Curious about how you can find the frequency dependance of ceramic capacitors. I was looking at ceramic capacitor datasheets and found these charts of capacitance frequency dependance,


https://product.tdk.com/en/search/capacitor/ceramic/mlcc/info?amp=&part_no=C1608JB1C106M080AB


https://product.tdk.com/en/search/capacitor/ceramic/mlcc/info?part_no=C1608X5R1C106M080AB

Both of these are from TDK manufacturer.

Clearly, the depedance deviates strongly a bit over 1MHz which could make a cap totally unsuitable for an application.

I hadn't seen these charts before and couldn't find any equivalent charts or specs in datasheets from other manufacturers. This data seems important, Should I be able to calculate this deviation from other datasheet specs?
Can I predict is capacitance will rise or fall (as seen in charts above)?
Is this just common knowledge about ceramic caps?

I read that X7R doesn't deivate until higher frequencies. However, the X7R caps at TDK show just the same behaviour as X5R (I didn't look at all X7R parts).

Only noticed this as I was looking for a suitable replacement for a ceramic capacitor with obscure dialectric type, JB. Decided X5R is fine and looked at some specs. I'm only using the cap at a lower frequency so this deviation doesn't matter for me, but would like to know more about this.

 

It would be very unusual to use a 10uF capacitor at MHz frequencies. I suspect what you are seeing in those charts is the self-resonance frequency.

 

Curious about how you can find the frequency dependance of ceramic capacitors. I was looking at ceramic capacitor datasheets and found these charts of capacitance frequency dependance,

View attachment 347618
https://product.tdk.com/en/search/capacitor/ceramic/mlcc/info?amp=&part_no=C1608JB1C106M080AB

View attachment 347619
https://product.tdk.com/en/search/capacitor/ceramic/mlcc/info?part_no=C1608X5R1C106M080AB

Both of these are from TDK manufacturer.

Clearly, the depedance deviates strongly a bit over 1MHz which could make a cap totally unsuitable for an application.

I hadn't seen these charts before and couldn't find any equivalent charts or specs in datasheets from other manufacturers. This data seems important, Should I be able to calculate this deviation from other datasheet specs?
Can I predict is capacitance will rise or fall (as seen in charts above)?
Is this just common knowledge about ceramic caps?

I read that X7R doesn't deivate until higher frequencies. However, the X7R caps at TDK show just the same behaviour as X5R (I didn't look at all X7R parts).

Only noticed this as I was looking for a suitable replacement for a ceramic capacitor with obscure dialectric type, JB. Decided X5R is fine and looked at some specs. I'm only using the cap at a lower frequency so this deviation doesn't matter for me, but would like to know more about this.

That's perfectly normal for those types of devices. That's why you use more than one ceramic capacitor for bypass and decoupling applications. Select the needed capacitance (a 10 uf for energy storage and a 0.1 uf for high frequency chip bypass/decoupling) for the application (these capacitors serve different uses, so they might be physically separated on the board design), check the chart, if you need higher frequency specs, at a particular spot, select another to parallel the first so their charts overlap. This is very old stuff we learned in the tube era. You kids don't learn this in Maker school today? Even if the capacitors are perfect, there is still trace inductance and other factors that will require correct placement of each capacitor on the board.

https://www.ti.com/lit/an/scba007a/scba007a.pdf

 

Curious about how you can find the frequency dependance of ceramic capacitors.

The datasheet should say but the often do not.
At work I use a Vector Network Analyzer that cost more than a car. At home I use a $100 version. (before the dam Tariffs)

Next you will need a $10 board like this. I connect directly to the VNA and use short coax.

You should probably get this board also. $12 It has many examples of R, C, L circuits. This is better than a manual. You can see what a RC low pass filer looks like.

I one of the modes you will get graphs like this. Here is a 0.1uf cap showing its capacitor effect to the left and its inductance to the right. The resonant frequency is at 5mhz. It has 0.66 ohms if ESR.

 

Certainly the effective value of a capacitor will depend on the frequency! This means that to select an effective capacitor for an application, there must be an understanding of both the purpose and the frequency range involved.
Fortunately the manufacturers of capacitors , and also some distributors, are able to provide useful data.
This is why it is important for a circuit designer to provide the capacitor selection to the PCB layout drafter person.

 

Here is a 0.1uf cap showing its capacitor effect to the left and its inductance to the right. The resonant frequency is at 5mhz. It has 0.66 ohms if ESR.

To increase the usable frequency range as a bypass capacitor, generate a small, negative series inductance that negates the ESL of the 100nF capacitor.

 

An alternative is to select a different value of capacitor.

 

To increase the usable frequency range as a bypass capacitor, generate a small, negative series inductance that negates the ESL of the 100nF capacitor.

With what, a gimmick inductor? We would use a gimmick cap from short lengths of wire to neutralize tube PA amplifiers but it was frequency dependent in the VHF or UHF range. Circuit fringe reactance, mainly in the tubes, would make a phase shift network causing oscillations at high power.

 

"Neutralizing" an RF amplifier with capacitors was actually adding capacitive negative feedback to cancel the effect of capacitive (usually) positive feedback. While the theory of neutralizing is quite sound and solid, the implementation was sometimes moore mystical. It is difficult for most of us to see high frequency electric potential fields.

 

"Neutralizing" an RF amplifier with capacitors was actually adding capacitive negative feedback to cancel the effect of capacitive (usually) positive feedback. While the theory of neutralizing is quite sound and solid, the implementation was sometimes moore mystical. It is difficult for most of us to see high frequency electric potential fields.

Yes, a little wire from the grid near the plate for negative feedback, if balance to the positive feedback. When we did a PA tube (on a 10KW exciter and sometimes on a much larger power final PA), the second step after the power-on smoke test was to adjust negative feedback via neutralizing capacitors, stage by stage.



A final PA tube.

Trying to neutralize (I know it's not exactly the same thing) a bypass capacitor by adding inductance gave me a good laugh.

 

Trying to neutralize (I know it's not exactly the same thing) a bypass capacitor by adding inductance gave me a good laugh.

That's what happens when you don't pay attention to what you read.

With what, a gimmick inductor?

Your "gimmick inductor" gave me a good laugh too.

 

At least some tube type power amps started the process by adjusting for minimum fed thru power prior to switching on the high voltage.

 

That's what happens when you don't pay attention to what you read.



Your "gimmick inductor" gave me a good laugh too.

The entire set of posts was in jest my friend. What's possible and what's practical are often two separate things. It's much easier just to pick the correct component in the first place.

 

That's perfectly normal for those types of devices. That's why you use more than one ceramic capacitor for bypass and decoupling applications. Select the needed capacitance (a 10 uf for energy storage and a 0.1 uf for high frequency chip bypass/decoupling) for the application (these capacitors serve different uses, so they might be physically separated on the board design), check the chart, if you need higher frequency specs, at a particular spot, select another to parallel the first so their charts overlap. This is very old stuff we learned in the tube era. You kids don't learn this in Maker school today? Even if the capacitors are perfect, there is still trace inductance and other factors that will require correct placement of each capacitor on the board.

https://www.ti.com/lit/an/scba007a/scba007a.pdf

That's perfectly normal for those types of devices. That's why you use more than one ceramic capacitor for bypass and decoupling applications. Select the needed capacitance (a 10 uf for energy storage and a 0.1 uf for high frequency chip bypass/decoupling) for the application (these capacitors serve different uses, so they might be physically separated on the board design), check the chart, if you need higher frequency specs, at a particular spot, select another to parallel the first so their charts overlap. This is very old stuff we learned in the tube era. You kids don't learn this in Maker school today? Even if the capacitors are perfect, there is still trace inductance and other factors that will require correct placement of each capacitor on the board.

https://www.ti.com/lit/an/scba007a/scba007a.pdf

Reviving a pretty old thread, but since I've not seen it mentioned here, Murata's online SimSurfing tool is an invaluable tool for this type of information.