Author Topic: 10uF cap measures 6.2-7uF - Normal ?  (Read 2246 times)

0 Members and 1 Guest are viewing this topic.

Offline Signal32Topic starter

  • Frequent Contributor
  • **
  • Posts: 251
  • Country: us
10uF cap measures 6.2-7uF - Normal ?
« on: September 28, 2021, 05:45:16 am »
Hello,

Ordered a bunch of 0402 10uF/10V capacitors from AliExpress. These ones: [DigiKey].
Measuring them with my U6-61E they come up at 6.1-7.1uF. The tolerance spec says +/- 20%, so I would expect them to measure in at at least 8uF.
The multimeter seems good, measuring other caps you get exactly the values that you would expect. Ex: 2.19uF for 2.2uF, 0.99uF for 1uF, etc ....

Is this normal / expected with smaller caps such as this ? Is this just a bad batch ?

Thanks !
 

Offline Chalcogenide

  • Regular Contributor
  • *
  • Posts: 156
  • Country: it
Re: 10uF cap measures 6.2-7uF - Normal ?
« Reply #1 on: September 28, 2021, 06:02:22 am »
MLCCs suffer of a very strong DC bias dependence on the capacitance value. That specific part number is already 60% low at just 2V of applied voltage! http://weblib.samsungsem.com/mlcc/mlcc-ec-data-sheet.do?partNumber=CL05A106MP5NRN . Portable DMMs typically measure the time required to charge the capacitor to a given voltage to compute its capacitance, but this means that an MLCC suffering of DC bias effect will read low, and your part appears to be really quite bad in this regard.
 
The following users thanked this post: tooki, Rod

Offline The Electrician

  • Frequent Contributor
  • **
  • Posts: 743
  • Country: us
Re: 10uF cap measures 6.2-7uF - Normal ?
« Reply #2 on: September 28, 2021, 06:08:12 am »
Hello,

Ordered a bunch of 0402 10uF/10V capacitors from AliExpress. These ones: [DigiKey].
Measuring them with my U6-61E they come up at 6.1-7.1uF. The tolerance spec says +/- 20%, so I would expect them to measure in at at least 8uF.
The multimeter seems good, measuring other caps you get exactly the values that you would expect. Ex: 2.19uF for 2.2uF, 0.99uF for 1uF, etc ....

Is this normal / expected with smaller caps such as this ? Is this just a bad batch ?

Thanks !

Watch this: https://www.eevblog.com/forum/projects/mlcc-capacitors-and-my-experiments-of-dc-bias-on-capacitance/msg2601327/#msg2601327

You shouldn't be using a multimeter to measure capacitors (particularly ceramic caps)--use an LCR meter (and for measuring ceramic caps the meter should have an ALC mode).
 

Offline Signal32Topic starter

  • Frequent Contributor
  • **
  • Posts: 251
  • Country: us
Re: 10uF cap measures 6.2-7uF - Normal ?
« Reply #3 on: September 28, 2021, 06:36:07 am »
Makes sense, thanks !
 

Offline Gyro

  • Super Contributor
  • ***
  • Posts: 9410
  • Country: gb
Re: 10uF cap measures 6.2-7uF - Normal ?
« Reply #4 on: September 28, 2021, 08:48:42 am »
10uF is a very large value for an 0402 package. I would expect that it would have 'compromises' several areas, including pretty extreme voltage coefficient, as you have found. I doubt the Ali ones would have as close a tolerance as the Digikey ones you referenced either, I wouldn't be at all surprised at -50%.
Best Regards, Chris
 

Online wraper

  • Supporter
  • ****
  • Posts: 16792
  • Country: lv
Re: 10uF cap measures 6.2-7uF - Normal ?
« Reply #5 on: September 28, 2021, 09:02:36 am »
There is such thing as capacitance aging. Capacitors made with type 2 and 3 ceramics start losing their capacitance since the moment they are produced. Heating during reflow resets the process, capacitance is regained and starts dropping again. Considering that capacitors stayed on shelf quite a while, measured capacitance seems reasonable for 10uF cap. You can heat it again with hot air or soldering iron, let it cool and measure again.
If capacitor uses type 3 ceramic, like Z5U or Y5V, then you will also have strong temperature dependence. Just a heat from your fingers may cause a significant portion of the capacitance to be lost.
Capacitance voltage dependence should not cause much of measurement difference unless your meter causes significant DC bias.
 

Online wraper

  • Supporter
  • ****
  • Posts: 16792
  • Country: lv
Re: 10uF cap measures 6.2-7uF - Normal ?
« Reply #6 on: September 28, 2021, 09:12:24 am »
PDF about ageing process.
 
The following users thanked this post: thm_w, tooki, RoGeorge, iMo, Vovk_Z, mawyatt

Offline RoGeorge

  • Super Contributor
  • ***
  • Posts: 6146
  • Country: ro
Re: 10uF cap measures 6.2-7uF - Normal ?
« Reply #7 on: September 28, 2021, 09:56:09 am »
PDF about ageing process.

I didn't know that, thank you!   :o
Quote
It’s not uncommon for customers to receive capac-itors and have them sit on the shelf for many months or even years before mounting to a PCB. At that point, the capacitor is far along the aging curve and may be reading lower capacitance than the specifi-cations. However, when the capacitor is mounted to the PCB, it will go through a solder reflow process in which it will see temperatures above the curie point.  By going above the curie point, the capacitor will go through the crystal structure change again and reset back to its original aging rate start time (t = 0).  The process that the capacitor goes through when resetting the aging process is called deaging.

Offline Vovk_Z

  • Super Contributor
  • ***
  • Posts: 1395
  • Country: ua
Re: 10uF cap measures 6.2-7uF - Normal ?
« Reply #8 on: September 28, 2021, 04:49:37 pm »
Is this normal / expected with smaller caps such as this ? Is this just a bad batch ?
If you want 'real' 10 uF you should buy at least 1210 size ones (X7R, but not Y5V).
 
The following users thanked this post: Rod

Offline TimFox

  • Super Contributor
  • ***
  • Posts: 7934
  • Country: us
  • Retired, now restoring antique test equipment
Re: 10uF cap measures 6.2-7uF - Normal ?
« Reply #9 on: September 28, 2021, 05:07:52 pm »
When dealing with the inherent properties of a dielectric in a capacitor (or a non-linear conductor in a resistor), the independent variable is the E field (voltage gradient), so the non-linearity will be worse in smaller packages (for a given applied voltage across the terminals).
 

Offline dietert1

  • Super Contributor
  • ***
  • Posts: 2018
  • Country: br
    • CADT Homepage
Re: 10uF cap measures 6.2-7uF - Normal ?
« Reply #10 on: September 28, 2021, 05:39:16 pm »
I have seen exactly the same -30 % low capacitance with MLCC. It's strongly temperature dependent.
If you want a capacitor with +/- 10% or +/- 5 % tolerance, it's a film cap.

Regards, Dieter
 
The following users thanked this post: Rod

Online Kleinstein

  • Super Contributor
  • ***
  • Posts: 14071
  • Country: de
Re: 10uF cap measures 6.2-7uF - Normal ?
« Reply #11 on: September 28, 2021, 08:43:32 pm »
10 µF in  0402 size is rediculous small. Don't expect more miracles from such a small part.

The Class 2/3 ceramics have rather non ideal properties. When needing a more ideal capacitor, more modern C0G ceramic caps are a real alternative to film caps, though only for smaller values (e.g. < 50 nF). Even in a relatively small package C0G can be pretty good, with low leakage and low loss.
 
The following users thanked this post: Rod

Offline bobbydazzler

  • Regular Contributor
  • *
  • Posts: 137
  • Country: au
Re: 10uF cap measures 6.2-7uF - Normal ?
« Reply #12 on: September 29, 2021, 12:47:53 pm »
Yes the smaller the package the worse the capacitor is affected by dc bias.  Also they're from Alixpress so may not even be genuine x5r parts, they could be y5v.  I got some 47uf 'x7r' 1210 mlcc capacitors from ebay that seemed to be genuinely 47uf but as soon as they warmed up they had almost no capacitance, so probably y5v or worse.
 
The following users thanked this post: Rod

Offline The Electrician

  • Frequent Contributor
  • **
  • Posts: 743
  • Country: us
Re: 10uF cap measures 6.2-7uF - Normal ?
« Reply #13 on: September 29, 2021, 03:03:55 pm »
Capacitance voltage dependence should not cause much of measurement difference unless your meter causes significant DC bias.

Even without DC Bias, the measurement difference can be quite large depending on the value of the AC excitation voltage applied by an LCR meter.  See the graph labeled "AC voltage characteristics" in this spec sheet: http://weblib.samsungsem.com/mlcc/mlcc-ec-data-sheet.do?partNumber=CL05A106MP5NRN

Also see this video:

 
 

Online Kleinstein

  • Super Contributor
  • ***
  • Posts: 14071
  • Country: de
Re: 10uF cap measures 6.2-7uF - Normal ?
« Reply #14 on: September 29, 2021, 04:07:00 pm »
With such a nonlinear part, does it still make sense to quote a capacitance for this. If nearly looks like this is more like a square hysteresis curve.
Getting the capacitance down nby 60% when measuring with a very low AC voltage is a really odd point. So the capacitance is no the small signal capacitance, but rather specific for a given amplitude.  Samsung uses 0.5 V in the DS and Kermet use 1 V -  so not even the same standard.  :wtf:
Maybe they should show hysteresis curves.

With so nonlinear parts it can also matter how the meter is measuring (sine voltage or sine current) or doing the math.
 
The following users thanked this post: Mecanix

Offline dietert1

  • Super Contributor
  • ***
  • Posts: 2018
  • Country: br
    • CADT Homepage
Re: 10uF cap measures 6.2-7uF - Normal ?
« Reply #15 on: September 29, 2021, 04:35:04 pm »
Like you wrote before it depends on the size. Just wired a 0805 4.7 uF MLCC and it reads 4.5 uF after soldering while still a bit warm and 4 uF at room temperature. This is with an old HP 4262A at 120 Hz, 10 Vrms idle, 370 mVrms on the cap. With higher test frequencies, where the remaining voltage gets smaller, it measures as low as 3 uF.
In the past i measured 0805 10 uF MLCCs many times, always wondering why they measure about 7 uF at 120 Hz. So 0603 or 0402 should be much worse. A pitty the Kemet expert doesn't tell the size, anyway a good lesson.

Regards, Dieter
« Last Edit: September 29, 2021, 06:04:25 pm by dietert1 »
 

Online Siwastaja

  • Super Contributor
  • ***
  • Posts: 8106
  • Country: fi
Re: 10uF cap measures 6.2-7uF - Normal ?
« Reply #16 on: September 29, 2021, 04:40:47 pm »
Remember there are a lot of bypass applications for low core voltages like just 0.8V, where a lot of capacitance is needed with small parasitic inductance, i.e., small package.

If anything, rated voltage is to be taken with grain of salt. It's guaranteed not to "blow up" at 10V. For useful capacitance, that is, near rated 10uF, this part is designed to bypass Vcore = 1.2V and less.
 

Offline TimFox

  • Super Contributor
  • ***
  • Posts: 7934
  • Country: us
  • Retired, now restoring antique test equipment
Re: 10uF cap measures 6.2-7uF - Normal ?
« Reply #17 on: September 29, 2021, 05:52:21 pm »
There is a mature theory of non-linear electronics, where a non-linear capacitor is defined by its hysteresis and functional relationship V(Q), which is not V = Q/C, in general.
See M Hasler & J Neirynck "Nonlinear Circuits" (Artech House).
 

Offline dietert1

  • Super Contributor
  • ***
  • Posts: 2018
  • Country: br
    • CADT Homepage
Re: 10uF cap measures 6.2-7uF - Normal ?
« Reply #18 on: September 29, 2021, 09:33:42 pm »
After testing a 4.7 uF MLCC i left the HP 4262 with 6 V DC on for several hours and when i happened to look again, capacitance measurement was a mere 1.2 uF! After turning off DC, measurement went up to 3 uF and appears stable or rising slowly. Parts are labeled: UMK212BBJ475KG-T (Multilayer Ceramic Capacitors MLCC - SMD/SMT 0805 50VDC 4.7uF 10% X5R) and their capacitance can be anywhere between 25 % and 100 % of nominal value. "It depends."

Regards, Dieter
 

Offline TimFox

  • Super Contributor
  • ***
  • Posts: 7934
  • Country: us
  • Retired, now restoring antique test equipment
Re: 10uF cap measures 6.2-7uF - Normal ?
« Reply #19 on: September 29, 2021, 09:40:20 pm »
Following Siwastaja, the "10 V" rating is merely for safety, where the manufacturer guarantees the device can withstand 10 V without failure or excessive current.  It is not a recommended operating voltage, but the details require knowledge of your own requirement and the datasheet curves.
 

Online Siwastaja

  • Super Contributor
  • ***
  • Posts: 8106
  • Country: fi
Re: 10uF cap measures 6.2-7uF - Normal ?
« Reply #20 on: September 30, 2021, 05:49:06 pm »
OTOH such 10V part may happily take 50V without blowing up in any way, ever. It just loses even more capacitance under that bias.

MLCC voltage ratings are even more arbitrary than their capacitance ratings.

My preferred method, when large amount of capacitance is required, is to find a few candidates, then look up actual capacitance under the applied voltage, then sort for $/uF to get the best deal. Rated capacitance or rated voltage are neither very useful here, they just offer starting points for sorting/finding. A 2.2uF part might offer 90% of the capacitance of the 4.7uF part on the same series, at half the price. I.e., both are roughly 1uF actual.

Of course, if it's a professional job I won't use a part at higher than rated voltage; for a prototype, I might, and have used some 25V rated parts for 30-something volt buses.

With low-voltage MLCCs (like 25V and down), voltage derating is really not needed, if I have a 10V bus and a 10V part happens to offer good capacitance at 10V DC bias, I can use it, no need to go for a 16V part.
« Last Edit: September 30, 2021, 05:50:57 pm by Siwastaja »
 

Offline CatalinaWOW

  • Super Contributor
  • ***
  • Posts: 5170
  • Country: us
Re: 10uF cap measures 6.2-7uF - Normal ?
« Reply #21 on: September 30, 2021, 06:04:26 pm »
All of the comments about why this measure low are related to measurement.  The real question is how these parts will operate in your circuit.  If your circuit operates down around two volts they don't meet your stated need.  If they are operated at a significantly higher voltage they may.  Component selection always involves many factors and trades.

Whether the low reading justifies a return for credit is another whole question.  Gets into what the data sheet says they will do vs what they are actually doing.  If they are performing to data sheet, then it's your problem for not understanding what you are buying.  The fact that you were misled by a simple value used to identify the part is irrelevant.   If they don't you can try to get a refund, and an honest source will give it to you. 
 

Online Kleinstein

  • Super Contributor
  • ***
  • Posts: 14071
  • Country: de
Re: 10uF cap measures 6.2-7uF - Normal ?
« Reply #22 on: September 30, 2021, 06:20:40 pm »
Using these caps with a relatively low voltage, like 2 V is normal, but the normal case is also to have a low ripple amplitude, usually way less than 0.5 or 1 V RMS. So in real world one would not see the nominal capacitance, but considerably less, even without the degradation from high DC.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf