6
\$\begingroup\$

I was calculating ESR for this ceramic capacitor, Samsung P/N CL21A226MQQNNWE.

Per the datasheet, DF is given as 0.1 max at 120 Hz.

Per formula, ESR = 0.1 / (2 * Pi * 120 * 22uF) = 6 Ohms.

What am I doing wrong, or is the ESR really 6 Ohms? That just seems way too high for a ceramic capacitor.

I feel a bit silly asking this, but hey, if I am confused someone else probably will be, too.

ADDENDUM Basically, the ESR for MLCC capacitors is surprisingly high at low frequencies. However, in my application, the ESR at higher frequencies is what was of interest. The graphs provided by several answers were super helpful and show how much the ESR changes with frequency. This is not something I knew about when I asked the question. I just knew that MLCC's supposedly had very low ESR (everybody knows this, so it must be true, right?). Anyway, now everything makes sense.

\$\endgroup\$
4
  • 2
    \$\begingroup\$ Your calculation is correct. And, ESR is higher at low test frequencies like 120 Hz. However, sometimes max limits on datasheets are way above what the actual part will measure due to tester limitations or lots of margin being added. \$\endgroup\$
    – John D
    Feb 26, 2018 at 17:36
  • 1
    \$\begingroup\$ X5R dielectrics are not that great. \$\endgroup\$
    – Andy aka
    Feb 26, 2018 at 17:42
  • 1
    \$\begingroup\$ At first I was thinking this was just a very conservative spec, but then I looked at some comparable high-capacitance ceramic and for the highest capacitance parts they give even higher DF values (e.g. 0.15), so I think it might actually be that bad. \$\endgroup\$
    – τεκ
    Feb 26, 2018 at 17:44
  • 1
    \$\begingroup\$ Looking at some Murata parts with comparable specs but with characteristic data available (ex), 2-3 ohms at 100 Hz is not unusual. \$\endgroup\$
    – The Photon
    Feb 26, 2018 at 17:52

4 Answers 4

Reset to default
2
\$\begingroup\$

This is correct only as a line f* diode bridge cap. Tan delta is std at 120Hz.

It is unwise to use as that*.

Unlike some or most e-caps this ceramic cap drops to 3 mohms near SRF ,~1MHz

enter image description here

Look for the PDF with my graph (re-search) hint Samsung site

Answer: char. Data.pdf http://www.samsungsem.com/global/support/product-search/mlcc/1205485_4290.jsp

\$\endgroup\$
3
  • \$\begingroup\$ I thought char data was "char" like text. Kind of stupid of me, but on the other hand, Samsung could make this a bit easier to find! Thanks, Tony! \$\endgroup\$
    – user57037
    Feb 26, 2018 at 19:38
  • 1
    \$\begingroup\$ char(dot) is rather brief isn't it. I wonder what the other files say? More treasure ... flash search engine. The Japanese know these parts and tricks but release to the ROTW years later. et tu Korea \$\endgroup\$
    – Tony Stewart EE75
    Feb 26, 2018 at 21:03
  • \$\begingroup\$ Samsung has moved things around. That link doesn't work any more. The data can be found here now: weblib.samsungsem.com/mlcc/… \$\endgroup\$
    – user57037
    Jun 16, 2021 at 6:26
1
\$\begingroup\$

There's an online calculator for Kemet caps Here that gives you an impedance and ESR plot.

Though your cap is from a different manufacturer, you can pick a similar Kemet cap like the one I picked below, 22uF 1206 X5R:

enter image description here

The calculator gives 1.5 ohms at 120Hz. With tester guardband and process variation 6 ohms looks reasonable if a little high. The graph shows that at SMPS frequencies the ESR is orders of magnitude lower.

\$\endgroup\$
2
  • \$\begingroup\$ My example cap is also in a smaller package (0805). So that may be one more reason why it is a bit higher. \$\endgroup\$
    – user57037
    Feb 26, 2018 at 18:27
  • 1
    \$\begingroup\$ The takeaway for me is that DF at 120 Hz is completely useless for determining ESR at, say 100kHz or 1MHz. A designer has to find a graph like this one to get that information. \$\endgroup\$
    – user57037
    Mar 1, 2018 at 6:08
1
\$\begingroup\$

DF can vary a lot between dielectric types. See this: -

enter image description here

It lists the poorer stability ceramic diectric DF as being 5% (Y5V) whereas the DF for NP0 is about 0.1%. Electrolytic DFs can be as high as 20%. Table acquired from here.

Regarding the total impedance of a 22 uF capacitor at 120 Hz (Xc = 60 ohms) and the ESR of 6 ohms, Z = \$\sqrt{60^2+6^2}\$ = 60.3 ohms!

enter image description here

As with any component, if the data sheet doesn't provide the performance above a certain frequency then don't make assumptions about it unless you are prepared to do extensize tests or there are generic data sheets from the supplier that gives higher frequency performance details across a range of components.

\$\endgroup\$
4
  • \$\begingroup\$ Drive-by downvoters. Oh well. \$\endgroup\$
    – user57037
    Feb 26, 2018 at 21:50
  • \$\begingroup\$ You can’t please everyone @mkeith lol \$\endgroup\$
    – Andy aka
    Feb 26, 2018 at 22:46
  • \$\begingroup\$ I am actually using the cap in the output filter of a DC-DC. The datasheet equations for calculating external compensation values require entering the output capacitor ESR. Not knowing any better, I calculated it using DF from the datasheet. But that is NOT the way to do it, because that DF is only valid at 120 Hz. Anyway, thanks for your answer. \$\endgroup\$
    – user57037
    Mar 1, 2018 at 6:27
  • \$\begingroup\$ This is good. Also if all they give is DF at 120 Hz , then it is a general purpose e-cap and the ESR*C=T will be greater than 100 us. If the spec gives ESR, then in most cases it will be a low ESR type with T < 10us which implies from rise time that the -3dB BW is 0.8/0.64 * 0.35 / 10us=370kHz \$\endgroup\$
    – Tony Stewart EE75
    Jun 16, 2021 at 12:55
0
\$\begingroup\$

Calculation may be right but the range does not apply for the capacitor type, it's quite common to reach such range mishaps.

\$\endgroup\$
1
  • \$\begingroup\$ This popped up because I edited it. But it is an old question, and, in my opinion, satisfactorily answered already. I do thank you for weighing in. But I recommend you answer newer questions which do not already have an accepted answer to help build reputation. \$\endgroup\$
    – user57037
    Jun 16, 2021 at 20:24

Your Answer

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

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