Question about ceramic caps readings

[eh la bas ma]

Hello,

Every time i use my DMM (VC97A) to read a ceramic cap, it always reads a higher value than the value written on the cap.

For example, 47p will read 72p. Every 47p read around 70p. Same thing goes for all small value ceramic  capacitors. 56p reads 80p, etc

I wonder which value should i trust ? The one written on the part, or my DMM ?

Thank you for your help !

[antonis]

I wonder which value should i trust ? The one written on the part, or my DMM ?

Average, perhaps..??

[FiveseveN]

I think you're going to be surprised if you try to measure the capacitance of "nothing", i.e. the probes shorted together.
Tools have limitations.

[eh la bas ma]

I think you're going to be surprised if you try to measure the capacitance of "nothing", i.e. the probes shorted together.
Tools have limitations.

I'm not sure i understand what you mean. Probes are crocodile clips, and i do read a stable value, no oscillations, etc. I carefully spread the cap's legs in order to avoid shorts...

What about you ? Do you often read something close to the written value when you measure a 47p ? No big difference ?

Am i the only one ?

[FiveseveN]

Take the crocodile clips and connect them together. No capacitor, just a dead short. What value does it display?

[Dormammu]

Hello,

Every time i use my DMM (VC97A) to read a ceramic cap, it always reads a higher value than the value written on the cap.

For example, 47p will read 72p. Every 47p read around 70p. Same thing goes for all small value ceramic  capacitors. 56p reads 80p, etc

I wonder which value should i trust ? The one written on the part, or my DMM ?

Thank you for your help !

In such small values — a lot of things can affect the accuracy of the measurement.
On my DMM the lower limit is 20nF, so I don't even try to measure such small values.
But the capacity of a guitar cable shows quite well, for example.

[amptramp]

Leave the probe leads in parallel close to each other with no capacitor attached.  This should give you a reading of the capacitance of the leads.  This will be added to every reading, so subtract this value and you should be OK.

I used to have an EICO 955B bridge that read about 70% of actual value at any capacitance because the internal bridge capacitor was out of tolerance.  Make a few measurements and determine whether you have a fixed amount added to the reading or a fixed percentage.  Then you can separate these effects from each other.

[GibsonGM]

Meter resolution as suggested.  I usually (always?) can't hear much difference between 56p and 100p in most any place they're used (such as opamp feedback paths).  I go with their marking and don't bother w/measurements.

[eh la bas ma]

Thanks for your replies !

Take the crocodile clips and connect them together. No capacitor, just a dead short. What value does it display?

First I get a blank screen, it switches to mF, then it reads 0.L

Leave the probe leads in parallel close to each other with no capacitor attached.  This should give you a reading of the capacitance of the leads.  This will be added to every reading, so subtract this value and you should be OK.

If i leave the probes alone, unconnected, it slowly oscillate between 0.007nF and 0.009nF.

I think that's 9p ? so i'll substract 9p to every values and i guess it will be the right number.

Thank you very much !

[mozz]

Cheap meter, don't expect too much at lower ranges.

Capacitance    40nF/400nF/4uF/40uF/400uF/4mF/40mF    ±(2.5%+5)

[R.G.]

Never trust your meter completely. Always assume that it's an indicator, and not completely reliable, and check the meter for reasonable accuracy and correct hookup whenever you get some reading that looks the slightest bit funny.
Measuring small things, like millivolts, pF caps, microHenry inductors, and sub-10-ohm resistors is fraught with errors because of the tiny differences being measured. In these realms, the probe wires and contacts themselves can contribute relatively huge offsets. The advice to measure just your probes is really, really good.
The digital meter gives a glossy appearance of definite-ness and expected accuracy that covers up the real world underneath the measurements. It fools you into thinking it's correct, just like anything printed out of a computer seems more correct than someone telling you or writing out the same thing by hand.
A digital meter simply can not be more accurate than +/- one smallest digit, because the internal analog to digital converter had to decide whether to display a higher number or a lower one if the reality was in the middle. On top of that, you have to guess at the accuracy of the internal analog circuits that did the signal work before the A-D converter. It's common for DVMs to have specifications of +/- a few percent, and then +/- one, two, or more (!) least digits. That's for mid-range and better meters. Companies competing for selling the cheapest possible meters have been known to ... um... make mistakes 8-) in their specs.
If your meter is telling you things that seem at odds with the real world, either the "real world" or the meter OR BOTH may be wrong. Learn the limits of your meters, and be prepared to investigate further. 

[Phend]

Read the manual for your meter for it's capacitance accuracy / resolution.
My el cheapo dmm isn't good until 150p, the manual isn't accurate either!
Even at 150p it is 20%, however my component tester is good at 47p. Quite accurate.
With 22p is is OL but put two 22p in // and it reads 42p.

Below is pic of el supremo cheapo tester that I use a lot.
Good for 40p.

[ElectricDruid]

Tiny capacitance values are best avoided all around, in my view. It's so easy for some stray cpacitance to start adding up to a few tens of pFs that I don't regard anything under 100pF as anything but a bit of a guess. Even if you put a 22pF/47pF cap in a circuit, there might be another 10pF of capacitance lurking between two traces unless you're careful about the layout. Those caps are probably +/-20% to start with, and then you're adding another 50%/20% of error. Much better to design it for a 220pF cap so that a 10pF variation isn't such a big deal. This is a habit I got into through designing on a breadboard where the problem is much worse, but it's served me well since then too.

[Mark Hammer]

I find it perplexing that we can stick several billion transistors in less than a square centimeter of plastic, for memory, but are unable to produce capacitors of similar size with a true value that corresponds to the value printed on the side,  What's the holdup?  Do the two sorts of devices impose different sorts of production requirements?

[antonis]

I wonder why we don't use MosFets as voltage controlled variable capacitors..

[Dormammu]

I find it perplexing that we can stick several billion transistors in less than a square centimeter of plastic, for memory, but are unable to produce capacitors of similar size with a true value that corresponds to the value printed on the side,  What's the holdup?  Do the two sorts of devices impose different sorts of production requirements?

Capacitors require a significant area to store charge. And this is — the dielectric area.

[Phend]

The smallest cap I have seen is 5pF (4.7) in the Rebote 2.5 Delay. Maybe I have seen them some where else too.  In the Rebote it is located across pins 6 and 7 of the TL072 chip.
How long would parallel traces or a couple of parallel wires have to be to effect this capacitance, lets say 10% ? Assuming the cap is 5pF and the traces / wires are close together.

[Rob Strand]

Hello,

Every time i use my DMM (VC97A) to read a ceramic cap, it always reads a higher value than the value written on the cap.

For example, 47p will read 72p. Every 47p read around 70p. Same thing goes for all small value ceramic  capacitors. 56p reads 80p, etc

I wonder which value should i trust ? The one written on the part, or my DMM ?

Thank you for your help !

It's normal.  The DMM's "raw" reading on the display is *not* the true part value for small capacitances.

The DMM circuit and leads have a small amount of "stray" capacitance.  That capacitance appears in parallel with the capacitance of the external capacitor.  The DMM reads the total.

For accurate measurement you need to do two measurements:
- measurement with leads *open* and spaced the same as when measuring the part
  Displayed value = C_open

- measurement with part
  Displayed value = C_total = C_part + C_open

To get the true part value you need to subtract the stray capacitance,

  C_part = C_total - C_open

You get a similar problem measuring low ohms values as well.  When you short the leads you might get 0.1 ohm to 1.0 ohm due to the meter, meter fuse, and DMM lead resistance (R_short).  That resistance appears in series with the true part value.  So in this case

  R_part = R_total -  R_short

Many LCR meters have way to zero the inductance, capacitance and resistance by opening and shorting the leads during a calibration process.
Note:
47p read 70p   => difference = 23p
56p reads 80p => difference = 24p

So you should see readings with the leads open in this range of values.  The closer the DMM leads are together or more they are twisted the higher the Copen value.   If you just poke the cap into cap terminals on the front of the meter then the value should be more consistent - as there are no DMM leads moving about to create variation in the open lead capacitance.
Go here,
https://savvycalculator.com/parallel-wire-capacitance-calculator/

Spacing 15mm  (0.015m)
Diameter 1mm (0.001m)
Wire length 1m
Relative permittivity:  1.1             (guess: air 1, plastic sheath 4)

Calculated capacitance of wires = 23pF

When I measure some cheap DMM leads just sitting naturally together (not 15mm) I get 12pF, moving closer together 15pF, light twist for even spacing18pF.

[ElectricDruid]

The smallest cap I have seen is 5pF (4.7) in the Rebote 2.5 Delay. Maybe I have seen them some where else too.  In the Rebote it is located across pins 6 and 7 of the TL072 chip.
How long would parallel traces or a couple of parallel wires have to be to effect this capacitance, lets say 10% ? Assuming the cap is 5pF and the traces / wires are close together.

Here's a calculator to work it out:

https://savvycalculator.com/parallel-wire-capacitance-calculator/

Sticking in 0.001 (1mm between tracks), 0.0005 (0.5mm "diameter" track), and 0.02 (2cm track length) comes back with 1.6e-12, so 1.6pF. A track is nowhere near as much conductor as a 0.5mm diameter wire though, so then I reduced that to 0.0001 (0.1mm diameter wire equivalent) and that bumps the pF up to 4.83pF.

So even a couple of centimetres of close parallel tracks could totally overwhlem a 5pF capacitance, especially if the tracks are narrow.


<edit>Sorry Rob, in your long post I missed that you'd posted the same link.

[Phend]

Testing a Rebote on a breadboard, with 0.1 inch (.00254 m) spacings. And .5 inch long (.0127) should not be a problem.
 Very Small stuff these pico's. Why are they there ?