingroup\$ Consider this: the capacitance changes depending on the applied voltage. You are applying a changing voltage to it during the test. You are therefore measuring the average capacitance through all the variations. \$\endgroup\$
    – JRE
    Commented Mar 1, 2018 at 9:20
  • 2
    \$\begingroup\$ There are detailed (typical) data for your particular capacitor: psearch.en.murata.com/capacitor/product/… Capacitance depends (among others) on both DC and AC voltage. I do not believe there is any universal formula or so. Anytime you use class 2 dielectric ceramic caps (anything other than NP0/C0G or so), be sure to check manufacturer's specs and/or assume actual capacitance can be many times less than nominal. \$\endgroup\$
    – Martin
    Commented Mar 1, 2018 at 10:36

3 Answers 3

Reset to default
1
\$\begingroup\$

That's the DC Bias Characteristic and is the reason for derating when a ceramic capacitor is used. That means that among ceramic capacitors, the capacitance, especially of capacitors classified as high dielectric constant (B/X5R, R/X7R characteristic), may differ from the nominal value when a DC voltage is applied.

enter image description here

The effect is linked to the crystal structure of the capacitor's dielectricum.

For more details regarding this topic there's e.g. a nice article on the Murata website: https://www.murata.com/support/faqs/products/capacitor/mlcc/char/0005

\$\endgroup\$
2
  • 1
    \$\begingroup\$ But I told about AC voltage! And I guess that in Smart Tweezers I choose the AC output voltage. And in Test Methods indicated AC Voltage level. \$\endgroup\$
    – KMM
    Commented Mar 1, 2018 at 7:52
  • \$\begingroup\$ @KMM 1 kHz is still a fairly low frequency, so you'll get similar effects as you would get with DC. \$\endgroup\$
    – jms
    Commented Mar 1, 2018 at 8:04
1
\$\begingroup\$

I would expect the capacitance to be higher at the lower voltage test, and you are observing the opposite. The capacitance will change significantly during the AC cycle, even with the +/-1.2V peak voltage you are applying. Maybe that's confusing it.

Maybe your "smart tweezers" need an IQ test. Try testing a 10-20uF film cap (which should not change appreciably with voltage) and see what you get.

\$\endgroup\$
1
  • 1
    \$\begingroup\$ Or a new battery. I've gotten really bad results before because the battery was near its end and couldn't supply the expected voltage anymore. \$\endgroup\$
    – Goswin von Brederlow
    Commented Mar 1, 2018 at 12:08
1
\$\begingroup\$

The problem that you are presenting here has relation with the characteristic of DC bias of the capacitors of type MMCs. When choosing a dielectric x5r and a capacitance of 22uF and an encapsulation 0603 you will have a poor performance, this is due to the variation of the capacitance according to the applied dc voltage. A simple way to solve this problem is that you change for a larger encapsulation capacitor I suggest one of type 0805 or even 1206 with dielectric X7R. This should improve the performance. It is vital that you review the datasheet or the online tools of each manufacturer. For example Murata and Kemet have two online software that allow to verify the DC bias characteristic. Although two capacitors have the same specification both in practice can present different values in their capacitance.

\$\endgroup\$

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