Thread: capacitor with varyable permittivity dielectric

Ok so I have a question , I have thought about it and I have heard some answers from other poeple so I would gladly hear some from the members of this forums.

Say we have a DC source and a transformer primary in series with that source , and then we add a capacitor in series with the transformer pirmary.Normally when you connect the dc source what happens is at first there is a current flow due to the charging up of the capacitor , once the capacitor is charged the current flow stops and the system reaches equilibrium, since a transformer can only transfer energy as long as the current in the primary is either changing amplitude and or polarity then after the cap charges up the transformer cannot induce a current in the secondary anymore.

Now what would happen if one could have a dielectric material between the plates such that one could alter it's dielectric constant while the cap is charged , so with varying the constant one would also change the capacitance of the capacitor , and with either lovering or increasing the capacitance and keeping voltage the same would result in charge flow back and forth from the capacitor and such charge flow could make the transformer do useful work even with a dc supply , am I correct or is there something wrong here ?

I would expect the charge on the capacitor to fluctuate and the current through the primary to do the same, so that some voltage would be generated across the secondary. A similar effect would be caused by varying the distance between the capacitor plates.

The equilibrium is not because of the current in the primary going to zero, it's because the current isn't changing. The same thing happens when a steady current is flowing in the primary - no voltage is induced. You can induce a voltage on the secondary in a much simpler way just by opening and closing a switch on the primary side. This is the principle of the induction coil in an automobile.

thanks for the replies Harold and JonG,
well yes I know that a transformer needs a time varying in amplitude and /or polarity signal for induction to happen and any transfer of energy between the coils to be possible.
I'm also aware that transistors in most of the SMPS do this , they chop the incomming DC so that it could be used to induce secondary current.

I was just wondering about some other possible way to achieve this assuming we have a dc supply, and we need a transformer to isolate/step down or up voltage/ current.

Also from a pure point of durability , an old spark gap may wear down much faster and be much les safe yet it deals just fine with overcurrents or current lagging or any other factor that comes in mind could blow up a good smps.
Also the fact that whenever a switch (semiconductor) is closed in series with the primary it has the full PD across that switch , I was thinking maybe there is a way in which to make the electric field due work instead of physical switches directing current paths.

I;m not saying this could work just for an example , imagine this varyable cap I was talking about has a special dielectric material which drops in dielectric constant whenever a certain level of voltage is reached across the capacitor plates, like the cap charges charges and then the constant drops and so drops the capacitance and some of the charge that went into the capacitor now has to flow back , but then after some moment the dielectric goes back to its initial constant and so the capacitance increases and the charge runs towards the capacitor once again.
In this sense this device would work similar to a spark gap , as whenever the voltage on the contacts reaches a certain value a arc forms and current travels for a brief moment of time until the voltage drops and the arc cannot be sustained anymore.

What do you think about this?

Originally Posted by Crazymechanic

thanks for the replies Harold and JonG,
well yes I know that a transformer needs a time varying in amplitude and /or polarity signal for induction to happen and any transfer of energy between the coils to be possible.
I'm also aware that transistors in most of the SMPS do this , they chop the incomming DC so that it could be used to induce secondary current.

I was just wondering about some other possible way to achieve this assuming we have a dc supply, and we need a transformer to isolate/step down or up voltage/ current.

Also from a pure point of durability , an old spark gap may wear down much faster and be much les safe yet it deals just fine with overcurrents or current lagging or any other factor that comes in mind could blow up a good smps.
Also the fact that whenever a switch (semiconductor) is closed in series with the primary it has the full PD across that switch , I was thinking maybe there is a way in which to make the electric field due work instead of physical switches directing current paths.

I;m not saying this could work just for an example , imagine this varyable cap I was talking about has a special dielectric material which drops in dielectric constant whenever a certain level of voltage is reached across the capacitor plates, like the cap charges charges and then the constant drops and so drops the capacitance and some of the charge that went into the capacitor now has to flow back , but then after some moment the dielectric goes back to its initial constant and so the capacitance increases and the charge runs towards the capacitor once again.
In this sense this device would work similar to a spark gap , as whenever the voltage on the contacts reaches a certain value a arc forms and current travels for a brief moment of time until the voltage drops and the arc cannot be sustained anymore.

What do you think about this?

There is no problem with transferring energy via a variable capacitance. You don't even need to invoke a special dielectric. Let's just use air: Imagine a parallel-plate capacitor whose spacing can be varied at will to vary the capacitance. If you separate the plates when there is voltage across it, you perform work against the electrostatic field, and that energy is reflected as an increase in stored capacitor energy. You can return to the initial plate spacing when there is no voltage across it, and do it all over again, increasing the net energy each time. This principle was exploited in early (pre-maser) amplifiers for radioastronomy (the capacitor was implemented by a semiconductor diode, whose capacitance is a function of applied voltage). Amplifiers of this type are often called "parametric" amplifiers, since some parameter -- in this case, capacitance -- is varied to achieve amplification.

hey thanks for the answers, well the idea was to have a capacitor like the one I described but a solid state device rather than a mechanical device with moving parts and so on.
I wonder is there such material which coul exhibit such properties to be used for a varyable permittivity dielectric, maybe there could be some technology implemented inside the dielectric to cancel the dipole alignment to the field of the plates which would cause the field between the plates to drop off or lower hence also change in current would happen if the cap would be attached to a circuit.

Originally Posted by Crazymechanic

hey thanks for the answers, well the idea was to have a capacitor like the one I described but a solid state device rather than a mechanical device with moving parts and so on.
I wonder is there such material which coul exhibit such properties to be used for a varyable permittivity dielectric, maybe there could be some technology implemented inside the dielectric to cancel the dipole alignment to the field of the plates which would cause the field between the plates to drop off or lower hence also change in current would happen if the cap would be attached to a circuit.

Did you miss this part of tk421's reply?

the capacitor was implemented by a semiconductor diode, whose capacitance is a function of applied voltage

Ok I got it , a device commonly known as a varicap probably.
the problem here would be that the change in capacitance is very small to be used to achieve any serious current flows, and the other would be a question , does the capacitance which is the function of the applied voltage could be made to vary itself like after a certain value of voltage is reached it drops off and then recovers or is it only possible to vary it by a " driver" circuit or so?

Originally Posted by Crazymechanic

Ok I got it , a device commonly known as a varicap probably.
the problem here would be that the change in capacitance is very small to be used to achieve any serious current flows, and the other would be a question , does the capacitance which is the function of the applied voltage could be made to vary itself like after a certain value of voltage is reached it drops off and then recovers or is it only possible to vary it by a " driver" circuit or so?

The change in capacitance can be as large as you want -- just parallel more caps! Commercial varicaps/varactors come in many sizes. And you can always parallel as many as you need.

Varactors can also be designed to have quite-large capacitance variations. Such "hyperabrupt" varactors can provide in excess of 10:1 capacitance variations.

ok I see, well ther is not that mcuh info on the net about such devices but basically I understand that there are semiconductrs like the ones I described , so would they change the capacitance without a driving circuit or no?
Like take for example an ordinary capacitor with two plates and a dielectric between those plates , is there any such material for the dielectric which could chnage the constant based on the applied voltage (electric field strength) ad then chnage it back again like having memory ? as such a material would not require a driving circuit for the device, and frequency stability could be achieved with keeping the voltage constant.

Originally Posted by Crazymechanic

ok I see, well ther is not that mcuh info on the net about such devices but basically I understand that there are semiconductrs like the ones I described , so would they change the capacitance without a driving circuit or no?
Like take for example an ordinary capacitor with two plates and a dielectric between those plates , is there any such material for the dielectric which could chnage the constant based on the applied voltage (electric field strength) ad then chnage it back again like having memory ? as such a material would not require a driving circuit for the device, and frequency stability could be achieved with keeping the voltage constant.

I'm not sure what you're asking for. The varactors I've described have a voltage-dependent capacitance. That's how you control the capacitance. I thought I'd made that clear. So I don't understand why you're asking about changing dielectric constants with applied voltage. Varactors do the equivalent (by varying the effective plate spacing).

Your comment about frequency stability is wrong, by the way. You've assumed that keeping voltage constant means a constant capacitance. However, you've overlooked temperature as a variable. If you need frequency stability, that needs to be considered, too.

I'm also mystified by your comment about not needing a driving circuit. Even in the description of your putative variable-dielectric-constant device, you talk about varying the constant via an applied voltage.

Perhaps you could state more clearly what you're trying to achieve, including what constraints you want to impose on the solution?

Ok let me put it this way, you do know how a smps works , it takes a steady dc supply and basically chops it so that there would be many impulses as that is the only way to get a trasnformer working - a amplitude and or polarity varying current.

Now Say I wouldn't like to use mosfets or other transistors for my power transfer project but make something different that I ahve been thinking about , imagine the same steady DC supply , either rectified and filtered or a battery doesnt matter that much, take that steady dc and put a transformer primary in series and also a capacitor in series with that primary, there is current induced in the secondary only as long as it takes the cap to charge up , after that nothing , so if I could change the dielectric constant of the dielectric material it would change the permittivity and so say decrease it so the charge per area on the plates should also decrese which would imply some current going away from the cap as charge flowing away from it , and then reset the dielectric constant to what it was before so now the charge is going back again to the cap , and so do this in cycles, and achieve induction in the transformer.

when I asked about the driving circuit I was asking whether the method which could alter the dielectric would need some additional mechanism which would be electrically controlled or maybe there would be such a material which as I said could change its permittivity based upon the field strength and once dropped could reset itself , like a spring having memory so to speak of.
I referred to a spark gap as that resembles the mechanism im speaking about in such a wa that a spark occurs everytime the voltage on the contacts has got back to the level needed to make the spark.

I guess by now my idea should be clear ?