Distinguishing ceramic from film capacitors electrically?

[mcovington]

I see that several microcontroller data sheets recommend ceramic bypass capacitors, presumably because of their lower ESR.

I have a few dozen 0.1-uF capacitors and don't know which kind they are.  They are blue and plastic-encased.  Obviously the usual green or red film capacitors are easy to distinguish from tan ceramic ones, but these aren't either one.

Is there a practical electrical test to tell me?  (Do ESR meters read low enough?)   Or should I cut one open and look at it under the microscope?

[Kleinstein]

There are quite different types of ceramic and film caps.  The good thing here is that the main ceramic type used for decoupling is class 2 ceramics like X7R and similar. These capacitors are relatively easy to identify: there capacitance is quite a bit voltage dependent.  With film caps the capacitance is essentially not effected by a DC voltage.

So one could build a simple oscillator (e.g. 74HC14 + resistors + cap) and see if the frequency changes with voltage at the cap:
It should be something like <0.01% change for a film cap and some >1% change for X7R if a reasonable voltage is applied. 2 Caps in series and a high resistor to apply the voltage (e.g.  0 V or 1/2 the supply) should do the trick.

The main reason they suggest ceramic is that these are small and cheap - if small enough it would be OK to use a film cap too.

[Yansi]

You are not probably going to distinguish C0G/NP0 ceramic from a decent foil one - easily. 

But what might give, is if you'd stick the cap to a VNA, to measure the impedance. Foil type shall have higher parasitic inductance. But still - in some cases such difference may not be that high.

//EDIT: Also, comparing the leakage current may lead somwhere (ceramic should have higher) and dielectric absorption. Polystyrene shall be superior.

[nfmax]

Even simpler - most ceramic capacitors around the 100nF mark will use a dielectric like X7R that is strongly temperature sensitive, i.e. not C0G Pop one in the capacitance meter and monitor the value (not the ESR) as you warm it up. If it decreases significantly as the temperature rises, it's ceramic!

[bob91343]

One of the distinguishing features of ceramic disc capacitors is the low ESL.  This results in a high self resonance frequency, important when working with high speed components.

A dis ceramic capacitor is basically a slice of ceramic with leads bonded on the faces.  If done well, there is very little opportunity to create inductance.

One should also realize that ceramic can be piezoelectric, causing circuit noise when subjected to movement.

Multilayer ceramic capacitors are a different animal.

[Gyro]

Regardless of the electrical methods, your OP blue plastic cased ones will be film type (assuming that they're rectangular). Hopefully there should be some markings, eg MKP, MKS etc to narrow down the film type.

[T3sl4co1l]

Easy: since we're talking small values for bypass, film caps are bigger!

The ESL for either kind is essentially lead length.  (Foil connection is not done at a point, but by spraying metal onto the entire exposed edge of the roll.)

Also, unless the ceramics are C0G or similar, they will be nonlinear so the capacitance will drop noticeably under modest bias -- maybe 20-100V.  C0G are usually distinguished by marking as such, or maybe a black dot on top, and are also very uncommon (and expensive!) in large values (>10nF?) and low voltages.

Tim

[SeanB]

Hit it with a hammer, and look at the shattered remnants. if it is all brittle it was ceramic in a case, but a foil type will have some plastic film with metallisation on it, and that is easy to see under low magnification of the remains.

[Yansi]

The ESL for either kind is essentially lead length.  (Foil connection is not done at a point, but by spraying metal onto the entire exposed edge of the roll.)

Also, unless the ceramics are C0G or similar, they will be nonlinear so the capacitance will drop noticeably under modest bias -- maybe 20-100V.  C0G are usually distinguished by marking as such, or maybe a black dot on top, and are also very uncommon (and expensive!) in large values (>10nF?) and low voltages.

Tim

Almost all foil caps these days yes, but there are heaps of older foil types without the side metallization on the side of the roll, so beware.

I have never seen a COG cap marked with a dot. Certainly not a surface mount one. Do you have any photos of those?

[Kleinstein]

Most foil caps are more rectangular shape.  The tricky shape is a capacitor dipped in some kind of epoxy: these may be tantalum (usually marked for polarity), ceramic (especially X7R or similar) but also foil type. Still the shape my give away the type.

NP0 ceramic is rather difficult to distinguish from a good foil type by the electrical performance. 100 nF NP0 is high value and would be rare and large, unless rather new as a MLCC type. Leakage, loss and DA can be well comparable to PP film caps. The temperature dependence of NP0 ceramic should be less than most foil types.

[T3sl4co1l]

I'm assuming the context is THT, since OP mentioned "plastic cased", which I'm guessing are dipped since the boxed are almost always film (not sure right now, if I've seen boxed ceramic?)...

Chips are easy, films are shiny on top, while ceramics are ceramic of whatever white to drab color; neither is almost ever marked.  (Have rarely seen ceramic chips with markings; I've seen film chips at all about as often as marked ceramics...)

Anyway, here's a variety of markings from my bin --



Top: Vishay/BC 1nF (50V?) C0G, no identifier, good luck figuring it out if you salvaged it.

Bottom, left to right: EPCOS/TDK 22p, generic 10p C0G, someone (possibly Xicon, I forget) 33p 50V C0G ("NP0" being the old designation, still seen frequently today), and 180p 63V TDK.

I don't know if the colored top is actually standard or listed anywhere, but I've seen it often enough to suspect it's what they mean.

Tim

[MagicSmoker]

I'm assuming the context is THT, since OP mentioned "plastic cased", which I'm guessing are dipped since the boxed are almost always film (not sure right now, if I've seen boxed ceramic?)...

There are, indeed, boxed MLCC. The infamous CK05 series comes to mind. E.g. - like this little doodad: https://www.newark.com/kemet/ck05bx104k/ceramic-capacitor-0-1uf-50v-x7r/dp/87F4662

***

To the OP: polyester film and the X7R ceramic dielectric are roughly comparable, while polypropylene film is quite similar to the NP0 ceramic dielectric (as far as AC losses and ESR go). NP0 has a much higher usable temperature, however (125C vs. 85C).

[schmitt trigger]

It would help if you posted some photos.

But by your very brief description, they likely are stacked film types.

One thing that ceramic capacitors have, specially the MLCC types, is that they have very high volumetric efficiencies. For instance, one can get a 1uF, 10V device in a 0201 SMT package!

[David Hess]

In a decoupling application I would not worry too much about the difference.  Film capacitors work just as well in decoupling applications.  Ceramic capacitors are less expensive now but in the past that was not always the case.

[T3sl4co1l]

Oh yeah, of course, I've got a bunch of M39014 style ceramics.

Tim

[mcovington]

OK, the ones I'm curious about are the blue ones here.  (These are two from the same batch, showing both sides.)  The red one is obviously film.

The blue one looks a lot like a Murata ceramic capacitor, but I can't find "A5K" in the catalogue.

[Yansi]

MLCC ceramic.

[schmitt trigger]

Agree, MLCC ceramics with leads.

[mcovington]

How do you know?

BTW, I just did some testing.  Temperature stability is good (capacitance is down 1% when heated too hot to touch with a hot air gun) and there is no appreciable leakage even at 600 volts.

Can anyone identify the manufacturer?

[schmitt trigger]

As mentioned previously, a hunch based on  volumetric efficiency.
But could be wrong.

But if you are willing to sacrifice a device in the name of knowledge, cut one in half.

[David Hess]

How do you know?

I can tell from the form factor and size for a 0.1 microfarad part.  A disc ceramic or film would be larger at that value.  The film example shown is a polyester stacked metal film capacitor; they work fine in place of ceramic capacitors in decoupling and filtering applications.

Both of those are constructed by welding leads to a common rectangular surface mount part and then dipping it in epoxy.  If you heat the part and twist the leads, they will separate exposing the inside construction.

[MT]

I wonder how much of fan, harddisk etc vibrations creeps into pow supply via SMT decoupling cap's? Any studies around?

[T3sl4co1l]

"A5K" sounds familiar and ceramicy, though I don't know what it means (it's not a type code).

The shape suggests some odd ceramic, but they could be stacked film as well.

The tempco is definitely(?) too low for X7R or worse.  If they're PET, they'll have somewhat higher losses than PP or C0G, which you might perform an experiment on.

You'd do that by setting up a few watts from a signal generator or oscillator, and tune a resonant tank to it.  Most of the losses will probably be in the inductor, but there should be enough that you can feel some in the cap, and perhaps compare to known types.

I wonder how much of fan, harddisk etc vibrations creeps into pow supply via SMT decoupling cap's? Any studies around?

A few mV, give or take.  Yes, search for "ceramic capacitor microphonic".

Tim

[MT]

A few mV, give or take.  Yes, search for "ceramic capacitor microphonic".
Tim

I im aware of the standard piezo microphonics what im wondered about was "studies" made on fans and hardisk vibrations. Surely this depends om cap material, mounting material PCB, distance of vib source, microphonic bandwidth, resonant frequencies of/on the board, etc, etc, etc.

Even in Daves EVVBLOG some time ago in which he taps the front of various scopes made a HUGHE difference and not some mV!

[bob91343]

I would guess A5K means 5 pF and 10% tolerance.  Dunno about the A.  But that disagrees with the other side.  Still, the K tells me 10%.

I recall the days with the six or three painted dots on mica capacitors.  I have some and not only can the coding be confusing but the old paint changes color.