PDA

View Full Version : winding a balun: which would you use?



WA4AVD
01-02-2010, 03:13 PM
Which would you use as the winding core: a toroid or a piece of pvc?

Also, there are rubber coated baseball bat "donuts" with metal core that players use as weights. I am wondering if you could use on of those to wind a balun?

G3TXQ
01-02-2010, 03:58 PM
What will you be using it for?

Steve G3TXQ

AG3Y
01-02-2010, 04:00 PM
I doubt seriously that the metal core baseball bat weight would work very well above a very low frequency. To go higher, you need ferrite material. There are just too many losses in a piece of iron to handle RF !

A very common 1:1 "balun" can be made by winding several turns of thinner coax ( RG 8x ) on a fairly large piece of PVC tube. Look at manufacturers web pages to see the minimum diameter that your coax can be bent around. You might be able to make a tighter loop then the specs, but you do risk the chance of the inner conductor "migrating" away from the center and shorting to the inside of the shield. You don't want that to happen !

73, and good luck! Jim

KD0CAC
01-02-2010, 04:10 PM
I can not tell you a lot about the subject , but will try , just to watch & learn for my self .
Needing to know what use you have for the balun , to start with .
Generally , what I think of from your question , is for an antenna , so what frequencies , how much power .
Are you talking at the antenna to match feedline , or at your shack ?
From what I have gathered , is the coax raped PVC can work & may handle high power , but then I've read that with the high power & raping the coax tighter than the manufacture recommended for a radius , that it can over heat the insulator , and the center can migrate to the shield , making for a short .
As far as the base bat weights , you would need to calibrate its electrical use for this , so starting with a toroid , the has listed specs. may be easier .

KU3X
01-02-2010, 06:38 PM
You never say what frequency range or ratio?
Are you using it for all of HF, low end of HF or high end of HF or maybe VHF?
Do you want a 1:1, 2:1, 4:1 or even higher?

If it's for HF, if it's a 1:1 and you want to cover 10 to 160 meters, you may want to use a #43 ferrite core? Amidon FT 240-43.

All of this stuff is in the ARRL handbook.
You can make a 1:1 balun by slipping ferrite beads over the transmission line like W2DU does in his original balun design.

Bottom line is, if you are making a broad band balun you have to use ferrite.

Yet another question: Are you going to build a current or voltage balun?
You probably want a current balun but it is food for thought.

Barry, KU3X

WB2UAQ
01-02-2010, 06:46 PM
The baseball weights would have very high eddy current losses even at 60 Hz for power transformer use. A power transformer is laminated (sheets) to break up the eddy currents. Ferrite and iron powder cores isolate the magnetic materials from each other reducing the eddy current losses for use at RF. 73, Pete

WA4AVD
01-03-2010, 01:01 AM
Pardon me for not providing more details... I am looking at winding a 4:1 balun for matching ladder line to coax...will be on the exterior wall of the shack. Antenna will be a long dipole...trying to get it to half-wave for 160 if I can. Good points about the bat "donut" weight.

KU3X
01-03-2010, 01:18 AM
The correct way to do it is with a current balun. You could either use a pair of FT 240-K cores, as Jerry Sevick does, or use a pair of FT 240-43 cores as stated in the ARRL Handbook.
To do it correctly, you will need two cores, about 12 feet of teflon tubing and 14 feet of #14 insulated wire.
The way to test it to see if it is really working correctly, borrow an MFJ 259B antenna analyzer and a 200 ohm carbon (non inductive ) resistor. Put the resistor on the output of the balun and the 259B on the input to the balun.
Now sweep the frequency range. If you see the SWR climb above 1.5 to 1, it's not a good balun. It really should stay equal to or below 1.1 to 1 across the bands.
If you want to see a picture of a good one, go to google and search, "GU4-HF-5KW-A" . That will give you an idea of what it should look like.

Good luck with your project,
Barry, KU3X

VK2TIL
01-03-2010, 04:27 AM
We had a good discussion on measuring balanced line through a balun here;

http://forums.qrz.com/showthread.php?t=229892

You should also be careful of the notion that a 1:4 transformation will automatically match co-ax to ladderline.

I think that a "choke" (coiled co-ax or W2DU-type) would be a good start; the W2DU is likely to be more broadband, esp. if a variety of ferrite materials is used.

NA0AA
01-03-2010, 05:02 AM
We had a good discussion on measuring balanced line through a balun here;

http://forums.qrz.com/showthread.php?t=229892

You should also be careful of the notion that a 1:4 transformation will automatically match co-ax to ladderline.

I think that a "choke" (coiled co-ax or W2DU-type) would be a good start; the W2DU is likely to be more broadband, esp. if a variety of ferrite materials is used.

J. Sevick is not complimentary on the W2DU Balun - he points out that it attenuates the common mode current rather than forcing it into the antenna itself.

I would use the Sevick designs for any balun I wanted to use. They are certainly not difficult to fabricate, although you might need to order the toriod cores

VK2TIL
01-03-2010, 05:43 AM
Quite true; there is a difference between a true balun that "forces" the bal-un action and a "choke" or "sortabalun" that just "permits" the action.

But the W2DU choke is simple and would be a good start.

The OP's proposal is for a ... 4:1 balun for matching ladder line to coax ....

This, as I have gently reminded him, is not as simple as might be thought; from the description of his antenna it seems likely that there will be some impedance transformations going-on in the transmission lines and that the "balun" will be asked to handle reactive impedances ranging in value from high to low.

As we saw in the earlier thread to which I linked, baluns of the 4:1 kind do not behave well when out of their "comfort zone" (usually 200 ohms resistive: 50 ohms resistive). I have not read all of Sevick's very fine works but, in the ones that I have, he concentrates on matching pure resistances.

In the OP's circumstances I think that a "choke" would be preferable to a "balun".

KU3X
01-03-2010, 01:39 PM
If you are going to use the 160 meter antenna on only 160 meters and feed it with parallel feeders here is what you can do.
Cut your parallel feeders in 1/2 wave length multiples. What you see at the feed point of the antenna you will see at the other end of the transmission line. Example: 468/1.850*VF. Now put a 1:1 current balun at the shack end. With this method it does not matter if you use 75, 300, 450 or 600 ohms transmission line as long as you keep it multiples of 1/2 wave length.
If you plan on using this antenna on other bands, this method will not work.
To use the 160 meter antenna on other bands you will need a 4:1 balun at the shack end and cut the transmission line in odd multiples of 1/8 wave length.
For multiband operation, it would be a good idea to put a transmission line isolator between your tuner and your rig.
Barry, KU3X

KA5S
01-03-2010, 03:16 PM
As an experiment, Sevick used a 4 inch core number 2 powdered iron core (permeability 10) (T-400A-10) to build a low loss 1:4 voltage Balun intended for use on a tuner output with a balanced load (page 60), though he suggests that where imbalance is present, a 2-core current Balun be used. His investigations on Baluns for tuner outputs led him to recommend that antenna and feedline length be adjusted to provide impdnaces a balun can handle. With hardened Balun, a G5RV (or other reasonably long dipole) would do if the leg length plus feedline length were 85 feet, assuming open wire feed. For best performance on the three lowest bands, he suggests a "double sized" G5RV with leg length of 80 feet and a feedline of 100 feet of 450 ohm open-wire (90 feed poly window line or 82 feet of TV 300 Ohm twin lead).

Amidon sells assembled W2FMNI Balun or kits that have what's needed to wind them. The AB240-250 Kit K Mix Balun Kit (https://www.amidoncorp.com/items/60) is listed at $25, and includes wire, Teflon(tm) tube, a type K core and a transmission line transformer handbook. (Mix K is listed at initial permeability of 280.)

I'll add here I had a 1:1 Balun of this type (not from this manufacturer) heat up probably to 200F (by the time I got the cover off it was merely uncomfortable to touch, rather than boiling spit) with just 100 watts applied; Sevick writes 1:1 is not suited to high impedances and I suspect the high permeability aggravated things. Sevick's tuner Baluns are 4:1 designs, either a 2 core current Balun or a voltage Balun as above, and wound on much lower permeability cores.

Get the book (http://openlibrary.org/b/OL8444583M/Understanding_Building_and_Using_Baluns_and_Ununs) . It really helps.

Cortland
KA5S

G3TXQ
01-03-2010, 05:53 PM
J. Sevick is not complimentary on the W2DU Balun - he points out that it attenuates the common mode current rather than forcing it into the antenna itself.

Could someone please explain in engineering terms what that means? I don't have Sevick's book, I've never come across the comment before, and I don't understand what it means.

Thanks,

Steve G3TXQ

W5DXP
01-03-2010, 07:04 PM
Could someone please explain in engineering terms what that means? I don't have Sevick's book, I've never come across the comment before, and I don't understand what it means.

From page 9 of Building and Using Baluns and Ununs by Sevick, under section "M. W. Maxwell, W2DU - 1983":

"From very accurate insertion loss measurements, we learned that the losses were mainly in the magnetic medium - and that they were related to the voltage levels and the permeabilities. Maxwell didn't take into account these latter findings. ... I seriously question (Maxwell's) power ratings. Ironically, it is very likely that Maxwell's balun had more real loss than the so-called 'transformer-type' balun."

I don't think Sevick realized that the common-mode waves obey the rules of distributed network reflection mechanics. The common-mode current doesn't simply flow through the choke. The common-mode wave, incident upon the choke impedance, encounters an impedance discontinuity with its associated reflection coefficient. If the choke impedance is very high, most of the incident common-mode wave is reflected. If the reflection coefficient is a very large value, most of the common-mode incident wave reflects whether the choking impedance is resistive or inductive. The reflected common-mode wave doesn't contribute to dissipation in the choke. Only the part of the common-mode wave that makes it into/through the choke affects dissipation and that current is easily measured.

Seems to me, for the same number of turns, a toroid with a perm=2000 would reflect a lot more of the common-mode wave than a toroid with a perm=125 - whether the choking impedance is resistive or inductive.

KA5S
01-03-2010, 07:12 PM
Could someone please explain in engineering terms what that means? I don't have Sevick's book, I've never come across the comment before, and I don't understand what it means.

Thanks,

Steve G3TXQ

If dissipated as heat, common mode current represent power lost. If the winding has enough reactance to choke that current instead of attenuating it, and if core loss is small otherwise, the Balun is efficient. Sevick points out that a beaded Balun can't increase reactance by winding multiple turns on a core, so has to use high permeability beads, which are lossy. He cites VE2CV's June 1991 QST article, Transforming the Balun, which notes the beads get extremely hot at high power, and that this will limit use to 125 watts or less. Members Only Archived Article (http://p1k.arrl.org/cgi-bin/topdf.cgi?id=86896&pub=qst)

Something I gather from watching controversies over things like this* for a while is that they seem often to have arisen due to incorrect assumptions. Sevick had to point out that Belrose did not take account of the fact that he had compared the 50 Ohm sleeve Balun in that article with a wound Balun of 100 Ohm impedance (due to the construction of its windings) -- and that this would negatively affect SWR performance .

*We don't NEED to get into the arguments over how much current flows into and out of a loading coil here!

Mind, if the OBJECTIVE is simply to get rid of RF due to CM currents, attenuating works fine. But that doesn't create balance.

FWIW: A quick search using the terms w2du, balun, loss, brings up (surprise surprise) one of W8JI's informative pages (http://www.w8ji.com/Baluns/balun_test.htm).


Cortland
KA5S

W5DXP
01-03-2010, 07:40 PM
Members Only Archived Article (http://p1k.arrl.org/cgi-bin/topdf.cgi?id=86896&pub=qst)

Here's an interesting quote from that article: "If, however, you're installing a new antenna system or reworking an old one, you cannot go wrong if you use a ferrite-bead-choke current balun where a balun is required. If the feeder of which the balun is part operates at a high SWR, you must use a current balun - and a ferrite-bead-choke current balun is best. ... Whenever your MF/HF application calls for a current balun, a ferrite-bead-choke current balun - a type capable of 4:1 and 9:1 impedance transformation as well as the 1:1 ratio afforded by Walter Maxwell's original design - will work best."

G3TXQ
01-03-2010, 07:41 PM
Cecil & Cortland,

Thanks for the further detail.

What I couldn't understand was the implication that a resistive impedance was somehow less effective at impeding common-mode currents than a reactive impedance.

What I still don't understand is the subtle difference between "choking" a current and "attenuating" it? Don't both terms mean that the common-mode current is reduced?

I do understand that if the residual common-mode current is significant, and if the ferrite is a high-loss mix, there is a potential heating problem. That's a risk you run if you choose a lossy mix but don't have sufficient common-mode impedance.

Steve G3TXQ

G3TXQ
01-03-2010, 07:52 PM
Mind, if the OBJECTIVE is simply to get rid of RF due to CM currents, attenuating works fine. But that doesn't create balance.
Cortland,

I have a simple model in my mind of a dipole being centre-fed with coax. We place a common-mode choke at the feedpoint to minimise current flow back along the outer surface of the coax. If the choke impedance is sufficiently high that the common-mode current is negligible, surely we have achieved balance whether the choke impedance was resistive or reactive.

Or am I missing something obvious?

Steve G3TXQ

W5DXP
01-03-2010, 07:59 PM
What I still don't understand is the subtle difference between "choking" a current and "attenuating" it? Don't both terms mean that the common-mode current is reduced?

The following values are plucked from thin air. :) Assume the common-mode current through a 1000+j100 choke and a 100+j1000 choke is the same magnitude. The two chokes have the same absolute value of choking impedance. Which one will dissipate the most power? I think that's the point that is trying to be made. Whatever dissipation exists in the choke subtracts from system balance.

IMO, the most important point is not being discussed. The primary purpose of an RF common-mode choke is to reflect the incident common-mode wave, not to attenuate it. High resistance can reflect almost as well as reactance. Common-mode power/energy reflected by the choking impedance is not lost/transformed in the choke. The problem with some W2DU chokes is not that they are made from #77 material but that they don't contain enough #77 material to reflect most of the incident common-mode wave.

I'm not talking economics here - just physics.

G3TXQ
01-03-2010, 08:15 PM
IMO, the most important point is not being discussed. The primary purpose of an RF common-mode choke is to reflect the incident common-mode wave, not to attenuate it. High resistance can reflect almost as well as reactance. Common-mode power/energy reflected by the choking impedance is not lost/transformed in the choke. The problem with some W2DU chokes is not that they are made from #77 material but that they don't contain enough #77 material to reflect most of the incident common-mode wave.

Cecil,

That's exactly as I understand it.

Not wanting to go off-topic, but I just read the Belrose QST article and was confused to read several things that seem at odds with views I've heard expressed here and on the eHam forum by "balun gurus". For example the statement that the windings on a toroidal 1:1 current balun can saturate the core, whereas a bead balun is immune?

Steve G3TXQ

VK2TIL
01-03-2010, 08:18 PM
I have Transmission Line Transformers; I compared it with Building etc and decided that there was considerable overlap but I thought that TLT has more in it.

I also have several of Sevickĺs published papers.

We don't know enough about the OP's proposal; even he seems uncertain about whether he can get a half-wavelength on 160.

If it turns-out to be a half-wavelength and is only used on 160, the ladderline/balun-choke are superfluous.

If it's short and/or is to be used on other bands, it's a different game but, without knowing antenna length, feedline length/surge impedance etc, we can't be too specific; we can, however, be reasonably certain that some strange reactive impedances will be encountered at some frequencies.

So we (including yr hmble svt) must generalise; always dangerous.

I distinguished between "true" baluns and "sortabaluns/pseudobaluns" earlier; the W2DU device is not a true balun so I try to use the term "choke" or "choke balun".

It's certainly true that it dissipates energy; consider some random numbers in some crude examples.

Suppose a current A arrives at the co-ax/ladderline interface via the inside of the co-ax shield; suppose it "sees" 50 ohms in the ladderline leg and 50 ohms in the outside of the co-ax shield.

Divider action says that half A will continue to the antenna and half will return via the co-ax outer.

Suppose we then slip ferrite rings over the co-ax to produce a choking impedance of 500 ohms.

The divider action is now in favour of the antenna side and most of the current goes that way; a lesser amount is dissipated in the ferrite (and some sneaks-through) so the ferrite has not forced a perfect bal-un action.

But it's done a pretty good job.

There are , of course, considerations of power dissipation and ferrite composition to be addressed at QRO levels.

The quote from the ARRL article (I can't access it) reflects my view expressed earlier;

... a "choke" (coiled co-ax or W2DU-type) would be a good start. and

... the W2DU choke is simple and would be a good start.

G3TXQ
01-03-2010, 08:37 PM
I'm obviously having a bad evening - a couple of comments here have added to my confusion :)


If it turns-out to be a half-wavelength and is only used on 160, the ladderline/balun-choke are superfluous.

If you mean we could simply feed it with coax, surely we would still need a 1:1 current balun at the antenna feedpoint?


The divider action is now in favour of the antenna side and most of the current goes that way; a lesser amount is dissipated in the ferrite (and some sneaks-through) so the ferrite has not forced a perfect bal-un action.

I don't understand the concept of "and some sneaks through". Assuming the length of the bead balun is neglible compared to the wavelength - pretty likely on 160m - the common-mode current on the coax must be the same either side of the choke. Or did someone re-write Kirchhoff's current law when I wasn't looking ;)

Steve G3TXQ

KA5S
01-03-2010, 08:41 PM
...Seems to me, for the same number of turns, a toroid with a perm=2000 would reflect a lot more of the common-mode wave than a toroid with a perm=125 - whether the choking impedance is resistive or inductive.

If its resistive, it can turn RF into heat, not bad for getting rid of stray RF (that's how we use EMI ferrites) but not good for forcing a balanced condition.

There's a lot of information out there, but here's something by Jack Smith at Clifton Labs, Estimating Q of Ferrite Cores (http://www.cliftonlaboratories.com/estimating_q_of_ferrite_cores.htm).

Ferrite cores are NOT just reactive. At 8 Mhz, the 43 core he discusses above has 300 ohms of reactive impedance and about 125 ohms resistive. If Q = Ás' / Ás'' then the Q is 2.4.

WARNING: SWAG follows.

Fair-Rite (http://www.dextermag.com/uploadedFiles/Literature_Fair_Rite_Power_App_Material.pdf) shows a 77 material toroid of 18X10X6mm with initial permeability of 2000 has a complex permeability at 8 MHz less than 20 ohms reactive and about 450 Ohms resistive. If Q = Ás' / Ás'' then the Q is <.01, and allowing for reflection, assuming for SWAG purposes a 500 Ohm single wire transmission line, about half the common mode current could be absorbed and lost to heat, without about half passed to the next bead. It is a good match to 500 Ohms and in that case almost all the CM current would dissipated as heat. The antenna currents remain unbalanced.

Now consider a #67 core with permeability 40 (http://fair-rite.com/newfair/materials67.htm). This has quite a bit lower permeability, but it has also much lower resistive component. In fact it's off the chart (http://fair-rite.com/catalog_pages/material/67ms.gif) on the low end until 20MHz. With Q = Ás' / Ás'' then the Q is at least 400 and at 8 MHz (using Jack Smith's core and Al value for 61 material) if we wind about 22 bifilar turns of 50 Ohm pair (or coax) we've got our 2 K reactance with at most 5 Ohms loss resistance. The reactance is high enough that a majority of the common mode current gets reflected to the antenna, and balance is improved.

I'd like some HF, 400 Q, 2K beads though.

*(SWAG is of course short for Scientific Wild A** Guess)


Cortland
KA5S

G3TXQ
01-03-2010, 08:50 PM
Cortland,

Now you have added to my headache ;)


...... about half the common mode current could be absorbed and lost to heat, without about half passed to the next bead. It is a good match to 500 Ohms and in that case almost all the CM current would dissipated as heat. The antenna currents remain unbalanced.

On 160m the length of a bead is surely negligible in terms of wavelength. So how can the current one side of a bead be half the current the other side? Where did the missing half go?

Steve G3TXQ

W5DXP
01-03-2010, 10:10 PM
For example the statement that the windings on a toroidal 1:1 current balun can saturate the core, whereas a bead balun is immune?

Let's use small coax and wrap ten turns on an FT-77-1024 bead. For the same amplitude of common-mode current through it vs a single run through the same bead, the one with multiple windings will saturate first.

True, of course, but not the whole truth. One would need lots more power to get the same common-mode current through the multiple-turn bead. Assuming the same common-mode current through differing chokes/baluns can lead one down a primrose path.


If its resistive, it can turn RF into heat, not bad for getting rid of stray RF (that's how we use EMI ferrites) but not good for forcing a balanced condition.

Again, high resistance is almost as good as reactance for causing reflection of the incident common-mode wave. 2000 ohms of choking resistance from a perm=2000 #77 toroid is better than 200+j100 ohms of choking impedance from a perm=125 #61 toroid. If the resistive toroid reduces the common-mode current by almost a magnitude more than the partially reactive toroid, it can actually be less lossy.


... a 77 material toroid of 18X10X6mm with initial permeability of 2000 has a complex permeability at 8 MHz less than 20 ohms reactive and about 450 Ohms resistive.

Nobody in their right mind would use just one bead. Most hams use a dozen or so beads. What happens then?

G3TXQ
01-03-2010, 10:30 PM
For the same amplitude of common-mode current through it vs a single run through the same bead, the one with multiple windings will saturate first.

Cecil,

Agreed! But I question the use of the term "saturate" by Belrose - as you know it has a specific engineering meaning.

I did some calculations recently on the flux density and temperature rise of some typical ferrite mixes and common toroidal core sizes. In all the examples I worked, the temperature rise became excessive at flux densities way short of saturation; in every case the core would have shattered before it ever reached saturation.

Amidon say this about their Ferrite and Iron Powder cores:


Overheating of the core will usually take place long before saturation in most applications above 100 KHz.

Perhaps Belrose was using the term "loosely" ;)

Steve G3TXQ

W5DXP
01-03-2010, 10:43 PM
Perhaps Belrose was using the term "loosely" ;)

I suspect that, in the past, most hams didn't differentiate between saturation and thermal overload and considered any toroid that explodes to have been "saturated" according to the Webster's definition: "to destroy completely". :)

AI3V
01-04-2010, 02:46 PM
Which would you use as the winding core: a toroid or a piece of pvc?

Also, there are rubber coated baseball bat "donuts" with metal core that players use as weights. I am wondering if you could use on of those to wind a balun?

You are confusing 2 different things.

1- The material the core is made of,

2- The physical shape that material takes.


1- The material, can be almost any magnetic material, air, or nothing at all (vacuum). The advantage of magnetic materials is that they have a permeability greater than 1 (air=1), this means that the magnetic field will be concentrated by the magnetic material, and that a given length of wire will have a larger inductance. As noted, some magnetic material are more suited for certain frequencys than other magnetic materials. You must choose a suitable frequency "mix" of either iron powder or ferrite. Basically, as you go higher in frequency the individual magnetic particles need to become smaller to avoid eddy current losses, hence the use of iron in "powder" form at RF as opposed to solid(DC only) or sheets (power line frequency AC).

2- The physical shape, the toroid (donut) shape has the neat characteristic that it is self shielding, when a coil is wound into a toroid shape the magnetic flux is contained inside the toroid. You can place 2 toroidal wound inductors much closer together than 2 solenoid wound ones without any mutual coupling between the coils.

You could use a toroidal PVC form!:D It just wont work very well below VHF. Manufactures make toroids with permeability=1 (so called mix 0), usually out of phenolic for when you want the self shielding property of a toroid without any change in inductance.

You can also make a balun out of two solenoid shaped air core inductors!(old school)

Rege

KA5S
01-05-2010, 01:42 AM
Cortland,

Now you have added to my headache ;)



On 160m the length of a bead is surely negligible in terms of wavelength. So how can the current one side of a bead be half the current the other side? Where did the missing half go?

Steve G3TXQ

Not, of course, but the impedance of the bead there is enough to reduce it.

I've used single beads an inch long to attenuate RF at 50 MHz by 20 dB. Put 'em at quarter wave high current places on the wire where the local CM impedance is 1 or 2 Ohms. Of course, I didn't have a lot of RF to burn up and I didn't care if it was lossy (actually wanted it) because I was NOT trying to push current into an antenna to balance it.

Look at some of the Fair-Rite curves and you see that at some frequencies they are almost all resistance, though often not a lot. That does dissipate RF. If you pick the wrong mix and you have (say) 4 or 5 amps of CM current, you can get a bead fairly warm. This may not be bad if you don't mind the loss. Some folks do.



Cortland
KA5S

G3TXQ
01-05-2010, 04:22 PM
Cortland,

Perhaps I misunderstood your comment, and the similar "aside" from VK2IL:


Fair-Rite shows a 77 material toroid of 18X10X6mm with initial permeability of 2000 has a complex permeability at 8 MHz less than 20 ohms reactive and about 450 Ohms resistive. If Q = Ás' / Ás'' then the Q is <.01, and allowing for reflection, assuming for SWAG purposes a 500 Ohm single wire transmission line, about half the common mode current could be absorbed and lost to heat, without about half passed to the next bead.

It seemed to be saying that the current would somehow be absorbed along the length of a bead balun. That's what I couldn't understand.

The only mechanism I can see for the common-mode current being higher at one end of the balun compared to the other end is its standing-wave pattern, and that's not an absorption effect. Not wishing to get Cecil onto one of his favourite topics, but on 160m I would expect a bead balun to be so small a fraction of a wavelength that it could be considered a lumped (R+jX) component; in which case Kirchoff's Current Law would apply.

73,
Steve G3TXQ

W5DXP
01-05-2010, 07:35 PM
Not wishing to get Cecil onto one of his favourite topics, but on 160m I would expect a bead balun to be so small a fraction of a wavelength that it could be considered a lumped (R+jX) component; in which case Kirchoff's Current Law would apply.

Reflection mechanics are also at work. The impedance discontinuity can cause a step function in the forward current and/or the reflected current. Kirchoff's Current Law requires that the total current (the sum of the forward and reflected currents) be the same on both sides of the lumped impedance discontinuity but not that the forward current component or the reflected current component be equal on both sides of the discontinuity. I believe this could be modeled as an Inv-V dipole in EZNEC with a lumped impedance bead in the center of each leg of the dipole and the ends of the dipole grounded using mininec ground. Different lengths of the dipole could be used to simulate the changing length of a transmission line. Hint: Antenna currents on a wire dipole can be considered to be common-mode current.

I modeled two chokes as lumped loads in the middle of each element of a 75m dipole. I ran the simulation on 28 MHz. As expected, there are large standing waves on the source side of the load and small standing waves on the ground side of the load.

I modeled a 6000+j0 ohm resistive choke. It dissipated 15.6% of the source power. With one amp of source power applied, the current through the choke was 0.19 amps. The net current on each side of the choke was equal.

I modeled a 4243+j4243 reactive choke. It dissipated 11.2% of the source power. With one amp of source power applied, the current through the choke was 0.16 amps. The net current on each side of the choke was equal.

The resistive "choke" dissipated 4.4% more source power than the resistive+reactive choke. That's about a 0.4dB difference in dissipation between the two chokes, 1. 6000 ohms at 0 degrees, 2. 6000 ohms at 45 degrees.

The current waveforms on each side of the above chokes were virtually identical.

W5DXP
01-05-2010, 07:43 PM
I've used single beads an inch long to attenuate RF at 50 MHz by 20 dB.

I think Steve's point is that the current on both sides of the bead is attenuated by 20 dB. Thus there is no current drop through the (lumped-circuit) bead since the currents on both sides of the bead are the same. The current on one side of the bead is NOT 20dB lower than the current on the other side of the bead.

KA5S
01-06-2010, 08:10 AM
Reflection mechanics are also at work. The impedance discontinuity can cause a step function in the forward current and/or the reflected current. Kirchoff's Current Law requires that the total current (the sum of the forward and reflected currents) be the same on both sides of the lumped impedance discontinuity but not that the forward current component or the reflected current component be equal on both sides of the discontinuity. I believe this could be modeled as an Inv-V dipole in EZNEC with a lumped impedance bead in the center of each leg of the dipole and the ends of the dipole grounded using mininec ground. Different lengths of the dipole could be used to simulate the changing length of a transmission line. Hint: Antenna currents on a wire dipole can be considered to be common-mode current.

I modeled two chokes as lumped loads in the middle of each element of a 75m dipole. I ran the simulation on 28 MHz. As expected, there are large standing waves on the source side of the load and small standing waves on the ground side of the load.

I modeled a 6000+j0 ohm resistive choke. It dissipated 15.6% of the source power. With one amp of source power applied, the current through the choke was 0.19 amps. The net current on each side of the choke was equal.

I modeled a 4243+j4243 reactive choke. It dissipated 11.2% of the source power. With one amp of source power applied, the current through the choke was 0.16 amps. The net current on each side of the choke was equal.

The resistive "choke" dissipated 4.4% more source power than the resistive+reactive choke. That's about a 0.4dB difference in dissipation between the two chokes, 1. 6000 ohms at 0 degrees, 2. 6000 ohms at 45 degrees.

The current waveforms on each side of the above chokes were virtually identical.

I've been browsing the Fair-Rite site. Their 31 material snap on core (http://tinyurl.com/Half-Inch-31-core)is from an EMI POV a very useful item. At 5 Mhz, though, its Q is quite low and it offers MUCH less impedance (~70 Ohms) than my SWAG of a 500 ohm single wire transmission line. This is probably the kind of thing people would clip onto a coax (the links are for 13mm ID). We're apt to use surplus 43 cores (http://tinyurl.com/43-core), at slightly higher Q.

And then, if the coax shield does act like said single wire line, it's not matched to the dipole element, and can't be expected to deliver power to it even when it IS reflected from a beaded Balun.

Has anyone actually put a current probe/meter on a coax with EMC cores to measure the CM current when it's run vertically into a slightly imbalanced dipole? Lots of people do for horizontally run wires in a test lab; I've done that myself -- but that's not our situation. There's no denying Belrose noted heat at high power.


Cortland
KA5S

KA5S
01-06-2010, 08:24 AM
I think Steve's point is that the current on both sides of the bead is attenuated by 20 dB. Thus there is no current drop through the (lumped-circuit) bead since the currents on both sides of the bead are the same. The current on one side of the bead is NOT 20dB lower than the current on the other side of the bead.

Yes, of course, (flat forehead moment). However, it's the power we worry about. The reflection, power down the coax shield ("forward" power?) is reduced. How much is reflected (and re-reflected, since 500 Ohms is still much higher than the half-dipole's 36 Ohms) and how much is dissipated? What I am thinking here is (for lossy beaded core Baluns) that if shield current is feeding the imbalanced dipole's half connected to shield as a 500 Ohm line, not much pwoer will be delivered, the dipole will still be unbalanced, and due to re-reflection, much of the CM power will still be dissipated.

Cortland
KA5S

G3TXQ
01-06-2010, 09:26 AM
Yes, of course, (flat forehead moment). However, it's the power we worry about. The reflection, power down the coax shield ("forward" power?) is reduced. How much is reflected (and re-reflected, since 500 Ohms is still much higher than the half-dipole's 36 Ohms) and how much is dissipated? What I am thinking here is (for lossy beaded core Baluns) that if shield current is feeding the imbalanced dipole's half connected to shield as a 500 Ohm line, not much pwoer will be delivered, the dipole will still be unbalanced, and due to re-reflection, much of the CM power will still be dissipated.

Cortland
KA5S
Indeed! But as was said earlier, if you choose to use a "lossy ferrite mix" you need to ensure there is sufficient choking impedance that there is very little common-mode current to cause dissipation.

It seems to me that's one of the trade-offs in choosing the mix. Lossy mixes have certain attractions over less-lossy mixes; but if you happen to encounter a common-mode situation where the balun has insufficient choking impedance, the balun with the lossy mix would be the first to overheat.

Call me cynical, but I believe the widespread inappropriate use of Type-2 Iron Powder material (and even Type-6) for 160m thru 10m baluns is because manufacturers would rather risk poor balance than risk a core fracture - customers generally spot the latter, but may not recognise the former ;)

Steve G3TXQ

G3TXQ
01-06-2010, 09:32 AM
I think Steve's point is that the current on both sides of the bead is attenuated by 20 dB. Thus there is no current drop through the (lumped-circuit) bead since the currents on both sides of the bead are the same. The current on one side of the bead is NOT 20dB lower than the current on the other side of the bead.

Cecil,

Exactly so!

Having said that, I'm left with the nagging question: "Why did Belrose find that the beads nearest the balun's balanced output heated most?"

73,
Steve G3TXQ

KA5S
01-06-2010, 12:48 PM
Indeed! But as was said earlier, if you choose to use a "lossy ferrite mix" you need to ensure there is sufficient choking impedance that there is very little common-mode current to cause dissipation.

It seems to me that's one of the trade-offs in choosing the mix. Lossy mixes have certain attractions over less-lossy mixes; but if you happen to encounter a common-mode situation where the balun has insufficient choking impedance, the balun with the lossy mix would be the first to overheat.

Call me cynical, but I believe the widespread inappropriate use of Type-2 Iron Powder material (and even Type-6) for 160m thru 10m baluns is because manufacturers would rather risk poor balance than risk a core fracture - customers generally spot the latter, but may not recognise the former ;)

Steve G3TXQ

I suspect we have been talking about 2 different aspects of the situation. It is often not enough to merely get rid of CM current. Often, one has to enforce balance.

I'd love to do some experiments; a pair of remote sensors with current to frequency converters and LED optical outputs might be the tool -- read 'em with a telescope and see if the tone from each is the same.


Cortland
KA5S

G3TXQ
01-06-2010, 01:24 PM
I suspect we have been talking about 2 different aspects of the situation. It is often not enough to merely get rid of CM current. Often, one has to enforce balance.

Cortland,

Could you say a bit more about what you see as the difference between the two things - it's obviously too subtle for me :)

If I centre-feed a half-wave dipole with coax, and put sufficient choking impedance at the feedpoint to ensure that the common-mode current on the outer surface of the braid is negligible, doesn't that ensure that the currents into the dipole legs are balanced?

I've always assumed that if there is no common-mode current on the feedline - coax or ladderline - only the differential-mode currents remain; and because they are equal amplitude/opposite phase the antenna currents must be balanced?

Steve G3TXQ

G3TXQ
01-06-2010, 02:02 PM
Cortland,

I just modelled a 20m half-wave dipole fed with coax, with a pretty severe common-mode problem. The coax was approx a half-wavelength long (common-mode) and well grounded at the shack end.

Using the current in one dipole leg as the reference, the current in the other leg was:

No balun: Current = 0.87 @ -23 degrees
Balun = 1000+j0: Current = 0.97 @ -0.1 degrees
Balun = 0+j1000: Current = 1.0 @ 1.94 degrees

I can't see anything there which suggests the "lossy balun" achieves significantly poorer balance.

Edit:

Adjusting the coax length to produce an extreme case:
No balun: Current = 0.3 @ 81 degrees
Balun = 1000+j0: Current = 0.97 @ 0 degrees
Balun = 0+j1000: Current = 1.003 @ 1.6 degrees

So the resistive balun has improved the current balance from -26dB to -0.26dB and the recative balun has improved it to -0.02dB.

73,
Steve G3TXQ

W5DXP
01-06-2010, 02:32 PM
I'm left with the nagging question: "Why did Belrose find that the beads nearest the balun's balanced output heated most?"

I experienced the same thing back around 1990, thought long and hard about it, and am therefore allowed to speculate. :)

I used 12 FT-77-1024 cores which are 3/4 inches long, i.e. 12 of them are 9 inches long. The total resistance, e.g. ~1000 ohms, is distributed over that 9 inches. That means the reflections are distributed over that 9 inches, i.e. each bead causes a separate and distinct reflection. "The bead nearest the balun's balanced output" has the reflected current from 11 cores flowing through it in addition to the transmitted current. The bead nearest the balun's unbalanced input only has the transmitted current flowing through it.

Given the distributed nature of the 12 reflections, it seems only logical that the bead with the most current flowing through it will heat up more than the one with the least current flowing through it after all 12 reflections have gone the other direction. (I'm deliberately ignoring re-reflections for the sake of simplicity.)

One can resolve the confusion by mentally placing a one-wavelength piece of lossless 50 ohm coax between each bead. Conditions on each side of each bead are identical but measure/calculate what happens between the beads. The feedline in the following thought experiment is 50 ohms and lossless with 1WL between beads.

Common-mode Source---1WL---bead---1WL---bead---1WL---bead---etc.---Gnd

Note that the reflection coefficient seen by the decreasing common-mode wave decreases from bead to bead. :eek: The logical fly-in-the-ointment is that each 1WL section radiates. :(

I have simulated a wire with three 5000 ohm chokes installed up and down the wire. I think the current magnitudes and distribution will open some eyes.

http://www.w5dxp.com/commode.JPG


So the resistive balun has improved the current balance from -26dB to -0.26dB and the recative balun has improved it to -0.02dB.

Not to mention that an approximately 1000+j0 ohm choke is easy to implement while an approximately 0+j1000 ohm choke is virtually impossible to implement. Even the best reactive choke material has a considerable resistive component. I have concluded that for the same number of turns on the same size toroid, a resistive choke with a perm=2000 is considerably superior in every way to a resistive/reactive choke with a perm=125.

W5DXP
01-06-2010, 03:20 PM
I suspect we have been talking about 2 different aspects of the situation. It is often not enough to merely get rid of CM current. Often, one has to enforce balance.

Yes, the absence of a balun at a BALanced to UNbalanced junction can be the cause of conducted common-mode current. Installing a good balun in such a system results in balance at least for conducted common-mode. Installing an effective choke at the BALanced to UNbalanced point accomplishes exactly the same function.

If a good choke is installed at the coax to balanced antenna junction, there is simply no other path for the current to take except to flow, balanced into the antenna feedpoint. I'm talking about conducted common-mode current, not inducted common-mode current which is often a different separate problem.

However, if the choke is installed at the shack end of the coax instead of at the antenna, it reduces RF in the shack, but feedline radiation and antenna feedpoint unbalance can still be high.

Take a look at http://www.w5dxp.com/commode.JPG to see what effect chokes have in a distributed network system.

I just simulated the coax outer braid on a dipole using two types of chokes.

1. High-perm resistive material, choking impedance = 5000+j0 ohms. The dissipation in the choke is 0.3 watts.

2. Lower-perm resistive/reactive material, choking impedance = 500+j500 ohms. The dissipation in the choke is 1.5 watts, five times higher than in the high-perm material.

Obviously, the high-perm resistive material does a better job of reducing the magnitude of common-mode current.

Conclusion: The higher permeability (2000) of #77 material causes less dissipation than the lower permeability (125) of #61 material for the same number of turns on the same size ferrite toroid.

G3TXQ
01-06-2010, 05:03 PM
I experienced the same thing back around 1990, thought long and hard about it, and am therefore allowed to speculate. :)

Cecil,

That current distribution is certainly interesting!

You prompted me to split the choke in my 20m dipole CM model into two: I put one R=250 ohm choke right at the feedpoint, and another one about 8" lower on the coax feedline. As you would expect from the CM current distribution on 20m, across an 8" length the difference in currents was tiny: 0.06076A versus 0.06066A.

So I understand and accept your theory, but I'm still puzzled why the current variation in a practical situation would be sufficient to cause a noticeable difference in the heating. I find it difficult to envisage any CM current distribution, even on 10m, where the current would be significantly different across the length of one - or even a few - beads.

73,
Steve G3TXQ

W5DXP
01-06-2010, 06:26 PM
I'm still puzzled why the current variation in a practical situation would be sufficient to cause a noticeable difference in the heating.

We know that open-ended dipoles, open-ended stubs, and loading coils all have end effects. I suspect the warmer first bead is an end effect associated with w2du bead baluns where the forward current flux generating effect extends into the inside of the first bead.

I am guessing that the flux in the first bead is somewhat proportional to the forward current while the flux in the subsequent beads is mostly proportional to the net current. Perhaps the forward current encroaches upon the first bead more than any other?

I'll bet if we created pick-up loops to detect the flux in each bead, that the first bead would indicate more flux than any of the others even though such a measurement might contradict the measured current through the beads.

G3TXQ
01-07-2010, 05:44 PM
Cecil,

I tried a short experiment this afternoon. I built a bead balun using a couple of inches of RG174 with ten #43 mix toroids slipped over it. The end toroid had a 10 turn secondary connected to a sensitive wideband mWatt meter. I put a 50 ohm load at one end of the balun and connected the coax centre conductor to Ground to unbalance it. Then used my Palstar ZM-30 as a signal source to drive the other end. I noted the toroid secondary current at 2MHz, 7MHz and 28MHz. Then I reversed the source/load and repeated the measurement.

The secondary current was obviously lower at 7MHz than 2MHz, and lower again at 28MHz, but I didn't see any change in the currents when I swopped the balun around. I reckon my measurement accuracy would have been sufficient to spot a >10% difference in current.

73,
Steve G3TXQ

W5DXP
01-07-2010, 08:35 PM
I didn't see any change in the currents when I swopped the balun around.

Did you put enough power through it to heat up the first bead? I suspect that first bead goes non-linear at a certain point.

G3TXQ
01-07-2010, 08:52 PM
Did you put enough power through it to heat up the first bead? I suspect that first bead goes non-linear at a certain point.

Cecil,

No, I didn't - I was only using my analyser as a signal source. If I'm brave, tomorrow I'll try it again with some real power.

But I'm not sure what you mean be a bead going "non-linear". The data sheets show that these toroids overheat at flux densities way below the level at which they saurate; in many cases they would shatter before they saturated.

73,
Steve G3TXQ

W5DXP
01-07-2010, 11:56 PM
But I'm not sure what you mean be a bead going "non-linear".

The hot bead seems to be violating the laws of linear math and physics. Therefore, it must be responding in a non-linear manner.

KA5S
01-08-2010, 03:01 AM
Cortland,

Could you say a bit more about what you see as the difference between the two things - it's obviously too subtle for me :)

If I centre-feed a half-wave dipole with coax, and put sufficient choking impedance at the feedpoint to ensure that the common-mode current on the outer surface of the braid is negligible, doesn't that ensure that the currents into the dipole legs are balanced?

I've always assumed that if there is no common-mode current on the feedline - coax or ladderline - only the differential-mode currents remain; and because they are equal amplitude/opposite phase the antenna currents must be balanced?

Steve G3TXQ

I think that in the case where an antenna itself is not balanced, dissipating CM RF will reduce our Balun's ability to make it balanced; whatever is radiated away as heat is no longer available for that purpose. If the choking impedance is non-dissipating (reactive) none of that power will be lost and the antenna will be restored to balance by reflection and re-reflection of power due to CM current on the shield.

Sevick talks about dielectric core losses due to the difference in voltage beween Balun inputs and outputs. But I may not have understood him!



Cortland
KA5S

KA5S
01-08-2010, 03:09 AM
We know that open-ended dipoles, open-ended stubs, and loading coils all have end effects. I suspect the warmer first bead is an end effect associated with w2du bead baluns where the forward current flux generating effect extends into the inside of the first bead.

I am guessing that the flux in the first bead is somewhat proportional to the forward current while the flux in the subsequent beads is mostly proportional to the net current. Perhaps the forward current encroaches upon the first bead more than any other?

I'll bet if we created pick-up loops to detect the flux in each bead, that the first bead would indicate more flux than any of the others even though such a measurement might contradict the measured current through the beads.

A lot of the yelling about mobile loading coil current was because it doesn't take much capacitance to affect that circuit. Might there be a similar effect here? If we really have 1K-Ohm choking impedance (reactive or resistive) we need something approaching 10K-Ohm isolation to the measuring sensor to keep things clean.

Cortland
KA5S

G3TXQ
01-08-2010, 11:23 AM
I think that in the case where an antenna itself is not balanced, dissipating CM RF will reduce our Balun's ability to make it balanced; whatever is radiated away as heat is no longer available for that purpose. If the choking impedance is non-dissipating (reactive) none of that power will be lost and the antenna will be restored to balance by reflection and re-reflection of power due to CM current on the shield.

Sevick talks about dielectric core losses due to the difference in voltage beween Balun inputs and outputs. But I may not have understood him!
Cortland
KA5S

I see it differently. If significant common-mode current is flowing on the outside surface of the coax feedline, power will be "lost" even if the choke is purely reactive. "Lost" in the sense that there will be radiation from the feedline rather than from the dipole.

However, even if power is not "lost", a purely reactive choke doesn't guarantee balance, if significant common-mode current is flowing. It's very easy to see the effect in an EZNEC model. A Resistive choke tends to cause a difference in magnitude of the dipole leg currents; whereas a Reactive choke of the same impedance tends to produce same-amplitude leg currents but with a phase difference. The assymetry in the azimuth pattern which this produces is clearly evident in the EZNEC plots, and is more severe with the Reactive choke - albeit the total power radiated is slightly higher because none is lost in the choke.

Here are the comparitive azimuth plots for a 20m half-wave dipole with a severe common-mode problem; the blue trace is a mediocre 100 ohms Resistive choke and the Red trace is a mediocre 100 ohms Reactive choke. The total power radiated with the Reactive choke is slightly higher, because no power is lost in the choke; however the imbalance it causes is plain to see:

http://www.karinya.net/g3txq/temp/chokes.png

With the Resistive choke there was a 7 degree phase difference between the leg currents, with the Reactive choke the phase difference was 75 degrees!

73,
Steve G3TXQ

W5DXP
01-08-2010, 01:14 PM
A lot of the yelling about mobile loading coil current was because it doesn't take much capacitance to affect that circuit. Might there be a similar effect here?

That's an interesting idea. For a traveling wave through a w2du choke/balun, what is the amplitude and phase difference between the flux in the first bead vs the flux in the last bead?


With the Resistive choke there was a 7 degree phase difference between the leg currents, with the Reactive choke the phase difference was 75 degrees!

And in case someone doesn't know, the greater the differential phase difference in a transmission line, the greater the common-mode amplitude. In an antenna, that phase difference equates to radiation cancellation in the far field.

G3TXQ
01-08-2010, 02:12 PM
The hot bead seems to be violating the laws of linear math and physics. Therefore, it must be responding in a non-linear manner.

Cecil,

Either that, or they are still linear but there's some factor we haven't yet considered.

I repeated the experiment just now with a bit more power - 10Watts. I used a probe thermometer to measure bead temperatures. The beads settled at about 25 deg C above ambient - all beads were the same temperature within the resolution of my thermometer - 1 degree.

73,
Steve G3TXQ

KA5S
01-08-2010, 06:36 PM
I see it differently. If significant common-mode current is flowing on the outside surface of the coax feedline, power will be "lost" even if the choke is purely reactive. "Lost" in the sense that there will be radiation from the feedline rather than from the dipole.

However, even if power is not "lost", a purely reactive choke doesn't guarantee balance, if significant common-mode current is flowing. It's very easy to see the effect in an EZNEC model. A Resistive choke tends to cause a difference in magnitude of the dipole leg currents; whereas a Reactive choke of the same impedance tends to produce same-amplitude leg currents but with a phase difference. The assymetry in the azimuth pattern which this produces is clearly evident in the EZNEC plots, and is more severe with the Reactive choke - albeit the total power radiated is slightly higher because none is lost in the choke.

Here are the comparitive azimuth plots for a 20m half-wave dipole with a severe common-mode problem; the blue trace is a mediocre 100 ohms Resistive choke and the Red trace is a mediocre 100 ohms Reactive choke. The total power radiated with the Reactive choke is slightly higher, because no power is lost in the choke; however the imbalance it causes is plain to see:

http://www.karinya.net/g3txq/temp/chokes.png

With the Resistive choke there was a 7 degree phase difference between the leg currents, with the Reactive choke the phase difference was 75 degrees!

73,
Steve G3TXQ

Iff CM power is a substantial pert of the applied power, and if it's dissipated as heat, we could have a problem. IMO, it's not often that bad. But in any event, with sufficient near-feed point Balun impedance , coax radiation is not an issue.

A distributed beaded Balun where each low-Q bead by itself has a fairly low (one turn) impedance allows the opportunity for power to be dissipated, versus a higher Q wound Balun of the same (fairly high) impedance where dissipation is small doesn't. But if there's less than 100 watts dissipated out of 1500 or so, outside of the pattern distortion you show in the nulls, and if beads don't smoke or fall apart due to heat, it might be unnoticeable.

How many angels CAN dance on the head of a pin? :)

Cortland
KA5S

G3TXQ
01-08-2010, 07:05 PM
A distributed beaded Balun where each low-Q bead by itself has a fairly low (one turn) impedance allows the opportunity for power to be dissipated, versus a higher Q wound Balun of the same (fairly high) impedance where dissipation is small doesn't. But if there's less than 100 watts dissipated out of 1500 or so, outside of the pattern distortion you show in the nulls, and if beads don't smoke or fall apart due to heat, it might be unnoticeable
Agreed! Perhaps I was taking too many words to make the simple point that a non-dissipative choke doesn't gaurantee antenna balance :)

73,
Steve G3TXQ

W5DXP
01-08-2010, 09:00 PM
I repeated the experiment just now with a bit more power - 10Watts.

Try it with 100 watts. That was the condition when I first detected it.

KA5S
01-08-2010, 09:37 PM
Agreed! Perhaps I was taking too many words to make the simple point that a non-dissipative choke doesn't gaurantee antenna balance :)

73,
Steve G3TXQ


I suspect it could -- if it had much higher impedance than the coax outer shield surface acting as a single-wire transmission line. What does it look like at 2000 or 5000 ohms?

Cortland
KA5S

G3TXQ
01-08-2010, 09:51 PM
I suspect it could -- if it had much higher impedance than the coax outer shield surface acting as a single-wire transmission line. What does it look like at 2000 or 5000 ohms?

Cortland
KA5S

At 2000 ohms the imbalance in the dipole currents is negligible whether it's resistive or reactive, and the azimuth patterns look symetrical.

When the choking impedances get down to about 250 ohms the azimuth assymetry begins to appear, and it is more pronounced with the reactive choke.

Although it's not a factor in the model I'm using, a potential problem with the reactive CM choke is that it can resonate with the CM impedance to enhance the CM current.

73,
Steve G3TXQ

G3TXQ
01-08-2010, 09:54 PM
Try it with 100 watts. That was the condition when I first detected it.

That's tricky - these are tiny T37-43 toroids!

Before searching for an explanation why the end bead should get hotter, are we sure it does? Are there any other references to the phenomena other than Belrose's article?

Steve G3TXQ

W5DXP
01-09-2010, 04:58 PM
Before searching for an explanation why the end bead should get hotter, are we sure it does? Are there any other references to the phenomena other than Belrose's article?

I experienced it while helping a ham in Tucson set up his antenna in the early 90's. I reported the experience on r.r.a.a and was told that the most likely explanation was that the first bead was a different ferrite mixture than the others. But since Belrose reported it also, I tend to believe it actually happened. I have 25 FT-77-1024 beads with which I could run some tests and I have previously measured 40 ohms for one wire using an MFJ-259B.

Incidentally, Amidon says the impedance for one turn is 135 ohms for that bead but it is technically not true. 135 ohms is for one turn with two wires running through the center hole. When the bead is slipped over coax, there is only one wire running through the center hole with no looping turn on the bead. I wonder how many hams think N of those beads slipped over coax multiplied by 135 ohms is the impedance of their choke when it is actually N x 40 ohms?

Later: I put close to 100 watts into my tuner. The output of the tuner was connected to a wire that ran through ten FT-77-1024 beads. The MFJ current meter read 0.5 amps of RF on both ends of the choke. The beads heated up but I could detect no appreciable difference in their individual temperatures.

I don't remember what band we were on in Tucson when I noticed the different temperatures. It was close to the peak of the next-to-last sunspot cycle so we may have been on 10m where the w2du choke-balun may have been an appreciable portion (5%) of a wavelength.

KA5S
01-09-2010, 10:11 PM
...Incidentally, Amidon says the impedance for one turn is 135 ohms for that bead but it is technically not true. 135 ohms is for one turn with two wires running through the center hole. When the bead is slipped over coax, there is only one wire running through the center hole with no looping turn on the bead. I wonder how many hams think N of those beads slipped over coax multiplied by 135 ohms is the impedance of their choke when it is actually N x 40 ohms?...

A lot of EMC beads were called "100" Ohm beads (etc.) -- but that was true only at 100 MHz for the -43 mix. Specs are a lot more informative nowadays than that. Even if folks like me look at the wrong graph!

Cortland
KA5S

KA5S
01-10-2010, 07:07 PM
Thinking about forcing balance. If the antenna is imbalanced, one needs to change the impedance of the feed line connected to each side so an equal current flows in each side. There is a well-known way for doing this with arrays of verticals, a quarter wave feed line forcing equal currents into different impedance elements.

Two quarter wavelength lines, fed 180 degrees out of phase with a 50:100 Ohm Balun, could, if each were connected to one terminal of a less tan perfectly balanced dipole, result in equal currents in its elements.



Cortland
KA5S

ad: Rohn-1