Which small ceramic capacitors have the worst microphonics?

[Joerg]

Does anyone know which tiny MLCC have the worst microphonics? Brand, series?

I want to do some tests to see how well these things can sense low pitch
sound, mainly via changes in capacitance but also piezo effects. Main
reason is that I need a fairly low impedance, high capacitance and a
very small size. Ideally less than 0.010" width and height but for right
now I can test with larger ones. 0.050" or so would be fine for testing
the concept, maybe even a little larger. A capacitance higher than
0.01uF would be nice. The frequency response doesn't have to go beyond
100Hz.

 

[John S]

Does anyone know which tiny MLCC have the worst microphonics? Brand, series?

I want to do some tests to see how well these things can sense low pitch
sound, mainly via changes in capacitance but also piezo effects. Main
reason is that I need a fairly low impedance, high capacitance and a
very small size. Ideally less than 0.010" width and height but for right
now I can test with larger ones. 0.050" or so would be fine for testing
the concept, maybe even a little larger. A capacitance higher than
0.01uF would be nice. The frequency response doesn't have to go beyond
100Hz.

My suggestion, Joerg, is to get the greatest capacitance in the smallest
size to start with.

 

[John S]

My suggestion, Joerg, is to get the greatest capacitance in the smallest
size to start with.

I should have given some reasoning behind this. Sorry.

The high capacitance in small size can only be achieved, AFIK, with high
dielectric constants. That means a lot of variation with applied voltage
and with mechanical variation as well. They are transducers that were
not designed as such.

Cheers

 

[whit3rd]

Does anyone know which tiny MLCC have the worst microphonics? Brand, series?

Well, the worst microphonics will be for quartz crystals and ceramic resonator
equivalents (because they are poled). Microphonic response at frequency F
depends on the crystal being asymmetric (so that it knows up from down).
A non-polarized MLCC would see sound input at F and give feeble electrical
output at 2F. Raise the bias voltage on the MLCC , maybe beyond the recommended
applied voltage, to see electrical output at F.

Quartz crystals and ceramic resonators are usually packaged so you cannot
stress them with external soundwaves, but the internal elements are sensitive.
Ceramic resonators are permanently 'poled' with no need of a bias voltage source,
and alpha-quartz has asymmetric crystal structure which amorphous polycrystalline
quartz lacks. Sadly, you can't make a halogen lamp envelope into a good
microphone.

 

[Joerg]

I should have given some reasoning behind this. Sorry.

The high capacitance in small size can only be achieved, AFIK, with high
dielectric constants. That means a lot of variation with applied voltage
and with mechanical variation as well. They are transducers that were
not designed as such.

Yes, that's what I probably do. But it is often surprising, one brand
has 2x the microphonics and all else is pretty much the same. So I was
wondering if people knew something, such as "brand XYZ always has tons
of microphonics" or something like that.

 

[Joerg]

Well, the worst microphonics will be for quartz crystals and ceramic resonator
equivalents (because they are poled). Microphonic response at frequency F
depends on the crystal being asymmetric (so that it knows up from down).
A non-polarized MLCC would see sound input at F and give feeble electrical
output at 2F. Raise the bias voltage on the MLCC , maybe beyond the recommended
applied voltage, to see electrical output at F.

Quartz crystals and ceramic resonators are usually packaged so you cannot
stress them with external soundwaves, but the internal elements are sensitive.
Ceramic resonators are permanently 'poled' with no need of a bias voltage source,
and alpha-quartz has asymmetric crystal structure which amorphous polycrystalline
quartz lacks. Sadly, you can't make a halogen lamp envelope into a good
microphone.

I can't have any bias voltage in this case. But to my surprise I just
read a report where a C0G cap is as "bad" in microphonics as the
ferroelectric X7R sibling. This turf seems to be full of surprises.

 

[Joerg]

Wouldn't that most likely be X7R?

One would think so because it's ferroelectric but I found a Kemet study
showing it may not be so in all cases.

 

[Klaus Kragelund]

One would think so because it's ferroelectric but I found a Kemet study

showing it may not be so in all cases.

Perhaps the brands from the unknown (read Chinese) manufactors may be worse since they might not have inhouse control to minimize that parameter.

Is this for a one-off or a production item. I wouldn't sleep nights if it were for a production item.

Regards

Klaus

 

[Joerg]

Perhaps the brands from the unknown (read Chinese) manufactors may be
worse since they might not have inhouse control to minimize that
parameter.

Is this for a one-off or a production item. I wouldn't sleep nights
if it were for a production item.

Long term it is for production but right now only a feasibility study.
If it works we can then determine a pathway to obtain caps of consistent
behavior. If necessary a custom run. Some deviation is fine because
every unit will be calibrated.

I always sleep well ... Drives others crazy, I can say good night,
turn around, and 15 seconds later ... zzzzzzz

 

[[email protected]]

Long term it is for production but right now only a feasibility study.

If it works we can then determine a pathway to obtain caps of consistent

behavior. If necessary a custom run. Some deviation is fine because

every unit will be calibrated.



I always sleep well ... Drives others crazy, I can say good night,

turn around, and 15 seconds later ... zzzzzzz

I'm the same way, even during periods of high stress on the job.

But, I have been told that could be a signal that one has sleep apnea, and it could be dangerous.

Some specialist say that if you can fall asleep fast, that is because you get to little sleep and you should get more sleep. I get along fine with 6 hours per night, but I may just be kidding myself....

Regards

Klaus

 

[Joerg]

I'm the same way, even during periods of high stress on the job.

I thought all Vikings could do that

But, I have been told that could be a signal that one has sleep
apnea, and it could be dangerous.

Yep, especially if you have cases in the immediate family. The only way
to find out would probably be in one of those sleep diagnosis centers
that some hospitals run. Essentially you have to stay overnight.

Some specialist say that if you can fall asleep fast, that is because
you get to little sleep and you should get more sleep. I get along
fine with 6 hours per night, but I may just be kidding myself....

6h is quite low. I try to get in 7h which isn't so easy anymore. Since
California had a business exodus most of my clients are now 2-3
timezones ahead. This means I now have to get up latest by 6:30am. And
then the dogs want to be walked as well.

 

[mook johnson]

Yes, that's what I probably do. But it is often surprising, one brand
has 2x the microphonics and all else is pretty much the same. So I was
wondering if people knew something, such as "brand XYZ always has tons
of microphonics" or something like that.

hasn't been my expereince. I once has some large like .5X x 1" stacked
X7Rs from AVX in the lab as motor bus capacitors. When I ran the motor
I could hear the RPM of the motor from the cap back across the lab.

Replaced them with Kemet, Novacap and another custom cap supplier, and
they were very similar.

Can you gain some sensitivity by tuning the mounting structure (pcb I
assume) to have a high mechanical Q at the frequency if interest?

 

[Joerg]

hasn't been my expereince. I once has some large like .5X x 1" stacked
X7Rs from AVX in the lab as motor bus capacitors. When I ran the motor
I could hear the RPM of the motor from the cap back across the lab.

Replaced them with Kemet, Novacap and another custom cap supplier, and
they were very similar.

I've had differences and those most likely had to do with the number of
layers and things like that. Some caps were taller than others.

Can you gain some sensitivity by tuning the mounting structure (pcb I
assume) to have a high mechanical Q at the frequency if interest?

Unfortunately not in this case because I have to measure from DC for
30-40Hz.

 

[miso]

Does anyone know which tiny MLCC have the worst microphonics? Brand, series?

I want to do some tests to see how well these things can sense low pitch
sound, mainly via changes in capacitance but also piezo effects. Main
reason is that I need a fairly low impedance, high capacitance and a
very small size. Ideally less than 0.010" width and height but for right
now I can test with larger ones. 0.050" or so would be fine for testing
the concept, maybe even a little larger. A capacitance higher than
0.01uF would be nice. The frequency response doesn't have to go beyond
100Hz.

Much like chip design, you will probably find different parts of the PCB
have more flex, which in turn effect the cap. Usually the corners are
considered to be the quietest.

If you can't put the critical cap in a corner and if the PCB is
rectangular. there is probably more flex in the long direction of the
board, so place the caps perpendicular to the long edge.

There may be a way to parallel caps so the microphonics cancel.
Obviously that would take more caps of a smaller value with different
orientations.

If you dig into chip design books and papers, you run into this kind of
black magic.

Perhaps flexible (elastomer) mounting of the PCB?

 

[Joerg]

Much like chip design, you will probably find different parts of the PCB
have more flex, which in turn effect the cap. Usually the corners are
considered to be the quietest.

If you can't put the critical cap in a corner and if the PCB is
rectangular. there is probably more flex in the long direction of the
board, so place the caps perpendicular to the long edge.

There may be a way to parallel caps so the microphonics cancel.
Obviously that would take more caps of a smaller value with different
orientations.

If you dig into chip design books and papers, you run into this kind of
black magic.

Perhaps flexible (elastomer) mounting of the PCB?

I actually want microphonics in this case and the cap won't be on a PCB.
Some tests yesterday with what I have here showed that while slight
pushing onto the surface of the cap does generate a detectable signal I
could hardly measure any capacitance change. Capacitance change would be
the only way to get DC out of it.

 

[Klaus Bahner]

On 14-04-2013 21:54, Joerg wrote:
[snip]

I can't have any bias voltage in this case. But to my surprise I just
read a report where a C0G cap is as "bad" in microphonics as the
ferroelectric X7R sibling. This turf seems to be full of surprises.

All of these ceramic materials *are* ferroelectric, otherwise you
wouldn't get the high epsilon_r. Slight changes in the composition of
the compound affects temp coefficients and other properties, hence the
classification into X7R, C0G etc., but they are all ferroelectrics.

I think it is impossible to deduce microphonics from these general
material properties at all. For example I would assume that manufactures
use tricks like varying the orientation of the layers in a multilayer
cap to minimize the overall microphonics effect.

So most likely you have to contact the manufacturers and try to get into
contact with one of the component design engineers knowing the gory
details - which most likely is easier said than done :-(

Another loose ideas could be the IEEE Trans. Ultrasonics, Ferroelectrics
and Frequency Control. I seem to remember an article on capacitor
microphonics there, but a quick search didn't find anything. Most likely
my memory doesn't serve me well here.

Or look for PVDF (Polyvinylidene flueride) sensors, usually used as
piezo, but the should make a good microphone, too. Perhaps there is
something on the market which can be mounted on a PCB. The price tag
might be much higher than a capacitor, though.

Regards,
Klaus

 

[Joerg]

On 14-04-2013 21:54, Joerg wrote:
[snip]

I can't have any bias voltage in this case. But to my surprise I just
read a report where a C0G cap is as "bad" in microphonics as the
ferroelectric X7R sibling. This turf seems to be full of surprises.

All of these ceramic materials *are* ferroelectric, otherwise you
wouldn't get the high epsilon_r. Slight changes in the composition of
the compound affects temp coefficients and other properties, hence the
classification into X7R, C0G etc., but they are all ferroelectrics.

I think it is impossible to deduce microphonics from these general
material properties at all. For example I would assume that manufactures
use tricks like varying the orientation of the layers in a multilayer
cap to minimize the overall microphonics effect.

So most likely you have to contact the manufacturers and try to get into
contact with one of the component design engineers knowing the gory
details - which most likely is easier said than done :-(

My experience with that is that they won't release much, this is usually
closely guarded internal reseach.

Another loose ideas could be the IEEE Trans. Ultrasonics, Ferroelectrics
and Frequency Control. I seem to remember an article on capacitor
microphonics there, but a quick search didn't find anything. Most likely
my memory doesn't serve me well here.

Good idea. I am a member of that society but since the Transactions no
longer come on paper it's easier to miss stuff when you don't make
regular checking a habit. A paper kind of pushes itself under the nose
for the evenings but the computers are off in our house after 6:30pm or so.

Or look for PVDF (Polyvinylidene flueride) sensors, usually used as
piezo, but the should make a good microphone, too. Perhaps there is
something on the market which can be mounted on a PCB. The price tag
might be much higher than a capacitor, though.

In this case it can be several Dollars per capacitor. But PVDF would not
offer very large capacitances when small (a few hundred um by 1mm or so
in real estate). I remember some PVDF experiments we did in medical
ultrasound. We never warmed up to that material. It had great bandwidth
which is important in that market but shows quite poor acoustic
efficiency. So we kept PZT.

 

[mike]

Does anyone know which tiny MLCC have the worst microphonics? Brand, series?

I want to do some tests to see how well these things can sense low pitch
sound, mainly via changes in capacitance but also piezo effects. Main
reason is that I need a fairly low impedance, high capacitance and a
very small size. Ideally less than 0.010" width and height but for right
now I can test with larger ones. 0.050" or so would be fine for testing
the concept, maybe even a little larger. A capacitance higher than
0.01uF would be nice. The frequency response doesn't have to go beyond
100Hz.

You don't disclose the application, and the devil is always in the details.

I have a couple of concerns.

A device that small and stiff is gonna have about zero sensitivity
below 100Hz. To make it work, you'll have to "bend it" by attaching
it to something bigger and flexible. Operational bending is bad enough.
Shock/vibration/resonance is gonna be a reliability nightmare.

I'm skeptical about component usage far outside the design parameters.
I've had enough production problems when the vendor actually improved
a published spec.
Exploiting an unspecified characteristic that's undesirable for
the typical application is asking for trouble.

 

[Joerg]

You don't disclose the application, and the devil is always in the details.

Sensing a local pressure in liquid. It has a DC value and then I'll have
to measure up to 20-30Hz worth of changes. Very cramped space, hence the
0.010" width.

I have a couple of concerns.

A device that small and stiff is gonna have about zero sensitivity
below 100Hz. To make it work, you'll have to "bend it" by attaching
it to something bigger and flexible. Operational bending is bad enough.
Shock/vibration/resonance is gonna be a reliability nightmare.

I can't work on bending (which I am aware is the normal modus operandi
of a "singing cap"). I can only work with thickness changes which will
eb quite miniscule. Ideally with capacitance change because the signal
coming out of it via piezo generation is very tiny, at least here on the
bench.

I'm skeptical about component usage far outside the design parameters.
I've had enough production problems when the vendor actually improved
a published spec.
Exploiting an unspecified characteristic that's undesirable for
the typical application is asking for trouble.

Oh, I've done that a lot

One could buy half a truckload of reels and store them in a gigantic
nitrogen cabinet. That would last nearly forever. But the better method
is to first establish principle of operation. Then it's time to sit down
with a manufacturer. It would not be the first time where the response
is "You want to do WHAT?!". After they heard the Dollar numbers involved
that usually changed to "Oh, wow, let's see how we can do this".

 

[Joerg]

You can easily test samples yourself.

A couple of years ago I submitted a TL431 based electret mic booster to
Elektor magazine, which was published.

Among my experiments; I stripped a spare electret mic to get at the
special JFET inside, I examined the microphonic properties of various
components including bog standard capacitors of various types.

In case you have difficulty locating a copy of the article:

Give the TL431 a Vcc of 9 - 12V and a cathode load of at least 200 Ohms,
connect the electret from cont ip to GND and set the K to 1/2 Vcc with a
47k from K to cont ip.

Measure the resistance - the final resistor gives large nfb, both AC &
DC, you need to shunt the AC nfb to get very high gain with stable DC
set point. connect a 100uF cap to GND and a 1k2 from K to the + of the
cap, subtract that 1k2 from the measured resistance and put a fixed
resistor equal to the remainder from cap + to cont ip.

I don't know what "cont ip" is. But I have a working measurement setup
here, that's not the problem.

If yoy break open the electret mic capsule, you can remove the electret
dielectric and attach wires to connect your capacitor.

Pretty much all types of non-electrolytic capacitor are microphonic when
tapped with a screwdriver - you will soon see, some more than others.

I know, that's the reason for my posting. Which ones are the worst?
Because worst is best in my case.

If the finished circuit picks up radio stations, your electrolytic has
low ESR - put a 100 Ohm resistor in series with it.

EMI is not a problem, we'll run this in a modulated fashion.