Please help understanding relative permiivity

[watchmaker]

I am using the Newnes books as desk companions in my study of capacitors.  I do not understand pg 22 of Circuit Design, Know it all.  (Below).

If absolute permitivity is the product of relative and absolute permitivities  and the only relative permitivity changes while all other variables remain constant, I do not understand how polystyrene can have a higher dielectric strength than pyrex glass.

What am I missing?  Thanks.


 Permitivity Page.pdf (294.04 kB - downloaded 78 times.)

[Benta]

I don't understand how you are connecting permittivity to dielectric strength.

[Sensorcat]

The attached page has two errors and is misleading in at least one point. I don't know this book, but three issues on one page makes me skeptical about it.

• First equation must be ε = ε0 ⋅ εr instead of ε = ε0 ⋅ ε.
• Dielectric strength of vacuum depends heavily on residual pressure (which is in practice never 0), and other factors; '∞' does not make much sense here.
• The top part of the page (including the equations) discusses one thing, the parallel plate capacitor and permittivity ε. Then, without any hint to the reader, a new topic is started, dielectric strength E.

Think of the two quantities like elastic modulus and ultimate tensile strength: One tells you a property of a material when used non-destructive, in linear zone, the other under which stress it will break. You cannot calculate one from the other. If pedantic, one could argue that this mechanical example does not really match the case, but you get the idea.

[watchmaker]

Thank you!  There is no way I would know the errors.  And I agree, if a knowledgeable person catches one (let alone three), then almost everything else must be viewed with caution (possibly ignored?).

It may well have misled me in my misunderstanding.

I once reviewed a book on chronometers that had erroneous captions regarding some instruments I knew first hand.  I concluded none of the captions could be trusted.

The sad part is this chapter was edited by Robert Pease, which is why I picked up the book.  Nuts.

Regards,

Dewey

[RoGeorge]

My personal impression about Robert Pease is that he was very good at practicing electronic, rather than being good at explaining it.  His writing style was very verbose, and also too colorful for me to understand what exactly is the core idea.  Unless I was already familiar with the topic, his articles were never clear to me.  Usually his articles ware almost clickbait titles, and the style reminds me of a rant.  Unless already familiar, a reader might misunderstand what he was saying.

Another trap when reading Bob Pease articles, beware of his rants about SPICE and simulators in general.  His despise for simulation was probably a good attitude for the 70s-80s, but not for today.  Nowadays simulation is an invaluable good tool for learning and/or designing.  Same for the catch phrase "My favorite programming language is soldering".  Nowadays, by each year electronics is more and more about programming, rather than soldering.

Don't get me wrong, he was very good and knew his things.  He was also famous/popular, which is rare for an EE.  Though, not particularly good at teaching, in my opinion.  Saying this only as a personal opinion after reading some of his articles (I didn't know him).  Also, I didn't read the book you have.

[watchmaker]

Thank you.  We had a writer like that in watchmaking.  Some practices he wrote about (very few) were excellent.For others he had never even tried them and they could not, would not work (thought experiments?)  Bragged he was paid by the word.

The problem was, he was the one left behind when NASA poached all the real watchmakers in the late 1950s to the 1970s.  So he got traction.  The trade is still trying to unwind the damage he left behind.

The good news is I caught the discrepancy and was able to ask here.

THANK YOU BOTH!

Regards,

Dewey

It is not only on the Internet that you have to be circumspect about what you read.

[tggzzz]

Another trap when reading Bob Pease articles, beware of his rants about SPICE and simulators in general.  His despise for simulation was probably a good attitude for the 70s-80s, but not for today. 

Not true, even thought there is some validity in your comment.

The problem with simulators is that they number-crunch models of reality. As was famously pointed out in a different context, "all models are wrong, but some are useful".

It should be noted:

• theoretical semiconductor models haven't changed significantly since the 70s
• the theoretical models are known to be deficient in some important respects. MOSFETs are particularly bad, e.g. see TAoE section 3x5.5, 3x5.6, 3x5.7. Fig 3x.45 shows one improved model which contains the standard Spice MOSFET plus 14 other components. How many manufacturers are going to bother with tuning that lot?!
• the quality of individual device models depends on the person/company creating the model; some are very distinctly "sub-optimal". That's especially true for opamps. I've declined to be paid for making models where the client just wanted anything that could be used to "tick a box"; they didn't care about the accuracy
• if all that is too esoteric, please show me the spice model for a 10nF capacitor  - where the dielectric is air, teflon, polystyrene, polyproplyene, electrolytic, X7R, etc (if you have an accurate model of dielectric absorbtion, please publish it; you will become world famous, admittedly within a small circle )
• then there's the circuit into which model devices are inserted. Those circuit models can be grossly deficient, e.g. even a one transistor one resistor circuit
• and then there are the many different types of RF simulator. That is A Big Clue that none of them are sufficient

Summary: what isn't modelled is just as important as what is modelled.

Nowadays simulation is an invaluable good tool for learning and/or designing. 

Simulation can indeed be invaluable, but in the end it is the physical reality that matters. Some people forget - or don't understand - that.

Same for the catch phrase "My favorite programming language is soldering".  Nowadays, by each year electronics is more and more about programming, rather than soldering.

Yup.

[RoGeorge]

If absolute permitivity is the product of relative and absolute permitivities  and the only relative permitivity changes while all other variables remain constant, I do not understand how polystyrene can have a higher dielectric strength than pyrex glass.

Meanwhile browsed the first chapter.  The book is not verbose in the Bob Pease stile, so all my remarks I made before does not apply.  The book itself looks good.  It looks to me more like a condensed cheat sheet, a collection of formulas, typical values, rules to apply and such.  It doesn't explain much.

The missing r in that formula is a typo, on the page before the same formula in words is stated correctly.  That typo is easy to spot, probably not a book to throw away, but as you said, this bring doubt.  Might be other mistakes out there in those 1200+ pages, and they might not be so easy to spot.


Now, about permittivity and dielectric strength.  Those are material properties, but they are unrelated, they are not derived from the formula in the pdf page you attached in the OP.  If we re-arrange the capacitor formula from that page we can get, indeed, a unit measured in V/m (I guess that made you think that permittivity and dielectric strength are tight together by that formula - they are not).  Those V/m after rearranging the formula are about the intensity of the electric field, so the E existing inside a charged capacitor.  That formula doesn't tell if the dielectric material will be capable to withstand to that E field, or if it will break.

Dielectric Strength is also measured in V/m (just like the unit for measuring the electric field E), except the dielectric strength tells the maximum E field a material can withstand.  If, for a given material, you apply an even bigger E field than the dielectric strength specified in that table, then that dielectric material will break down.  It will lose it's properties of being an electric insulator, and an electric arch will pierce the material.

Now what's the permittivity?  It's another property of materials.  In regards to capacitors, it's an indicator of how much energy can be crammed inside a dielectric material.  Vacuum (as in imaginary empty space) is the worst in terms of "storage" capacity.  That's the epsilon zero.  Dielectric materials are many times better.  Epsilon r tells how many times better a dielectric material is (at storing energy) when compared to absolute vacuum.


You may wonder how come that an insulator material (dielectric materials inside capacitors are electric insulators) stores electric energy after all?  And that's the funny part. 

Many will say, in a charged capacitor the energy is stored by the charges that accumulates on the capacitor's plates.  This is wrong.  A common misconception but dead wrong.

In a charged capacitor,  the energy is stored in the dielectric material between the plates, not in the plates of the capacitor.  If that seems hard to believe, here's a funny experiment:  what happens if you first charge a capacitor, then you dismantle it (remove the dielectric), then short-circuit the two plates, then assemble the capacitor back?

Well, total surprise, once assembled back the capacitor is still charged, as you can see at the end of this video:

Dissectible Capacitor
TSG Physics


 

The energy remained stored inside the glass material (the dielectric), so after short-circuiting the plates, and after the glass and the plates were assembled back, the energy was still there (proven by that last spark).

Even more funny about electricity (and a very common misconception) the electric energy is not in volts, not in amps, not in the coulombs, not in the wires at all.  The electric energy is in the fields, in the electric and/or magnetic fields.

For the charged capacitor above, the energy is in the electric field, E (which energy was stored by the glass material during the time the capacitor was disassembled and short-circuited).

Another funny thing (derived from the fact that energy is in the fields), electric energy is not transported through the wires.  Electric energy is transported through the electric and magnetic fields that surround the wires, so the energy travels through the air, NOT through the wires.  The wires only guide the E and H fields, so they dictate the path of the energy.

For example, the energy from the nearby power plant to your house does not travel through the wires.  The energy travels through the air, through the electric and magnetic fields (which fields are not inside the wires, the fields are outside of the wires).  The energy is always in the E and H fields, so in the space between the wires.

This is very unintuitive and hard to accept.  Shouldn't be a surprise for physicists, but most of the EE people will reject the idea of electric energy traveling through air (which air is known as an electric insulator, insulators do not conduct electricity, therefore electricity can not travel through air - well, air is an insulator for electric charges, but not an insulator for the electric and the magnetic fields, and the energy is in the fields, not in the electric charges, such the misconception).

[watchmaker]

Thank you for taking the time RoGeorge.  I need to print out your post and read it.

But yes, I am using it as a way to look up relationships quickly.  I saw others viewed Pease as a trusted source and the series seems to get some respect.  But as a learner, I have to be careful about its use.

I do rely mostly on Boylestad, MIT and Real Analog plus the LAOE lab book.  Plus the labs from the first 3.  Ihave been buiser than Iintended with watches and I am just getting back into it.

I want to replicatesomeof Dave's demonstrations to get used to my scope and AWG.

It is a little interesting without a live teacher to validate my "learning".

Again, thanks to all of you.

Regards,

Dewey

BTW, at this point in my life "mental masturbation" is about as good as it gets.  Need Viagra to avoid pissing on my toes!

[MrAl]

I am using the Newnes books as desk companions in my study of capacitors.  I do not understand pg 22 of Circuit Design, Know it all.  (Below).

If absolute permitivity is the product of relative and absolute permitivities  and the only relative permitivity changes while all other variables remain constant, I do not understand how polystyrene can have a higher dielectric strength than pyrex glass.

What am I missing?  Thanks.


(Attachment Link)

Hi,

Where do you see that?  The attachment shows pyrex glass to be higher than polystyrene.

There are a bunch of cautions here though.  The way I understand it now is that there are two types of pyrex glass.  One is borosilicate and the other is soda-lime.  I did not check into the two different electrical specifications, but they are probably different.  This would mean you would have to specify which one you are talking about.

The other thing is that the dielectric strength of a vacuum is very high maybe that is why they refer to it as infinite.  It must involve only a static charge too because a vacuum also has other properties when it comes to a changing field.  This is most likely a highly theoretical value as there are impurities and can mess that up.

Another thing is, I do not see any reference between the permittivity and dielectric strength so I cannot see why anyone would want to relate them with a formula.  We can have one value low and the other high, or we can have that one value still low but the other also low, and that one value still low and the other value somewhat at a median value.  So we can see a single value for one spec and multiple values for the other spec.  That seems to suggest they are unrelated.
Also, if you look at capacitor specs you might see a 1uf capacitor that has a voltage rating of 25 volts, while another capacitor might be 1uf and 1000 volts.  The capacitance (F) indicates how it can store energy while the voltage rating (V) tells us when it might stop being able to do that <ha ha>.

An interesting little side issue is that ordinary hot glue seems to have a very high dielectric strength, but it can crack if flexed which creates channels for moisture which of course means channels for current flow which in turn would lower the dielectric strength a lot.

As to Spice and those issues, we always have to remember that every formula we have comes from a model that simulates reality.  Every physical formula we have is from a model that we form it's not the actual real universe.  We try to emulate reality we don't yet know what it really is, or even if there is a fixed reality.  Spice models are no different and are made to model the most important characteristics just like any other physical formula.

[radiolistener]

You need to separate the wheat from the chaff and flies from the cutlets.

Relative and Absolute value is one thing.
Permittivity and permeability is another thing.
Dielectric strength is a third thing.
They all have different meaning.

A relative value is measured with respect to a reference value, which is expressed as an absolute value.

An absolute value is measured according to a common standard or reference.

In other words absolute value is a value measured in common units, like Volts, Ampers, meters, etc.
While relative value is a coefficient which shows difference with some known reference which is expressed as an absolute value.

Permittivity is a measure of a material's or environment's property that indicates how it reacts to an electric field, specifically its ability to polarize in response to an externally applied electric field.

Permeability is a measure of a material's or environment's property that indicates how it reacts to a magnetic field, specifically its ability to become magnetized in response to an externally applied magnetic field.

Absolute permittivity is the permittivity of a material or environment measured in absolute terms, typically in units such as Farads per meter (F/m).

Relative permittivity is the ratio of the permittivity of a material or environment to the absolute permittivity of a vacuum.

The same thing with Absolute/Relative Permeability - absolute permeability is measured in units such as Henries per meter (H/m), while relative permeability is a ratio to the absolute permeability of a vacuum.

And don't confuse Permittivity with Dielectric strength.

Dielectric strength is a measure of a material's or environment's ability to withstand an electric field without undergoing electrical breakdown, typically expressed as the maximum electric field that the material can endure before losing its insulating properties. It is commonly measured in units such as volts per meter (V/m) or kilovolts per millimeter (kV/mm).

As you can see Permittivity and Dielectric strength are measured in different units and have completely different meanings.

[SteveThackery]

Brilliant summary - thanks, @radiolistener.

[tggzzz]

As to Spice and those issues, we always have to remember that every formula we have comes from a model that simulates reality.  Every physical formula we have is from a model that we form it's not the actual real universe.  We try to emulate reality we don't yet know what it really is, or even if there is a fixed reality.  Spice models are no different and are made to model the most important characteristics just like any other physical formula.

The standard models of devices embody the best understanding of how devices behave, but they are imperfect in some respects. Those standard model have been translated into Spice models, so Spice simulations have imperfections. On top of that there are many parameters in the standard (and hence Spice) models, and it is questionable how well those parameters are tweaked for individual devices.

Then there are the macromodels, especially those for opamps produced by opamp manufacturers. They are of wildly varying quality and completeness, and their suitability must be assessed before using them in your simulation.

Whether Spice models or macromodels have their "important characteristics" well enough modelled for an application is up to the engineer to decide. E.g. often Spice capacitor models are sufficient, but not if your design has to take dielectric absorbtion into account. Ditto MOSFETS operating in the sub-threshold region, e.g. for energy harvesting. And so on...

[TimFox]

When a material dielectric is added to a capacitor structure, and a voltage applied to the plates, the extra stored energy (compared with the vacuum case) is due to polarization in the dielectric, which is net separation of positive charge (towards the negative plate) and negative charge (towards the positive plate), within the dielectric.
If you go to elementary electricity textbooks, you will find two field variables:  E and D.
Unfortunately, the textbooks differ in which units they use; there used to be four or more, but now there are only two in common use:
"Rationalized mks" is what electrical engineers normally use, with Volts and Amperes and SI units (meters, etc.)
"Gaussian cgs" is one of the centimeter-gram-second systems, often used by physicists, since the equations are simpler.
The equations relating E and D are different in the two systems.
Ignoring the different constants, the "displacement" D is the sum of the potential gradient E and the polarization density P.
Since this is rather complicated, I refer you to freshman-level E&M books.
https://www.physics.umd.edu/courses/Phys263/wth/fall04/downloads/EDBH/edbh.pdf  gives a reasonable description for rationalized mks units.
In Gaussian units, in vacuum, E = D; in rationalized mks, in vacuum you multiply E by the permittivity of free space e_o to get D.

[MrAl]

Just to add a little to the Spice discussion...

Some models do not even attempt to model some characteristics.  Those would be left out of the model entirely.

[tggzzz]

Just to add a little to the Spice discussion...

Some Many models do not even attempt to model some characteristics.  Those would be left out of the model entirely.

FTFY

I wouldn't mind if they told you what the models included/omitted, but you have to find that out yourself.

[mawyatt]

My personal impression about Robert Pease is that he was very good at practicing electronic, rather than being good at explaining it.  His writing style was very verbose, and also too colorful for me to understand what exactly is the core idea.  Unless I was already familiar with the topic, his articles were never clear to me.  Usually his articles ware almost clickbait titles, and the style reminds me of a rant.  Unless already familiar, a reader might misunderstand what he was saying.

Another trap when reading Bob Pease articles, beware of his rants about SPICE and simulators in general.  His despise for simulation was probably a good attitude for the 70s-80s, but not for today.  Nowadays simulation is an invaluable good tool for learning and/or designing.  Same for the catch phrase "My favorite programming language is soldering".  Nowadays, by each year electronics is more and more about programming, rather than soldering.

Don't get me wrong, he was very good and knew his things.  He was also famous/popular, which is rare for an EE.  Though, not particularly good at teaching, in my opinion.  Saying this only as a personal opinion after reading some of his articles (I didn't know him).  Also, I didn't read the book you have.

Tend to disagree with Bob's assessment. He and Jim Williams were unique with an intuition into electronics, especially anything semiconductor and analogish in nature, that was remarkable. Bob (Jim as well) tended to discuss and work at a more intuitive level rather than mathematical, thus some didn't view his ability as well.

Yes, Bob did despise SPICE. He had seen many youngsters just run a SPICE simulation rather than try to understand the circuit at hand. We had seen the same behavior especially creating and teaching Grad school as an Adjunct, and recall telling students they should be able to draw basic circuit waveforms and show bias voltages. However very few did, and suspect that Bob was irritated by this behavior also.

What most don't realize is SPICE simulations are only as good as the underlying models utilized, which as mentioned are limited in most cases. Without quality models the results are tainted which is often seen in some simulations.

When doing modern IC design one is totally dependent on SPICE or some flavor of such, and must have a thorough understanding of how the SPICE Engine works, as well a Semiconductor Physics, otherwise an expensive and career limiting failure awaits.

Digital IC designers have it better than Mixed-Signal and Analog/RF/MW types, as much of the detailed device behavior and modeling has been done to an acceptable level since the general signal parameters of input/output is well defined. They generally don't deal with CMOS device levels unless pushing the speed limits, then things can get quite a bit more involved leaning more towards the analogish world!!

Anyway, models are the achilles heal of SPICE simulations and one must pay close attention to how these models were created, under what conditions, and what limitations are expected, otherwise one is "treading on thin simulation ice"!!

Best,
 

[tggzzz]

Yes, Bob did despise SPICE. He had seen many youngsters just run a SPICE simulation rather than try to understand the circuit at hand. We had seen the same behavior especially creating and teaching Grad school as an Adjunct, and recall telling students they should be able to draw basic circuit waveforms and show bias voltages. However very few did, and suspect that Bob was irritated by this behavior also.

That phenomenon greatly irritated my father, a chemical engineer specialising in two-phase flow. Clearly that wasn't with Spice, and was even with the models and programs they had created themselves.

Examples: what happens if you knock the bottom off a large PWR, and what happens if a small PWR is tilted at quite a steep (unspecified) angle For the large PWR, he used the equations for water going down a bathtub plughole, and verified them experimentally. Then, several years later... you know the rest.

Eventually he relented, and got an HP85 for simple number crunching. On one occasion he asked me (a schoolkid!) how he could find where a curve crossed an axis. Since he knew the rough location (naturally; see above!) the algorithm was a simple search. Didn't take him long to code it up!

What most don't realize is SPICE simulations are only as good as the underlying models utilized, which as mentioned are limited in most cases. Without quality models the results are tainted which is often seen in some simulations.

When doing modern IC design one is totally dependent on SPICE or some flavor of such, and must have a thorough understanding of how the SPICE Engine works, as well a Semiconductor Physics, otherwise an expensive and career limiting failure awaits.

Such in-house designers may well have very well characterised processes and hence models. Doesn't necessarily reveal pattern-sensitive behaviour, nor microcode and electromigration problems

Digital IC designers have it better than Mixed-Signal and Analog/RF/MW types, as much of the detailed device behavior and modeling has been done to an acceptable level since the general signal parameters of input/output is well defined. They generally don't deal with CMOS device levels unless pushing the speed limits, then things can get quite a bit more involved leaning more towards the analogish world!!

Anyway, models are the achilles heal of SPICE simulations and one must pay close attention to how these models were created, under what conditions, and what limitations are expected, otherwise one is "treading on thin simulation ice"!!

I prefer the "building castles on sand" metaphor

RF engineers cannot be under any illusion whatsoever that their simulation models and even techniques are anywhere near complete. Then there's the "rusty bolt problem" found on warships, where usable frequencies have, to some extent, be determined before and after the warship is commissioned.

[MrAl]

Just to add a little to the Spice discussion...

Some Many models do not even attempt to model some characteristics.  Those would be left out of the model entirely.

FTFY

I wouldn't mind if they told you what the models included/omitted, but you have to find that out yourself.

Hi,

Ha ha, yeah, and maybe we should change "Many" to "Most"

I have seen a couple mention a thing or two in the past, like "does not model power supply current" or something like that, but then again if we have say a four unit op amp chip we only get one op amp per model anyway, so we'd have to add up for all four inside anyway.

I guess we are stuck with what we have, unless we care to make our out models that reflect what we see in measurements.

I noticed a diversion from reality with the output voltage of a simple LM358.  Spice model says one thing, spec sheet says another thing, measurements say yet another thing.  It's when we want to be ultra precise that we run into trouble with the models and even the spec sheet sometimes.  I had to be very precise in that one application.

[mawyatt]

Anyway, models are the achilles heal of SPICE simulations and one must pay close attention to how these models were created, under what conditions, and what limitations are expected, otherwise one is "treading on thin simulation ice"!!

Best,

Here's an example of a SPICE model for a 0.1uF SMD 0402 Capacitor, and one can see how involved a model can be for a simple capacitor.

Now consider what's involved with something like a Bipolar or MOS device, take a look at MEXTRAM/HICUM Bipolar, or PSP MOS models for example.

LTspice models are usually nowhere near as detailed, often just simplified behavioral models with highly limited fidelity.

Seasoned designers know how/where to apply models, to what extent to apply details, and what limitations underly, with long term experience usually backed up by teeth marks   

Best,

Code: [Select]

*----------------------------------------------------------------------
* SPICE Model generated by Murata Manufacturing Co., Ltd.
* Copyright(C) Murata Manufacturing Co., Ltd.
* Description :0402M(01005)/X6T/0.1uF/4V
* Murata P/N :GRM022D80G104ME15
* Property : C = 0.1[uF]
* Data Generated on Nov 6, 2018
*----------------------------------------------------------------------
* Applicable Conditions:
*   Frequency Range = 100Hz-6GHz
*   Temperature = 25 degC
*   DC Bias Voltage = 0V
*   Small Signal Operation
*----------------------------------------------------------------------
.SUBCKT GRM022D80G104ME15_DC0V_25degC port1 port2
C1 port1 11 9.06e-8
L2 11 12 1.20e-10
R3 12 13 2.60e-2
C4 13 14 5.31e-6
R4 13 14 750
C5 14 15 6.61e-6
R5 14 15 94.4
C6 15 16 7.83e-6
R6 15 16 8.61
C7 16 17 1.31e-5
R7 16 17 7.50e-1
C8 17 18 6.87e-6
R8 17 18 2.96e-1
C9 18 19 3.48e-6
R9 18 19 4.50e-2
C10 19 20 4.74e-3
R10 19 20 6.84e-4
L11 20 21 5.05e-12
R11 20 21 1.65e-1
L12 21 22 2.96e-11
R12 21 22 1.23e-1
L13 22 port2 4.56e-11
R13 22 port2 2.48e-2
R100 port1 11 5.00e+8
.ENDS GRM022D80G104ME15_DC0V_25degC

[tggzzz]

To use "argument by reference to authority", here's a few points Horowitz and Hill make in TAoE x-chapters.

I haven't copied the points they make about the strange results from the IntuSoft QN5088/2N5088 BJT - which aren't present in the Fairchild 2N5088 model.


MOSFET gross modelling failure in the sub-threshold region.




Complexity of the best available MOSFET model




General comments

[MathWizard]

Many will say, in a charged capacitor the energy is stored by the charges that accumulates on the capacitor's plates.  This is wrong.  A common misconception but dead wrong.

In a charged capacitor,  the energy is stored in the dielectric material between the plates, not in the plates of the capacitor.  If that seems hard to believe, here's a funny experiment:  what happens if you first charge a capacitor, then you dismantle it (remove the dielectric), then short-circuit the two plates, then assemble the capacitor back?

Well, total surprise, once assembled back the capacitor is still charged, as you can see at the end of this video:

Dissectible Capacitor
TSG Physics


 

The energy remained stored inside the glass material (the dielectric), so after short-circuiting the plates, and after the glass and the plates were assembled back, the energy was still there (proven by that last spark).

So what happens if you charge up an air gap capacitor, and remove and short the plates ? Does the energy just go straight to increasing thermal motion/kinetic energy/temperature of the air ?

I could try some calculations, but how much energy is involved in rearranging and moving some collection of electrons on a conductor, like if a bunch of electrons are built up on 1 side end of a rod. I guess there's some solid state chemistry/physics eqn's, for how those electrons diffuse back to a more electrical balanced state, and some energy must get lost as heat from collisions, and then the average kinetic energy to move the electrons.

If you charged up the dielectric material from a capacitor, how many new sets of uncharged plates could it charge up (with each pair being removed to charge the next pair) ? I have to read all this stuff some day, but not tonight.

But how does that jive with the energy eqn's I know for capacitor's, surely it won't.

[radiolistener]

So what happens if you charge up an air gap capacitor, and remove and short the plates ?

Why air? There is better choice - you can use vacuum capacitor... But I'm not sure - how you're planning to remove vacuum and replace it with another vacuum?   

[vk6zgo]

Just to add a little to the Spice discussion...

Some Many models do not even attempt to model some characteristics.  Those would be left out of the model entirely.

FTFY

I wouldn't mind if they told you what the models included/omitted, but you have to find that out yourself.

Hi,

Ha ha, yeah, and maybe we should change "Many" to "Most"

I have seen a couple mention a thing or two in the past, like "does not model power supply current" or something like that, but then again if we have say a four unit op amp chip we only get one op amp per model anyway, so we'd have to add up for all four inside anyway.

I guess we are stuck with what we have, unless we care to make our out models that reflect what we see in measurements.

I noticed a diversion from reality with the output voltage of a simple LM358.  Spice model says one thing, spec sheet says another thing, measurements say yet another thing.  It's when we want to be ultra precise that we run into trouble with the models and even the spec sheet sometimes.  I had to be very precise in that one application.

My problem with LtSpice is when "bright eyed, bushy tailed" Noobs set out to make a tuned RF amplifier using an Op Amp.
Spice is "happy as a sandboy" to simulate it, so they excitedly build it & end up with an oscillator.

Real world LC networks are not identical nor identically tuned, so slightly different tuning of the input & output tuned circuits change the negative feedback into positive feedback, just like an "old time" Tuned Plate Tuned Grid Oscillator.
They end up asking us "Greybeards" why it didn't work like the simulation!

That of course is the ultimate case, in many cases, it only takes a reactive load at the Op Amp output for them to take off---especially LM358s!!!

[MrAl]

Many will say, in a charged capacitor the energy is stored by the charges that accumulates on the capacitor's plates.  This is wrong.  A common misconception but dead wrong.

In a charged capacitor,  the energy is stored in the dielectric material between the plates, not in the plates of the capacitor.  If that seems hard to believe, here's a funny experiment:  what happens if you first charge a capacitor, then you dismantle it (remove the dielectric), then short-circuit the two plates, then assemble the capacitor back?

Well, total surprise, once assembled back the capacitor is still charged, as you can see at the end of this video:

Dissectible Capacitor
TSG Physics


 

The energy remained stored inside the glass material (the dielectric), so after short-circuiting the plates, and after the glass and the plates were assembled back, the energy was still there (proven by that last spark).

So what happens if you charge up an air gap capacitor, and remove and short the plates ? Does the energy just go straight to increasing thermal motion/kinetic energy/temperature of the air ?

I could try some calculations, but how much energy is involved in rearranging and moving some collection of electrons on a conductor, like if a bunch of electrons are built up on 1 side end of a rod. I guess there's some solid state chemistry/physics eqn's, for how those electrons diffuse back to a more electrical balanced state, and some energy must get lost as heat from collisions, and then the average kinetic energy to move the electrons.

If you charged up the dielectric material from a capacitor, how many new sets of uncharged plates could it charge up (with each pair being removed to charge the next pair) ? I have to read all this stuff some day, but not tonight.

But how does that jive with the energy eqn's I know for capacitor's, surely it won't.

Hi,

That's an interesting question.  Here's another one.

What happens if you create an air gap capacitor and place a fan outside of it (so that the air inside changes when you turn on the fan), then charge up the capacitor, then turn on the fan.  I would think that the air gets ionized, then moves away from the plates as more air gets ionized.

Now what if we had a second capacitor, uncharged, in close proximity to the first cap that was charged, then turn on the fan.  I would think the ionized air would flow into the other capacitor and then that capacitor would become charged, even if not exactly to the same level as the first.

How about a glass dielectric capacitor.  Charge it up, then remove the glass, then place the glass inside another set of capacitor plates.  the second set of plates should be charged, I would think.  And there we have an isolation capacitor.  Now set up a long line of capacitors with only the first one with a glass dielectric.  Charge up the first one, move the glass to the second one, then to the third one, then to the fourth one, etc.  The final capacitor should be the only one charged unless you keep charging the first one over again with a new glass element inside in which case they would all eventually be charged with their own glass elements inside.
The only caveat I think is in the process of moving any of the glass elements there could be some loss due to leakage currents.  It would have to be performed under precise laboratory conditions.

How about a voltage doubler?  Charge up two caps with glass dielectric, then take the two glass elements and stack them one on top of the other, then replace one set of plates.  I would think the voltage across this new cap would be twice that of any single cap.

Now how about the dual of the capacitor, the inductor.
Use a coil around a metal core creating an inductor.  'charge' up the inductor, then remove the core, then place it inside a different coil.  If the second coil has the same number of turns the turns ratio would be 1:1 (ignoring losses for now again).  If the second coil had twice as many turns, the turns ratio would be 1:2, and if it had less turns, the turns ratio would be 2:1.
This is not that hard to imagine because we know the cores become magnetized.
The only caveat here is that we would have to switch coils very fast before the magnetization becomes lower than the original coil and core had.  We may have to use some kind of 'keeper' while we were moving the core to another coil.  We'd also have to consider how the core was inserted into the second coil because it would be moving and therefore I would expect it to try to create a current in the second coil.
Maybe a better scenario would be to have a whole line of coils placed end to end just separated a little, then 'charge' up the coil on one far end with the core inside, then slide the core down the line of coils and watch what happens.

In any of these we'd have to consider how the movement of any storage medium changes and its effects.

I suppose there could be a myriad of combinations too with caps and inductors.  Set a cap and inductor into oscillation then near the zero crossings swap cores with another set of plates and another coil.  Should continue to oscillate I would think.

Maybe a simpler scenario with the capacitors would be to have one dielectric but two plates on both sides.  Charge it up with one set of plates, then discharge it with the other two sets of plates.  I would think this would be something we could build.  It seems that the two plates would be isolated from each other, but that would have to be tested.  That would be a kind of capacitor transformer.

Bear in mind I've never performed any of these experiments.