Continue to Site

Welcome to

Welcome to our site! is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Rotary capacitor physics and design

Not open for further replies.


Full Member level 3
Jul 16, 2014
Reaction score
Trophy points
Activity points
Hi so my goal is to make a rotary capacitor, the purpose is rather complicated so right now i'm just looking whether I can make the capacitor at all with the intended capacitance.

given I use thin (as thin as mechanically possible for them to still be rigid) copper discs where one is stationary and the next one on a shaft rotating , all of the discs side by side , the question is what would be the closest practically possible spacing and what could be the best practically attainable dielectric material that I could coat each stationary plate with in order to have a strong dielectric to increase my capacitance ?
I read about Barium titanates and such materials as well as a class of materials called relaxor ferroelectrics, what could be the highest possible dielectric constants I could achieve? Could I be able to go above 1000K?

how high a capacitance do you want?
please expand on 1000K - is that capacitance? is it desired dielectric constant? units?

variable capacitors have been around for a long time
they were the first tuning element (for station selection) in old radios

I don't think that rotating variable capacitor with high Er is feasible. The capacitance would be more defined by the unavoidable air gap than the dielectric. We have low Er ceramic trimmer capacitors and mostly air dielectric variable capacitors

yes with K I was referring to dielectric constant, the reason I aim for the highest possible in life attainable dieletric constant is exactly because I have to compensate for the airgap being there,
as for the structure I would have stationary discs and rotary discs, the discs would have considerable size by size I mean diameter, say 30cm and 50cm and some maybe more , the capacitor part would be the outer part of the disc (because I need the inner part for other things) so just a stack of stationary circumferential discs and rotary discs and the overlapping area forms a capacitor, now since the center of the rotating discs would be held together by some plastic composite the outer part of the discs could be made very thin , overall I calculated one possible test scenario and came up with the area of parallel plates to be about 16000 cm2
further I would need to know how do you think what would be some real life maximum tolerances one could use for the airgap? 1mm or maybe bit less?
and given the two electrodes (discs) being so close together would it be possible to make some thin spray on film or otherwise on the stationary plates with some high dielectric constant material in order to maximize the capacitance?


Maybe pack your capacitor into liquid dielectric.


well i thought about that too, honestly i would need to know what best numbers i can get with superdielectrics in reality maybe i dont even need liquid or otherwise because i'm working with high frequency and so the reactance of the capacitor for given capacitance is lower.

the problem with liquid would be to contain it, it would want to move out from the place it is needed which is between the overlapping sections of the discs.


i'm working with high frequency
Maybe you could be a little bit more exact...
For one working with mains frequency ... 200Hz my be high frequency
Audio: 20kHz
RF: some GHz....


well i can't predict the exact range as that would have to be nailed down to simulation and or real life tests/calculations but overall in the Mhz range , say 1-100Mhz.

1 mm air gap? As a starting point, did you calculate the total effective Er of a combination of Er=1 and Er=1k or even 1000k?

If by Er you mean capacitance then yes, I did calculate the total capacitance given the same overlapping surface area and airgap with different dielectric constants.

so as an example if my airgap is 1mm and my area is 8000 cm2 with a dielectric of 1000K I can only get about 7 uF, having the same airgap but a dielectric of 10 000K, I already have 70 uF, that is already much much better , then by increasing the total area I can have almost satisfying numbers of capacitance.
the question is what is the highest probable dielectric constant one could achieve without some super sci-fi unobtanium and does 1mm airgap is too big or can I make a smaller one ?

Maybe an ambiguousity of the term air gap, I mean air actually. I'm assuming that the space between electrode plates won't be completely filled with dielectric, there's always an air gap between metal and dielectric. The higher the Er, the more the capacitance depends on the air gap.

sure, well that is why i'm thinking what is the smallest possible airgap I could have, given all tolerances are very tight and everything is made spot on.
I'm thinking ideally the real airgap which would be the empty space between the rotating disc circumference and stationary plate + the dielectric layer so the gap between the dielectric layer and copper plate should be much less than 1mm, say 0.3mm, the total gap including the dielectric layer could then be made some 0.5mm maybe, if the layer could be spray or otherwise coated in the stationary plate and made thin enough.

i guess the question also becomes which material could have the right mechanical as well as dielectric properties to serve as a strong dielectric because from the formula I see that just as much as the airgap also the K of the dielectric constant plays an important role here, having an extremely high constant makes thing better by alot.

when you calculated the capacitance, you talked about the area, 8000 cm^2
is that the area of 2 neighboring plates?

the rotary variable capacitor FvM has in post #3 consists of many pairs of neighboring pates
each pair of neighboring plates is in series with the next, so the net capacitance is

1/Cnet = N/C, where N is the number of neighboring plates and C is the capacitance from one set of plates

so Cnet = C / N

no in my case each plate is in parallel so I can simply add the capacitance, each two plate sets produce me about 400 cm2, so I count together all the plates I have in parallel and use their surface area to get the total area, since all plates are the same and will in theory have the same dielectric I can then calculate this as a single parallel plate capacitor.

so you're plan is to have 20 stationary copper discs, each about 400 cm2, side by side, (so you ave a plate area of about 8000 cm2) each with a rotary partner on its own shaft
so they are wired in parallel to add the capacitances

aluminum will work just as well, its less expensive, lighter, and easier to machine.
in fact, any metal will do.

since you're plan is to put some kind of dielectric on the stationary plate, you could make the dielectric about 1/2 mm thick, put a layer of
wax paper on the dielectric, and wax paper the mating surface of the rotating plate so the is essentially no air gap
when the rotating plate moves, the two layers of wax paper rub, the metal and the dielectric stays intact.

well I think it would have to be something different than wax paper, first of all because was paper doesn't introduce any large increase in dielectric constant and the other reason being is that the rotating discs would indeed rotate at a rather large rpm, see the thing here is this, the stationary plates are more like large circumferential planes with empty middles and the rotating discs are like discs but there is other machinery in the middle and the capacitor part is only at the outer edge where it meets the stationary side aka where they overlap. I think making the rotating outer disc parts thin enough to still be rigid and since they will rotate at some speed the angular momentum should fix the rest of stability issues, but overall I assume i would still need some very minor airgap, maybe there is some special liquid or liquous solution that would stick between the rotating plate and stationary dielectric which could increase the constant and at the same time serve as a sort of lubricant against the plates ever getting out of balance or rubbing.

the rpm could be something like 3000 rpm/min or so about.


Maybe some other things we should know?

How do you decide to handle the mechanical imbalance at 3000 RPM?
And fluid with 3000 RPM maybe is not a very good idea.

What is all this good for?


you haven't specified the capacitance range you're looking for

since 3000 RPM is 50 rev/sec, maybe there's a better or different way to produce a
50 Hz signal than this rotating capacitor

I read about Barium titanates and such materials as well as a class of materials called relaxor ferroelectrics, what could be the highest possible dielectric constants I could achieve? Could I be able to go above 1000K?

You want a variable capacitor? What is the minimum and maximum capacitance you are looking for?

They are commercially available; what is the special feature you are looking for?

The range is commonly between 10-500pf. The spacing determines the highest working safe voltage.

Well I don't think that there should be mechanical imbalance as the rotor should get balanced before assembly, now i'm not looking for 50hz as that would require very large capacitance in order to have very low reactance losses, but my working frequencies are much higher so the capacitance required is much less so I thought I could get away with such an arrangement.
i'm having an oscillator located on the rotating shaft so I thought I need to get it's output off somehow and using brushes makes the whole thing lossy and introduces problems so i thought since the frequency is rather high I could get away with capacitor type of pickup.
This is more like a thought idea for now because I am not still sure can i build a capacitor like this that would have large enough capacitance in order for the reactance to be low enough at the desired frequencies.
starting from 1 Mhz I would need about 400 uF in order to be good and going higher in frequency the capacitance requirement decreases of course.

Not open for further replies.

Part and Inventory Search

Welcome to