Capacitor Dielectrics for Transmitting Loop Antennas

[rickman]

Just pursuing some speculation about materials for tuning capacitors on
transmitting loop antennas. I was thinking of a capacitor formed by
fixing end plates on the two ends of the loop and adjusting by
stretching the entire loop to widen the gap. The main problem with this
is the limited high end of capacitance from the breakdown voltage
required, restricting the spacing which can be used to at least a few
mm. This in turn requires rather large plates. Then to get a
sufficient ratio of capacitance the gap must be widened to quite a few
inches.

One way to mitigate this is to use a dielectric material which will both
provide a higher permittivity as well as a higher breakdown voltage
allowing a smaller minimum plate gap and a much smaller plate size for
the same capacitance.

I found an example tuning capacitor using PTFE as a dielectric in a
trombone configuration. The owner says it works well. I'm not sure
where to get tubes of "virgin" PTFE and I'm not sure how stable this
material is over temperature and weather conditions.

Another material that appears to be a great dielectric is fused silica.
I can get that inexpensively in tube form in a variety of sizes and it
seems to have better properties than PTFE, especially the dielectric
coefficient temperature curve which is virtually flat from 25°C to 500
°C, so I assume it is also flat down to reasonable outdoor temps. It
does not absorb moisture and has a higher dielectric coefficient
allowing a smaller plate size than PTFE.

So I'm wondering why I can't find where anyone has used fused silica in
these capacitors. Maybe it would be a bit pricey in plate capacitors,
but it would be good and cheap in trombone style caps.

--

Rick

[Jeff]

On 29/01/2016 02:17, rickman wrote:
> Just pursuing some speculation about materials for tuning capacitors on
> transmitting loop antennas. I was thinking of a capacitor formed by
> fixing end plates on the two ends of the loop and adjusting by
> stretching the entire loop to widen the gap. The main problem with this
> is the limited high end of capacitance from the breakdown voltage
> required, restricting the spacing which can be used to at least a few
> mm. This in turn requires rather large plates. Then to get a
> sufficient ratio of capacitance the gap must be widened to quite a few
> inches.

There is a commercial loop that operates in this way (Ciro Mazzoni). It
does not use a single plate on each section, rather interleaving fingers
just like a normal air spaced capacitor.

I don't see why you could not add a sheet of dielectric to the plates,
it would also help with the alignment and prevent the plates from shorting.

Jeff

[rickman]

The question is which dielectric? Dielectrics tend to have thermal
dependencies to their dielectric coefficients and so will prevent you
from using an tuner presets. You also have to pick one with a low loss
tangent. Then it also needs to be available and affordable. That is
why I was asking about Teflon and fused silica. Teflon is available in
a variety of sizes and shapes, but I don't find tubing so it would have
to be machined. I'm not sure about the temperature coefficient of
Teflon. One person I know with experience with a trombone type
capacitor says it does not change tuning with temperature.

Fused silica has great characteristics across the board. I can find it
as tubes in sizes which may line up well with metal tubing you can find,
but sheets seem to be rather pricey. Not sure why. The dielectric
coefficient of Teflon is only around 2 while fused silica is more like
4. Both have a good breakdown voltage. So fused silica would be my
preference.

I have seen a video of the Cirro Mazzoni with interdigitated plates. I
figured a loop could be tuned this way, but I wasn't sure it was
practical until I saw that video. They use a linear actuator. I am
thinking for parallel plates this would need to be modified with a cam
to give very precise adjustment when the plates are close spaced and
less precision as they are pulled apart.

--

Rick

[Wimpie]

El 29-Jan-16 a las 03:17, rickman escribió:

The size reduction that you can get by using dielectric is limited,
unless you fully immerse the capacitor in oil (including any metallic
edges carrying RF HV).

The dielectric itself may have a high breakdown strength, but the
limitation will be air mostly. When making a parallel plate capacitor
with dielectric, there will be some air between the metallic plates and
the dielectric. With increasing voltage, the air between the dielectric
and the plate breaks down (partial discharge). That breakdown will
generate heat and destroys any plastic dielectric within short time.

Even with virtually no air between the plates and the dielectric, you
will have edges where you have an air interface between metal and
dielectric. Smooth round edges increase breakdown voltage, but in the
end you will get breakdown starting from the metallic edge towards the
dielectric.

Other issue is tracking, especially with wet contamination. In that case
breakdown goes via a contaminated surface (the creepage path).

I make my own parallel plate capacitors with PE sheet dielectric for
fixed matching networks and band-pass filters. When testing with RF HV,
It is always the air that limits the safe RF working voltage.

There is something nice, breakdown fieldstrength in air increases with
decreasing gap. You may search for Paschen Curve, peek's formula, corona.

Wim, PA3DJS
Please instruct your racing pigeon to remove abc when using email.

[Brian Howie]

I had a similar problem with a trombone type gamma match capacitor
on a 50MHz yagi, I tried a narrow bore pipe with both polythene and
PFTE taken from semi-rigid coax. I found that I got breakdown in the
small airgap much above 100W, which vapourised the dielectric . I
ended up using a larger diameter tube and some llarger diameter PTFE
dielectric I got from some UR 67 sized coax.

Silica would probably work quite well , but you may still get air
break down , depending on geometry.. You will have estimate the
voltage and the division of this voltage between the air and
dielectric regions .

I can see why vacuum variable Cs are used.

Brian GM4DIJ

[rickman]

I have no idea why you think this is a plausible scenario. The
dielectric reduces the strength of the field between the plates and so
increases the voltage before a discharge occurs. I know this because
others have built such caps not to mention adjustable caps are
commercially available with dielectric.

> Even with virtually no air between the plates and the dielectric, you
> will have edges where you have an air interface between metal and
> dielectric. Smooth round edges increase breakdown voltage, but in the
> end you will get breakdown starting from the metallic edge towards the
> dielectric.

You will only have a discharge if you exceed the maximum voltage of the
cap.

> Other issue is tracking, especially with wet contamination. In that case
> breakdown goes via a contaminated surface (the creepage path).

Yes, contamination can be an issue with any cap exposed to the elements
regardless of what dielectric is used.

> I make my own parallel plate capacitors with PE sheet dielectric for
> fixed matching networks and band-pass filters. When testing with RF HV,
> It is always the air that limits the safe RF working voltage.

When using a dielectric you can make the air gap as large as you wish.
There is nothing to set a limit on the length of the air gap except for
the space you have available.

> There is something nice, breakdown fieldstrength in air increases with
> decreasing gap. You may search for Paschen Curve, peek's formula, corona.

I am familiar with Paschen's law. But the narrowing of the gap is not
useful in STP air until a *very* small gap is achieved. One graph I've
found shows the voltage minimum at STP is around 40 microns. So you
need a gap smaller than this to get the voltages needed. Very hard to
do with parallel plates.

--

Rick

[rickman]

I'm not following this at all. If the dielectric is good for say, 25 kV
and you apply less than that to it, how would the air gap make the
dielectric break down?

--

Rick

[Wimpie]

El 29-Jan-16 a las 19:33, rickman escribió:

This is just from theoretical modelling, simulation and actual
measurement. We are discussing AC not DC, so you get a capacitive
voltage division across the dielectric and the air between the metal and
the dielectric.

If the breakdown voltage for a specific arc path is reached, the air
breaks down. depending on the geometry, it starts with a blue glow, or a
direct full air breakdown from metal to dielectric. The spark erodes the
dielectric and then you get a full breakdown from plate to plate. To
some extent, I like destructive testing...


So, adding dielectric helps (I do it also), but it doesn't give the
improvement many people think.

>
>> Even with virtually no air between the plates and the dielectric, you
>> will have edges where you have an air interface between metal and
>> dielectric. Smooth round edges increase breakdown voltage, but in the
>> end you will get breakdown starting from the metallic edge towards the
>> dielectric.
>
> You will only have a discharge if you exceed the maximum voltage of the
> cap.

Agree, but there can be a big difference between DC and AC voltage ratings.

>
>
>> Other issue is tracking, especially with wet contamination. In that case
>> breakdown goes via a contaminated surface (the creepage path).
>
> Yes, contamination can be an issue with any cap exposed to the elements
> regardless of what dielectric is used.
>
>
>> I make my own parallel plate capacitors with PE sheet dielectric for
>> fixed matching networks and band-pass filters. When testing with RF HV,
>> It is always the air that limits the safe RF working voltage.
>
> When using a dielectric you can make the air gap as large as you wish.
> There is nothing to set a limit on the length of the air gap except for
> the space you have available.
>

If your air gap is large (compared to dielectric thickness), adding the
dielectric has reduced to no benefit (except mechanically). You want as
much as possible field (Integral (E*ds) ) in the dielectric to have
maximum benefit from it.

>
>> There is something nice, breakdown fieldstrength in air increases with
>> decreasing gap. You may search for Paschen Curve, peek's formula, corona.
>
> I am familiar with Paschen's law. But the narrowing of the gap is not
> useful in STP air until a *very* small gap is achieved. One graph I've
> found shows the voltage minimum at STP is around 40 microns. So you
> need a gap smaller than this to get the voltages needed. Very hard to
> do with parallel plates.
>

Maybe we have some noise in our communication: reducing the path length
through air, increases the breakdown E-FIELD strength. Higher allowable
E-FIELD means larger D (D = E*eps0 for air). As D is continuous across
an interface, you also get more D in the dielectric (hence more stored
energy, more benefit).

If you start with trombone style capacitors, make sure you have large
rounded edges, otherwise they will fail soon under RF stress, even with
the best dielectric. I'm sure you know that small bending radius results
in high field strength.

Good luck with the kV capacitor design,

Wim, PA3DJS

[rickman]

I am thinking if I have a dielectric rated for voltage X and the
capacitor plates range from a separation which is the same as the
dielectric thickness to a larger value, it will provide at a minimum the
voltage X provided by the dielectric thickness.

What do you think people are thinking?

>>> Even with virtually no air between the plates and the dielectric, you
>>> will have edges where you have an air interface between metal and
>>> dielectric. Smooth round edges increase breakdown voltage, but in the
>>> end you will get breakdown starting from the metallic edge towards the
>>> dielectric.
>>
>> You will only have a discharge if you exceed the maximum voltage of the
>> cap.
>
> Agree, but there can be a big difference between DC and AC voltage ratings.

Do you have any data on this? I've yet to see any capacitors, air or
vacuum, rated for either AC or DC voltage only. If they don't specify,
how would you know which it was?

>>> Other issue is tracking, especially with wet contamination. In that case
>>> breakdown goes via a contaminated surface (the creepage path).
>>
>> Yes, contamination can be an issue with any cap exposed to the elements
>> regardless of what dielectric is used.
>>
>>
>>> I make my own parallel plate capacitors with PE sheet dielectric for
>>> fixed matching networks and band-pass filters. When testing with RF HV,
>>> It is always the air that limits the safe RF working voltage.
>>
>> When using a dielectric you can make the air gap as large as you wish.
>> There is nothing to set a limit on the length of the air gap except for
>> the space you have available.
>>
> If your air gap is large (compared to dielectric thickness), adding the
> dielectric has reduced to no benefit (except mechanically).

This is not clear at all. What are you trying to describe?

> You want as
> much as possible field (Integral (E*ds) ) in the dielectric to have
> maximum benefit from it.
>>
>>> There is something nice, breakdown fieldstrength in air increases with
>>> decreasing gap. You may search for Paschen Curve, peek's formula, corona.
>>
>> I am familiar with Paschen's law. But the narrowing of the gap is not
>> useful in STP air until a *very* small gap is achieved. One graph I've
>> found shows the voltage minimum at STP is around 40 microns. So you
>> need a gap smaller than this to get the voltages needed. Very hard to
>> do with parallel plates.
>>
>
> Maybe we have some noise in our communication: reducing the path length
> through air, increases the breakdown E-FIELD strength. Higher allowable
> E-FIELD means larger D (D = E*eps0 for air). As D is continuous across
> an interface, you also get more D in the dielectric (hence more stored
> energy, more benefit).

I'm not sure what you see as the benefit of the dielectric. I see it as
allowing a closer spacing of the plates which along with the higher
relative permittivity provide a higher capacitance for the same plate
size, or a smaller plate to get the same capacitance.

The material will have a breakdown voltage for a given thickness. I
can't see how widening the gap with an additional air dielectric (in
order to adjust the capacitance) could possibly cause a breakdown of the
dielectric.

> If you start with trombone style capacitors, make sure you have large
> rounded edges, otherwise they will fail soon under RF stress, even with
> the best dielectric. I'm sure you know that small bending radius results
> in high field strength.

Yes, that is a given regardless of the dielectric chosen, air, vacuum or
otherwise as well as any capacitor configuration, trombone, plate or
butterfly.


Maybe an illustration will be clearer. This is the plate type capacitor.

Plates together, high capacitance, some voltage rating X.

Dielectric
/
||X||
||X||
||X||
||X||
----------+|X|+-------------
||X|| Loop Conductor
----------+|X|+-------------
||X||
||X||
||X||
||X||
/
Capacitor Plate


Plates spread, low capacitance, a voltage rating > X.

Dielectric
/
||X| |
||X| |
||X| |
||X| |
----------+|X| +-------------
||X| | Loop Conductor
----------+|X| +-------------
||X| |
||X| |
||X| |
||X| |
/
Capacitor Plate

Is this more clear?

--

Rick

[Wimpie]

El 29-Jan-16 a las 21:44, rickman escribió:

Your ASCII art makes sense!

Some calculation to show you how the RF voltage divides between
dielectric and air. If I made a mistake, I hope someone will correct me...

Let us assume 1 mm PE dielectric (er = 2.54) and 1 mm air dielectric (er
= 1), surface area A is 1m2 (just to simplify calculations). Distance
between plates is 2 mm.

Capacitance formula (ignoring fringe effects):

C = 8.854e-12*er*A/s

Capacitance of dielectric is 22.5 nF
Capacitance of air is 8.85 nF
total capacitance (two in series) is 6.35 nF

Now assume 10 kV RF between the plates, how does the voltage divide?

Say we use 1 MHz for the calculation then we have two impedances in series:

dielectric: 7.07 Ohm (capacitive)
air: 18.0 Ohm (capacitive)
total: 25.1 (also capacitive)

So 2.8 kV appears across the dielectric, and 7.2 kV appears across the
airgap. 7.2 kV across a 1 mm airgap will breakdown (breakdown voltage is
around 4.5 kV). This breakdown will destroy the PE (or PTFE) dielectric.
In addition, you don't want partial discharge in an RF matching component.

If you do the calculation for 0.05 mm air gap, the voltage divides as:
8.6 kV across the dielectric, 1.35 kV across the air interface. This
interface will also breakdown (around 0.6 kV).

How to get something working?

you need a thicker dielectric! This changes the voltage division so that
more voltage is across the dielectric, and less across the air gap.

If you use 2.5 mm dielectric and 0.05 mm air, the gap may temporary
survive the 10 kV RFp (in a real application I would include more margin).

However 2.5 mm dielectric and 1 mm air (total 3.5 mm) will breakdown as
more then half the capacitor voltage appears across the 1 mm air gap. In
a real capacitor with large plate distance, you get a non-uniform field,
and that breaks down at a lower voltage then based on the Paschen curve.

conclusion:
The dielectric may have a strength of say 20 kV, but in the above
example, the capacitor wil breakdown below 10 kV.

When designing my RF capacitors, I focus on the voltage stress in air
gaps (especially near edges).



--
Wim, PA3DJS
When using a racing pigeon for mail, tell him/her to remove abc

[Roger Hayter]

rickman <gnu...@gmail.com> wrote:

snip

That's certainly what I thought you meant. However, at RF will you not
have a capacitative potential divider? The DC voltage which the
dielectric will resist is quite irrelevant, because at AC the larger
capacitance of the dielectric will have less AC voltage across it than
the lower capacitance of the air. So it is the breakdown voltage of the
air gap which will be relevant. So the dielectric will apparently *not*
enable you to reduce the spacing significantly below what would be safe
without the dielectric (but it will enable you to have a larger
capacitance for a given spacing). That makes sense to me, anyway.

Perhaps a vacuum capacitor would be best, as it would be low loss as
well as high voltage.






--

Roger Hayter

[rickman]

Ok, I understand what you are saying now. I'm not certain the result is
what you suggest. Working with real numbers, air has a breakdown
voltage of 3 kV/mm. The dielectric would be about 1 mm thick and have a
breakdown voltage of 25 kV. If the capacitor is spaced an additional 1
mm from the dielectric so the air has the same thickness as the
dielectric it would have about 1/4 the capacitance of the dielectric and
so "carry" 4/5 of the voltage applied. If this exceeds the 3 kV rating
of the air you are suggesting that the air would arc and somehow cause a
breakdown of the dielectric. That is what I don't follow.

Would this happen because of ions bombarding the surface of the
dielectric? That still does not explain how any current would flow
through the dielectric even if the surface is struck with ions. The
rating of the dielectric would prevent the flow of current through that
material, no? Without significant current flow there won't be any
significant damage to the dielectric.

What may happen is that with the air gap breakdown voltage exceeded, it
will no longer behave as a capacitor preventing the capacitance
adjustment I am expecting until the gap width increases to a value that
will stand up to the applied voltage.

If the principles you are describing would cause the breakdown of a
dielectric, a conventional trombone capacitor could not be constructed
with a narrow, fixed air gap and a plastic dielectric. I know this is a
very workable approach, so I can't believe your reasoning.

I'm much more concerned with the impact of the elements on capacitors
with moving parts. I imagine it doesn't take much dirt or other debris
to prevent the proper operation of moving parts with small gaps.

--

Rick

[c...@post.netunix.com]

rickman <gnu...@gmail.com> wrote:
> Just pursuing some speculation about materials for tuning capacitors on
> transmitting loop antennas. I was thinking of a capacitor formed by
> fixing end plates on the two ends of the loop and adjusting by
> stretching the entire loop to widen the gap. The main problem with this
> is the limited high end of capacitance from the breakdown voltage
> required, restricting the spacing which can be used to at least a few
> mm. This in turn requires rather large plates. Then to get a
> sufficient ratio of capacitance the gap must be widened to quite a few
> inches.

The correct capacitor for the job is a Vaccum Transmitting Capacitor.
Expensive but the correct item for the job.

[Roger Hayter]

rickman <gnu...@gmail.com> wrote:

> On 1/29/2016 4:41 PM, Roger Hayter wrote:

snip

> >
> > That's certainly what I thought you meant. However, at RF will you not
> > have a capacitative potential divider? The DC voltage which the
> > dielectric will resist is quite irrelevant, because at AC the larger
> > capacitance of the dielectric will have less AC voltage across it than
> > the lower capacitance of the air. So it is the breakdown voltage of the
> > air gap which will be relevant. So the dielectric will apparently *not*
> > enable you to reduce the spacing significantly below what would be safe
> > without the dielectric (but it will enable you to have a larger
> > capacitance for a given spacing). That makes sense to me, anyway.
> >
> > Perhaps a vacuum capacitor would be best, as it would be low loss as
> > well as high voltage.
>
> Ok, I understand what you are saying now. I'm not certain the result is
> what you suggest. Working with real numbers, air has a breakdown
> voltage of 3 kV/mm. The dielectric would be about 1 mm thick and have a
> breakdown voltage of 25 kV. If the capacitor is spaced an additional 1
> mm from the dielectric so the air has the same thickness as the
> dielectric it would have about 1/4 the capacitance of the dielectric and
> so "carry" 4/5 of the voltage applied. If this exceeds the 3 kV rating
> of the air you are suggesting that the air would arc and somehow cause a
> breakdown of the dielectric. That is what I don't follow.

I have no experience of this, so I'm inclined to believe the poster who
says that he has. I can see it being a Bad Thing for the reasons you
describe below, not least because while the capacitor isn't being a
capacitor the transmitter is probably seeing a significant mismatch.

>
> Would this happen because of ions bombarding the surface of the
> dielectric? That still does not explain how any current would flow
> through the dielectric even if the surface is struck with ions. The
> rating of the dielectric would prevent the flow of current through that
> material, no? Without significant current flow there won't be any
> significant damage to the dielectric.

Current flow in capacitors has always been a mystery to me, but I have
no doubt that capacitor current flows into the plates and out again.
And it will certainly be flowing in the arc, which will have a
non-reactive and dissipative component - it will get very hot. I can
see this singeing the surface of the dielectric and perhaps sputtering
stuff from the metal plate on it, perhaps carbonising an organic
plastic. Then this will be a hot spot in the next arc - I can see this
progressing. Probably your silica will last a longer then PTFE, but
localised thermal shock could cause spalling. I can certainly see the
possibility of breakdown though.

>
> What may happen is that with the air gap breakdown voltage exceeded, it
> will no longer behave as a capacitor preventing the capacitance
> adjustment I am expecting until the gap width increases to a value that
> will stand up to the applied voltage.
>
> If the principles you are describing would cause the breakdown of a
> dielectric, a conventional trombone capacitor could not be constructed
> with a narrow, fixed air gap and a plastic dielectric. I know this is a
> very workable approach, so I can't believe your reasoning.

The reasoning would seem to dictate an air gap/voltage rating sufficient
to avoid arcing even without the dielectric.

>
> I'm much more concerned with the impact of the elements on capacitors
> with moving parts. I imagine it doesn't take much dirt or other debris
> to prevent the proper operation of moving parts with small gaps.


--

Roger Hayter

[rickman]

I would believe data that is presented, but I think you are talking
about Brian's post about a trombone gamma match capacitor. He talks
about an applied wattage, but not the voltage on the cap or info on the
expected voltage breakdown of the cap. How can a valid conclusion be
drawn from that?

> I can see it being a Bad Thing for the reasons you
> describe below, not least because while the capacitor isn't being a
> capacitor the transmitter is probably seeing a significant mismatch.

I didn't say the entire capacitor would not act as a capacitor. I said
the air gap might not support the field so the entire field is across
the dielectric so the capacitance remains unchanged until the gap is
wide enough. This would simply limit the range of the capacitor
although that is a factor to be considered.

>> Would this happen because of ions bombarding the surface of the
>> dielectric? That still does not explain how any current would flow
>> through the dielectric even if the surface is struck with ions. The
>> rating of the dielectric would prevent the flow of current through that
>> material, no? Without significant current flow there won't be any
>> significant damage to the dielectric.
>
> Current flow in capacitors has always been a mystery to me, but I have
> no doubt that capacitor current flows into the plates and out again.
> And it will certainly be flowing in the arc, which will have a
> non-reactive and dissipative component - it will get very hot.

I think that is the error. There is no current flow until the voltage
on the entire path from plate to plate starts to conduct. The
dielectric won't conduct until it's breakdown voltage is exceeded. No
arc, no heating, no destruction.

> I can
> see this singeing the surface of the dielectric and perhaps sputtering
> stuff from the metal plate on it, perhaps carbonising an organic
> plastic. Then this will be a hot spot in the next arc - I can see this
> progressing. Probably your silica will last a longer then PTFE, but
> localised thermal shock could cause spalling. I can certainly see the
> possibility of breakdown though.
>
>
>>
>> What may happen is that with the air gap breakdown voltage exceeded, it
>> will no longer behave as a capacitor preventing the capacitance
>> adjustment I am expecting until the gap width increases to a value that
>> will stand up to the applied voltage.
>>
>> If the principles you are describing would cause the breakdown of a
>> dielectric, a conventional trombone capacitor could not be constructed
>> with a narrow, fixed air gap and a plastic dielectric. I know this is a
>> very workable approach, so I can't believe your reasoning.
>
> The reasoning would seem to dictate an air gap/voltage rating sufficient
> to avoid arcing even without the dielectric.

An air gap which is sufficient to operate properly. I'm not sure the
air gap will indeed cause a problem. The fact that trombone caps which
have some small space between the dielectric and at least one of the
conductors work just fine would indicate no big problems.

The impact of the combined dielectric and air gap is something to be
explored.

--

Rick

[Brian Howie]

On Fri, 29 Jan 2016 22:47:32 -0500, rickman <gnu...@gmail.com> wrote:

>I would believe data that is presented, but I think you are talking
>about Brian's post about a trombone gamma match capacitor. He talks
>about an applied wattage, but not the voltage on the cap or info on the
>expected voltage breakdown of the cap. How can a valid conclusion be
>drawn from that?

The only data I had was the power into the antenna. I was using a
Bird Thru-line to set the gamma match* for best SWR at low power. I
then started winding the power up to 400W. At 100W the gamma-match
failed . The power in will relate to the voltage at the capacitor ,
although it was not possible for me to estimate it. It's likely the
breakdown occured at the end of the tube where the corners are
sharpest and the electric field highest . Bevelling the end of the
tube might have helped a bit.

*strictly speaking it was a Clemens match.I started with UR58 but
ended up with bigger diameter tubing and thicker PTFE line from a
larger diameter coax.

http://www.iw5edi.com/ham-radio/files/clemens1.gif

Brian

[rickman]

I reread your prior post and noticed how you got it to work... "I

ended up using a larger diameter tube and some llarger diameter PTFE
dielectric I got from some UR 67 sized coax."

That would imply the dielectric used was thicker, is that correct? If
so, I think that shows the original attempt failed because the breakdown
voltage of the insulation was too low. I imagine you had a small air
gap in both cases. The details may be too incomplete to draw any
conclusions about why the first attempt did not work.

--

Rick

[Wimpie]

@Rickman:

I can ignite a fluorescent tube, without making any physical contact
with the lamp, just via capacitive coupling directly through the glass
envelope (= dielectric). It is just a matter of exceeding the breakdown
voltage of tha gas.

Once my father became angry. I showed him a small CW solid state Tesla
Coil setup. There was a wasted straight fluorescent tube standing in a
corner. When increasing the power fed to the coil, the tube ignited...

The difference with the ionization in the lamp is that in air you
quickly get a thin channel that carries all the current. The discharge
ends at the dielectric (and spreads out). In the end the discharge will
punch through.

Even if the dielectric wasn't distroyed, you don't want partial
discharge in an RF component.

I would suggest you to run some experiments. With LC resonant circuits
and inductive coupling, you can generate HV (like your proposed Loop
Antenna). Based on the VSWR bandwidth, inductance and RF input power,
you can calculate the voltage between the plates.


--
Wim, PA3DJS
Please delete abc before hitting the send button

[rickman]

On 1/30/2016 4:28 AM, Jeff wrote:
>
>>
>> Dielectric
>> /
>> ||X| |
>> ||X| |
>> ||X| |
>> ||X| |
>> ----------+|X| +-------------
>> ||X| | Loop Conductor
>> ----------+|X| +-------------
>> ||X| |
>> ||X| |
>> ||X| |
>> ||X| |
>> /
>> Capacitor Plate
>
>

> My suggestion was more like this, looking vertically down on the capacitor:
>
> +___________
> ----------+|XXXXXXXXXXX +-------------
> ||x-----------| Loop Conductor
> ----------+|XXXXXXXXXXX +-------------
> +----------
>
>
> +__________
> ----------+|XXXXXXXXXXX +-------------
> ||x -----------| Loop Conductor
> ----------+|XXXXXXXXXXX +-------------
> +----------
> There is not meant to be an air gap between the left-hand lower plate
> and the dielectric, and the RH plate should fill the space between the 2
> dielecrtics, that is a function of my ASCII drawing.
>
> The air gaps are minimized and provide a guide for the plates, the still
> may be the possibility for breakdown at the ends.

Until you get a spacing between the moving plate and the dielectric of
significantly less than 40 microns you are lowering the breakdown
voltage of the air. This is why I don't put much credence in the idea
that the air gap voltage will impact adversely the breakdown voltage
between the plates through both dielectrics (air and the other material).


> The idea may be extended to more plates if required.

I have considered a butterfly capacitor like this. The plates would be
as in a conventional capacitor with dielectric between each pair. The
adjustments are rather fine with that construction. Reduction gearing
is needed with backlash control. But I have never been convinced the
construction allows for low loss in the conductors. A multi-plate
capacitor is far too complex to analyze properly... at least for me.
The current flows want to stay on the surface, but many of the plates
would be fed by current through the bulk of the conductor which makes me
question just how that would work and whether the flow would be pinched
by a combination of the skin effect and the proximity effect resulting
in higher resistive losses.

A capacitor formed by single plates on the end of the loop conductor
would be ideal for current flow into the capacitor. The current can
spread radially outward from the loop conductor with less resistance
than any other method of distribution. That is why I am contemplating
this.

--

Rick

[Brian Reay]

That was one of my first thoughts but I assume rickman needs a larger
value Cap. than he can readily achieve with a vacuum Cap., possibly on
either size or cost grounds.

[Brian Reay]

If I've understood your diagrams correctly, when you are in the 'low'
capacitance state, haven't you effectively got two Cs in series?

One with your chosen dielectric and one with an air dielectric. The
being a notional extra 'plate' on the surface of the dielectric. If you
think about it, it is the charge on the plates which is relevant, no
that they are metal.

[mm0fmf]

Here's a TX loop using home made air gap capacitor.

http://moosedata.com/loopcap.jpg

The tube material is 75mm OD. A satelitte TV dish positioner screwjack
was used to stretch and compress the loop itself and adjust the
capacitor. That gave a fast to slow tuning speed depending on the
applied voltage and could be controlled from the shack. I though I had a
picture of the whole loop but can't find it.

[rickman]

On 1/30/2016 11:53 AM, Jeff wrote:

> The configuration that I am suggesting is not a butterfly capacitor. It
> is a normal design where there are only 2 sets of plates.
>
> I don not understand what you are getting at when you say that there is
> some sort of current flow problem. What I am suggesting is no different
> to using a multi plate capacitor in a conventional design. It is only
> that added dielectric that is different which would result in physically
> smaller capacitor plates for a given capacitance.
>
> There is nothing new in what I am suggesting, many tuning capacitors in
> MW receivers use multi plate capacitors with layers of dielectric
> between the plates.

Sorry, I wasn't clear. I didn't mean to say your idea is a butterfly
capacitor, but that I had looked at butterfly capacitors. What you show
above is what is done in the Ciro Mazzoni antenna except they just use
an air dielectric.

My point is simply that such designs have difficult to analyze current
flow. I have not tried one myself to see how high the efficiency can
be. Many dwell on optimizing the resistance of the loop and then use a
capacitor design that is very suboptimal including the vacuum variable
caps. Those capacitors are great, but they connect by clamps. While
many building transmitting loop antennas go to extreme lengths to
prevent any additional resistance no matter how small, the ignore the
relatively high resistance in making clamp connections to the capacitor!

My idea is about minimizing the connection and even the internal
resistance of every capacitor design I have seen. The interdigitated
plates may have a very low resistance, but I can't begin to analyze the
skin and proximity effects of that design. The simple parallel plates
on the ends of the loop conductor is much easier to analyze and I can
see it would provide the lowest possible resistive losses.

The parallel plates with dielectric may have a problem from the air
breakdown, I don't know for sure. I suspect this is not really a
problem because the same problem would be expected in many trombone
designs. However, unless those designs were analyzed in detail I can't
say for sure. It may be that with the very close spacing of the air gap
the capacitance of that portion of the gap is high enough to keep the
voltage across it low.

--

Rick

[rickman]

As I just posted elsewhere in this thread, I am wondering what so many
folks obsess over reducing the resistance in the loop and ignore the
resistive looses in the capacitor. It seems to me the clamp connections
of a vacuum variable is the worst possible resistively speaking. I am
trying to work out a capacitor which clearly has a very low resistive
component including the connection to the cap.

--

Rick

[rickman]

Yes, this is very similar to the Ciro Mazzoni antenna. I think this
design has the potential for having exceedingly low resistive losses.
But the configuration of the capacitor is very hard to analyze in a
meaningful way and may actually have a higher RF resistance than would
be desired when the loop itself is one piece bent tubing of a large
diameter.

Is this your antenna? Did you weld it yourself?

I'm thinking of an alternative configuration which would get around the
potential problems with the variable air gap. It would be like a
trombone, but instead of two parallel capacitors it would be a single
one with a rather large diameter and short length. It provides a very
simple, direct, low resistance connection to the loop conductor. The
mechanics would have to be worked out since expanding the loop would
change the angle of alignment and potentially make it jam. So the throw
would need to be kept short and enough gap provided to allow the loop to
be flexed to adjust the capacitance. It would likely be difficult to
find a large enough diameter dielectric, so may prove impractical.

Another arrangement I have considered would be dual tube like the
trombone, but not side by side, rather in line. There would need to be
a straight section in the loop at the capacitor site with a gap large
enough to withstand the breakdown voltage in air. A dual stator
capacitor is created by a slider of dielectric material a close fit
inside the straight section of loop with a tight liner. The slider tube
would move across the region around the gap similar to a trombone design
but rather than both series caps changing values together, one would
increase while the other decreases.

The issue I see with this design is finding dielectric materials of the
right size for the job. This might be easier to do with Teflon than
fused silica.

Anyone know how the voltage across the capacitor varies as the capacitor
is tuned across different frequencies with constant power? I'm thinking
there does not need to be a constant breakdown voltage at all
frequencies. Rather I expect a higher voltage rating is needed with
smaller capacitor values and so higher frequencies. There is an antenna
calculator, I'll have to check that.

--

Rick

[Wimpie]

El 30-Jan-16 a las 20:24, rickman escribió:

You will save yourself a lot of time (and maybe frustration) by just
assuming that the limitation in your design will be corona and/or air
breakdown.

Problems (corona discharge followed by full break down) arise at the
edges of the capacitor plates, as there you have the highest E-field and
longest path. So reducing the air gap in the plane or cylindrical
sections to a minimum, will not help you much. You will get an
increasing air gap when travelling towards the (rounded) edges.

I don't see your point with the "efficiency" of an interdigital parallel
plate capacitor with welded connections. If you start with some
ballpark calculation, you will see that the contribution of the
capacitor AC resistance is negligible (compared to the loop AC resistance).



--
Wim, PA3DJS
Feed your pigeon well, and let him/her remove abc from the address in
case of PM.

[rickman]

On 1/30/2016 5:38 PM, Wimpie wrote:
>
> I don't see your point with the "efficiency" of an interdigital parallel
> plate capacitor with welded connections. If you start with some
> ballpark calculation, you will see that the contribution of the
> capacitor AC resistance is negligible (compared to the loop AC resistance).

Can you explain what you mean by "ballpark calculation"? My point is
that the various effects involved in calculating the resistance of an
interdigital capacitor are too complex to even estimate a resistance.

The loop resistance can be easily calculated. It can be made very small
by using a large diameter conductor for the loop. Looking at the
various conductors used to design multiplate capacitors I see lots of
opportunities for large AC resistance to appear. Current flow in
rectangular conductors is primarily in the corners and so present a much
higher resistance than does a round conductor. All the multiplate
capacitors I have seen are made of these rectangular conductors from the
loop conductor to the plates and of course the plates themselves.

The one exception would be the vacuum variable capacitors which use
concentric cylinders providing the optimal path for current flow to and
along each of the cylinders. But of course they are connected to the
antenna loop by clamps which have a lot of resistance compared to
soldered or welded connections possible with other types.

--

Rick

[rickman]

On 1/31/2016 5:06 AM, Jeff wrote:
>
>> The one exception would be the vacuum variable capacitors which use
>> concentric cylinders providing the optimal path for current flow to and
>> along each of the cylinders. But of course they are connected to the
>> antenna loop by clamps which have a lot of resistance compared to
>> soldered or welded connections possible with other types.
>>
>

> For some reason you seem to be making the assumption that a clamp
> connection will be worse than a soldered one. There is a lot of evidence
> for example that crimped connections are lower resistance than soldered
> ones, and evidence that soldered joints are worse than clamp joints in
> loops (at least before exposure to the weather).

I have seen *no* evidence that clamped joints are better than soldered
joints much less welded joints. But largely this is an apples and
oranges thing. Clamp joints are the only way to connect VVCs without
doing damage from what I have seen so there are no soldered or welded
joints to compare. Conversely I have never seen anyone use or want to
use a clamp type joint for any other part of a transmitting magnetic loop.

As to the resistance of a crimped joint (which is not really relevant
since that is not used in transmitting antennas) the rating you are
talking about is a DC rating. The quality of the joint is purely from
the region of the crimp which requires current flow along a longer
surface path or one through the bulk of the connector which is resisted
by the skin effect. Have you seen crimp joints in RF transmitter apps?


> As to welded joints I suspect that the resistance is very dependant on
> the standard of the joint.

Not sure what you mean by "standard of the joint". If you mean quality,
isn't that true of all connections, quality is important.

--

Rick