Call Search
     

New to Ham Radio?
My Profile

Community
Articles
Forums
News
Reviews
Friends Remembered
Strays
Survey Question

Operating
Contesting
DX Cluster Spots
Propagation

Resources
Calendar
Classifieds
Ham Exams
Ham Links
List Archives
News Articles
Product Reviews
QSL Managers

Site Info
eHam Help (FAQ)
Support the site
The eHam Team
Advertising Info
Vision Statement
About eHam.net

   Home   Help Search  
Pages: [1] 2 Next   Go Down
  Print  
Author Topic: Building a butterfly capacitor  (Read 6755 times)
JAHAM2BE
Member

Posts: 277


WWW

Ignore
« on: September 04, 2012, 06:26:36 AM »

As part of a(nother) magnetic loop antenna project, I'm considering building my own butterfly capacitor. After reading several articles on the subject, I think it doesn't sound impossibly difficult, since the two stators and the rotor can be assembled separately and combined at the end, and actually looks like a bit of fun. I think a good butterfly capacitor may be the last obstacle standing between me and my dream of a compact, multi-band, efficient, and remotely-tunable antenna.

I'm wondering about the ideal dimensions of a butterfly capacitor in order to achieve certain goals, and how much loss might happen when we depart from the ideal dimensions.

The goals for my capacitor are:

1) Minimum loss (obviously), and

2) A very low minimum capacitance (ideally in the 10pF rance) and a high maximum capacitance (200pF or so).

My interpretation of what I've read is that the "ideal" form factor is cubical (equal side dimensions and balanced) for minimizing loss (goal 1). Two questions:

1) Is there any ideal form factor with regard to goal 2, ensuring low minimum capacitance and a large capacitance swing?

2) How bad is it to depart from the ideal cubical form factor? In particular, I may be cutting the capacitor plates by hand from sheet metal (copper or aluminum). In that case, I would like to make larger plates (e.g. 10cm diameter) so the handwork becomes easier. However this may lead to a stout capacitor with a small height (due to few stacked plates) but large width and depth (due to large-diameter plates). Will that degrade performance enough to worry about?

I'll be running 5 watts CW for now, so I expect a 2 kV plate spacing should be more than adequate. To reduce the risk of corona discharge, I'll probably sand down the plates with fine automotive sandpaper - not as good as weeks-long tumbling in walnut shells, but one must make do with the tools on hand. Also, I'll solder the stator plates to the mounting posts.

Finally, one last question: is it realistic to expect to be able to exceed 10,000 Q with a homebrew butterfly variable capacitor as described above?

In the absence of any data, I would probably arbitrarily pick a plate size, cut some plates, and hook up the capacitor to my loop to experimentally determine the frequency coverage, then add/remove plates as needed. But if there is a more systematic design procedure to minimize loss and maximize range, I'd love to hear about it. I suspect one answer may be, "model the physical system in FEM software and observe the simulated RF current flow," an avenue I would one day like to pursue.

I know some folks here on eham build their own variable capacitors, so I hope you might give me some advice. Thanks in advance.
Logged

N3QE
Member

Posts: 2137




Ignore
« Reply #1 on: September 04, 2012, 08:31:27 AM »

I've seen articles in QEX about homebrewing variable capacitors and I find a couple articles on the web too.

You know when I first started as a ham the "old timers" made fun of us young ones because we went to the store to buy resistors and capacitors instead of making them from scratch like they did in their heydays. It's kind of refreshing to see this true homebrewing coming back!

Typical spacing seems to be kinda wide because of plate-to-plate variations but maybe that's good if you need high voltage ratings. You probably don't really get true high voltage caps unless you have some way of microfinishing all the cut sharp edges so nothing rough is left. I don't think it's just a matter of the making the flat surfaces smooth but if you look at the good old air variables I can see some sort of very consistent rounding has been applied to the edge of each plate, I don't know exactly how this is achieved.

10 to 200 pF sounds like a wider range than what you could actually achieve. Those strays add up and likely you'd end up with 30 pF of stray capacitance on a 200pF max unit.

I think MFJ sells some new butterfly caps. 282-2016 and 282-2017. Seem too large for a 5W tuner but a homebrew unit would probably end up with really wide spacing too unless you were really good at micromachinig all the plates and parts.


Logged
W9GB
Member

Posts: 2613




Ignore
« Reply #2 on: September 04, 2012, 01:04:54 PM »

April 2003, eHam.net has an article on building air-variable capacitors
http://www.eham.net/articles/5217

August 2009,
http://www.instructables.com/id/air-variable-capacitor-from-scrap-aluminum-sheets/

Magnetic Loop
http://www.cvarc.org/tech/magnetic_loop_antenna.pdf
« Last Edit: September 04, 2012, 01:11:16 PM by W9GB » Logged
STAYVERTICAL
Member

Posts: 854




Ignore
« Reply #3 on: September 04, 2012, 05:34:34 PM »

Hello, great subject - I am currently in the process of figuring out which is the ideal capacitor as well.

From some research I have done, it seems that dissipation factor is very important when using dielectrics other than air.
Mainly this comes into play when using various materials for spacers which may contribute to the capacitance.
The maximum Q of a capacitor is determined by various losses, but it is interesting that the maximum Q is limited by the disspation factor of the dielectric.
The relationship is Max Q = 1 / DF (dissipation factor).

So if you are using PVC for example for dielectric it has a DF of around 0.016.
So if you used PVC between plates the maximum Q would be 62.

Alternatively, the Polyethylene in RG213 (HDPE) has a dissipation factor of 0.0002, giving a maximum Q of 5000.
Strangely Polystyrene has a DF of 0.00007 giving a maximum Q of 14000 !

The point of the above is that you have to be careful if you are making a really low loss capacitor that you don't
undo the good work by using lossy spacing material for holding the plates apart and so on.

I am also considering making a multigang trombone capacitor to lower stray inductance, and using the PE/inner core of RG213.
It will not be as good as air, but one nut, not fully tightened would lose more I think.

For your reference as well, there is a commercial magloop sold in Europe which uses a piston to squeeze the loop back and forth.
It has a couple of capacitor blocks attached to each side and it tunes from 40m to 10m.
The reason for the wide range is that at the extreme open situation, there is only a small area exposed and so you have a very low open capacitance.

This is the great thing about magloops, there is so much opportunity for experimentation.

73 - Rob
Logged
VK2FAK
Member

Posts: 87




Ignore
« Reply #4 on: September 05, 2012, 02:57:57 AM »

Hi all..

I used a good calculator when trying to find what is needed for a Cap...

http://www.standpipe.com/w2bri/software.htm

Try the Capacitor calculator on that site, its only for air variable though...but lets you play with the size and number of plates and also the spacing...

John
Logged
JAHAM2BE
Member

Posts: 277


WWW

Ignore
« Reply #5 on: September 05, 2012, 03:40:30 AM »

You probably don't really get true high voltage caps unless you have some way of microfinishing all the cut sharp edges so nothing rough is left. I don't think it's just a matter of the making the flat surfaces smooth

Noted. I read before (and hinted at in my original post) that air variable capacitor plates are often tumbled for days (weeks?) on end with walnut shells to polish them and remove surface burrs and the like. I don't know what kind of equipment is needed for this, but I imagine it isn't cheap or easily available.

Does anyone have suggestions for achieving the necessary micropolishing with simple DIY equipment?

Anyway I'll be running 5W for now so arcing probably won't be a problem, but one day I would like to run more power.

10 to 200 pF sounds like a wider range than what you could actually achieve. Those strays add up and likely you'd end up with 30 pF of stray capacitance on a 200pF max unit.

I'm thinking the capacitor geometry probably affects the amount of stray capacitance and thus the minimum capacitance. For example consider two extreme cases: the first case, having a single huge rotor plate and two huge stator plates; the second case, having a stack of 20 small rotor plates and two 20-plate stacks of small stator plates. Even if the total maximum capacitance is the same between the two cases, intuitively it would seem that the stray capacitance of the huge, few-plate capacitor would be less than that of the tall, multi-plate structure. That would seem to suggest fewer, larger plates to reduce minimum capacitance, but then the capacitor becomes short and stout, which I understand may be lossy. Thoughts on the above reasoning?
 
I think MFJ sells some new butterfly caps. 282-2016 and 282-2017. Seem too large for a 5W tuner but a homebrew unit would probably end up with really wide spacing too unless you were really good at micromachinig all the plates and parts.

I'm tempted now and again to buy those units, but I live in Japan and I think shipping here from MFJ costs about as much as the capacitor itself. And anyway, buying a prefabricated capacitor doesn't give me as much flexibility in capacitance range as a homebrew unit would.

Now, an important goal is high Q, so if a homebrew cap cannot possibly approach the Q of the MFJ unit, that would of course be an argument in favor of the MFJ unit. I hope that with balanced construction and soldered plates that a homebrew air-spaced unit can achieve Q equal to or greater than 10,000. But that's just my hope - if anyone has any data or experience, please share!
« Last Edit: September 05, 2012, 03:56:59 AM by JAHAM2BE » Logged

VK2FAK
Member

Posts: 87




Ignore
« Reply #6 on: September 05, 2012, 03:43:06 PM »

HI all..

I think your going a bit overboard on this polishing thing.....clean is good.

The thickness of the copper or whatever your using for plates, as to what you can do to them.....I have fairly thick plates so I just use a fine sand paper to clean then....
Is the loop going to be inside or out.......if out, how are you expecting to keep them clean after the tumble cleaning....your setup is not going to be air tight..

John
Logged
JAHAM2BE
Member

Posts: 277


WWW

Ignore
« Reply #7 on: September 07, 2012, 11:37:36 PM »

The thickness of the copper or whatever your using for plates, as to what you can do to them.....I have fairly thick plates so I just use a fine sand paper to clean then....
Is the loop going to be inside or out.......if out, how are you expecting to keep them clean after the tumble cleaning....your setup is not going to be air tight..

I'm no expert on this, but I understand that for high-voltage caps operated near their maximum voltage rating, small scratches or burrs can seriously degrade the voltage rating and cause arcing. So the polishing/tumbling isn't so much to clean the plates as it is to smooth out the tiny sharp edges to prevent arcing. I think outdoors operation with dust and other surface dirt won't matter as long as said dirt doesn't scratch the plates.

Also, I found out that the equipment used for this sort of thing is apparently called a "rotary tumbler". A search at Amazon yields several small units for less than $100.

The problem that's engaging me most right now is how to solder the stator plates to the mounting post. The typical construction method of using a threaded rod and nuts to hold the plates in place won't work well because the rods I can get are usually made of stainless steel (apparently a major no-no to use in a high-current, low-loss connection) or some other low-conductivity metal. So I'm now thinking to buy some small-diameter (1cm?) copper pipe and use copper flashing for the plates, and to solder the plates to the pipe. But figuring out how to equally and neatly space the plates and hold them in place for soldering is a bit challenging mechanically. I can't solder the plates one at a time because then the heat from soldering later plates would probably undo the joints from the earlier plates. Challenging, but not insurmountable.
Logged

STAYVERTICAL
Member

Posts: 854




Ignore
« Reply #8 on: September 08, 2012, 12:57:01 AM »

I am no expert on these things, but if you stack the copper plates between cardboard sheets or thin strips of wood, or even aluminium, you could drill them as a group.
Then you could put the copper rod through the hole and solder them as a group as well.
I would probably silver solder the copper plates around where the hole will be made, so that the rod will take quickly and well to the plates.
It may even be an idea to "tin" the rod first with a thin layer of silver solder.
Then when you put the rod through the stack and apply heat you will only need to put on the final silver solder to keep it mechanically solid.

Obviously you would move the stack of plates so that the cardboard (or whatever) was not near the hole when doing the final soldering.
I guess aluminium separators would not affect the plates, but heated cardboard may discolour them a bit.

I would not go crazy with the polishing thing, since if you separate the plates enough you won't have any flashover.
I understand you want the most capacitance, so closer spacing is better, but you can go a bit too far with the polishing.
I would just try to round the edges with a very fine sandpaper so they feel smooth and you should be ok.

I don't know if it's any use to you, but I am currently making a loop and building it like a prototyping test bed.
I am using kitchen chopping boards made of Polyethylene (HDPE) which has a dissipation factor of 0.0002, similar to Teflon.
The dielectric constant is also very close to Teflon as well, but they are much less expensive.

At the top of the loop I am making a platform where the loop terminates.
Then, I can simply bolt on my various capacitor/drive systems and evaluate how each works.
This means if something does not work it is easy to junk that idea and try something else.

My current project pipeline contains capacitors made of:

- Double sided PCB material.
- The inside HDPE insulated wire from RG213 coax sliding inside copper tubes (trombone type).
- Anything else I can think of.

Remember, I am not an engineer, just an enthusiastic experimenter, so I need to learn the theory before I make informed decisions.

If you want a loop to cover 40m to 10m, in my opinion, you will need fewer plates and have them parallel to each other (in the same plane).
This will provide the lowest minimum capacitance, since the interacting surfaces are limited to edge-on area.
This is why many butterfly capacitors are problematic.
Even when the moving plates are not meshed, the stationary plates have a good amount of surface area exposed to each other.
This limits the minimum capacitance.

If you look at vaccuum capacitors, it appears they have concentric nested cylinders which mesh.
The minimum capacitance is obtained when the two series of cylinders are edge on, giving low values of around 5pF.
So if we want similar low minimum capacitance, we should consider a similar system, but perhaps in a planar arrangement.
Also, a vaccuum capacitor seems to use a flexible metal bellows to connect one set of plates to the binding posts.
So, the mantra of always using butterfly capacitors is really only if you don't have a good connection system for the moving plates.
I am going to try a similar system made of copper sheet to connect the moving plates to its connection post.
If the connection is excellent it does not matter if it moves, otherwise vaccuum variables would be poor choices as well.

There is a commercial loop sold out of Europe which uses this system.
They use two capacitor blocks which are simply bolted onto the open ends of the loop.
The whole loop is squeezed together slightly (it is hinged at the bottom) to make the capacitors mesh at the top.
This loop covers 40m to 10m because of the loop conductor size, and the capacitive system having a low minimum capacitance.

This system is like my first ever loop where I just soldered two PCBs to the top and squeezed it together by hand (when not transmitting !).
The advantage of this is that you get twice the capacitance of a butterfly type system without the moving contact problem.

I personally will not implement this system since I like to have my loop conductor static, not subject to movement.
This is purely because my mechanical skills are non existent, but I include it to give an indication of what you will need to consider to get a wide range loop.

I have found getting more capacitance is not a big problem, it's the minimum capacitance which is the issue.
When I needed to cover 40m with my loop I just cut six inches of rg213 and clipped it across the capacitor to give 40m.
It worked as well as my other loops since RG213 with solid PE insulation has a low dissipation factor, and is only a "top up" capacitance for 40m when needed.
I would not leave it in place when using other bands though.

One other thing I have found with magloops - keep them as symmetrical as possible.
If you don't you will have SWR matching problems which are difficult to solve.

As soon as I restored symmetry to my loop, my previous SWR problems vanished.
This tip may save you many hours of head scratching.

Good experimenting,

73 - Rob
« Last Edit: September 08, 2012, 01:24:05 AM by STAYVERTICAL » Logged
STAYVERTICAL
Member

Posts: 854




Ignore
« Reply #9 on: September 08, 2012, 02:12:19 AM »

Just a correction above.

When I said butterfly capacitor, I meant split stator.
Butterfly capacitor configurations are designed to minimise the stray capacitance when open, at the cost of total capacitance.
Apologies for any confusion caused.

73 - Rob
Logged
VK2FAK
Member

Posts: 87




Ignore
« Reply #10 on: September 08, 2012, 03:03:46 AM »

HI all....

Well from experience I say you don't have to go to crazy about the cleaning and smoothing of the plates....just be reasonable with it...with no sharp corners on the plates.

John
Logged
AB1PA
Member

Posts: 3




Ignore
« Reply #11 on: September 27, 2012, 09:38:33 AM »

Hi,

I also would not go too crazy about the polishing of the plates, clean plates should do it. I also built a magnetic loop with my home-built butterfly capacitor. I am running 100W into the coax, 100 feet of RG6 by the way, so there should be approx. 70W at the antenna input with no problems at all. I am quite happy with the results. My capacitor is not square, it's slightly rectangular. Compared with my fan dipole at 25 feet the magnetic loop antenna performs similar, if not better. It's only installed on a tripod at approx 5 to 6 feet above the ground.
How i made my antenna: http://www.qrz.com/db/AB1PA

73
Jens
Logged
JAHAM2BE
Member

Posts: 277


WWW

Ignore
« Reply #12 on: October 02, 2012, 06:12:06 AM »

I also built a magnetic loop with my home-built butterfly capacitor. [...] How i made my antenna: http://www.qrz.com/db/AB1PA

I notice you use copper for your plates. That's something I'm also seriously considering, because copper would allow soldering the stator plates to the stator posts (which in my case will also be copper) for lower loss than a mechanical connection with tightened nuts.

However on another forum someone warned me that copper plates will corrode more quickly than, say, aluminum plates. Since a magnetic loop is often used outdoors, and since there is no way to clean corroded plates once they're soldered into place, I'm undecided whether I should go for copper plates or aluminum plates. Aluminum plates might be more weather-resistant, but are more difficult to solder.

Do you have any problems with corrosion of the copper plates?

Also I notice your capacitor only tunes your loop down to 20m. I was hoping to cover down to 40m (and up to 10m), which, based on the physical size of your capacitor, will require a pretty big capacitor - maybe impractically big.
Logged

AB1PA
Member

Posts: 3




Ignore
« Reply #13 on: October 04, 2012, 12:45:35 PM »

The copper looks slightly darker so far, but as far as I can say it didn't have an effect on performance yet. This surely can change in the future, especially since I am planning to weatherize the antenna so I can install it outside permanently. Any tips regarding weatherizing copper without affecting it's RF properties? Anyone?
Primarily I used the copper because I got it for free.

You are right, soldering the stator plates should result in lower loss and keeping the losses small is the key to success with a magnetic loop, right? Well I thought about soldering the plates, and maybe I will do it later. At the moment I will keep it like it is, so it can easily be disassembled. My experience with this antenna tells me that the loss can't be that large since it shows a pretty narrow bandwidth.

Maybe i will even put some more plates on the capacitor to cover 40m, but at the moment I just don't have the time for doing that, I rather spend my little ham radio time slot for having some QSO's. Right now the antenna is installed on a 2x4 stud, which is sturdy enough to hold it. It should even not be a big problem with the 40m plates installed.

For the future I am planning to build an antenna completely made from aluminum since we recently got an aluminum welder at work, but first I have to work on my welding skills. HI.

73
Jens

 
Logged
JAHAM2BE
Member

Posts: 277


WWW

Ignore
« Reply #14 on: October 04, 2012, 03:57:00 PM »

Well I thought about soldering the plates, and maybe I will do it later. At the moment I will keep it like it is, so it can easily be disassembled. My experience with this antenna tells me that the loss can't be that large since it shows a pretty narrow bandwidth.

If you do this, I would be very interested in the before-soldering bandwidth and the after-soldering bandwidth as an indirect indicator of how much improvement soldering the plates brings in your case.

We must be careful of course when using bandwidth as a measure of efficiency, but in your particular case it seems there is little unnecessary series inductance (though your use or paralleled tubes does complicate the analysis slightly), so perhaps bandwidth, and the change in bandwidth introduced by soldering, may give a somewhat realistic indicaton of efficiency and changes thereof.

My planned loop will use huge 10cm copper tubing (see recent eham thread for details), with 1m diameter. If the conductor is very low loss as expected, then this will make capacitor loss even more important. Some calculations indicate that on 40m, the most inefficient and loss-critical band, even fractions of a milliohm can pull down efficiency by a few percent (since the maximum 40m efficiency lies around 30-40 percent as I recall).
Logged

Pages: [1] 2 Next   Go Up
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.11 | SMF © 2006-2009, Simple Machines LLC Valid XHTML 1.0! Valid CSS!