Is the W2DU balun really a balun?

[W2DU]

I received an email from Shane yesterday describing a semantic discussion he’s been having with someone who claims the W2DU balun is not a balun. Over the years I’ve been confronted with this same semantic problem, so I’ve discussed it many times with many other hams. Consequently, I believe the answer I’ve given to Shane might be of interest to AMfoners, so the following is how I responded to Shane’s request for my opinion:

Hi Shane,
 
I understand your semantic problem concerning the balun, because it has been presented to me over and over again. In one instance, almost immediately after my QST article on the W2DU balun was published in March 1983, Lew McCoy took issue with me in a rather heated personal conversation, in which he refused to agree that my balun is a balun. I tried to persuade him that if it performed like a balun it is a balun, but he would not agree. So let me try to explain the issue as I see it.
 
First, there are many different configurations of baluns that perform the procedure required to make a transition from an unbalanced line or circuit to a balanced line or circuit, and vice versa. The coupling device that makes that transition in such a way that no significant undesired or extraneous voltages and currents appear on either the unbalanced or balanced line or circuit is a balun by definition, no matter what the configuation. The word 'significant' is operative here, because no balun configuration is perfect in the real world.
 
The W2DU bead balun satisfies the definition, so yes, the W2DU configuration is a balun.
 
To clarify the issue let's consider the circuit involving the W2DU configuration. The center conductor of an unbalanced feedline (the coax) is connected to one terminal (A) of the antenna, while the outer conductor, or shield, is connected to the opposite terminal (B) of the antenna. The two antenna terminals are considered to be balanced with respect to ground.

However, at the junction of the shield and antenna terminal B, due to skin effect, the current flowing on the inner surface of the shield sees two paths which it can follow; one path is into antenna terminal B and the other path is down the outside surface of the shield. How the current divides between the two paths is determined by the impedance appearing in each path. If the impedance appearing at the outside surface of the shield is fairly low (and with no balun that impedance will normally be low), current will flow on the outer surface of the shield.
 
On the other hand, if an impedance placed in series at the antenna-terminal end of the shield that is much greater than the impedance appearing at antenna terminal B, the current flowing into terminal B will be greater than that flowing onto the outside surface of the shield. If that impedance placed at the antenna-terminal end of the shield is made sufficiently high with respect to that appearing at terminal B, the current flowing onto the shield can be reduced to the value where it is insignificant.
 
In designing the W2DU balun I compromised cost, physical size, and the amount of shield current that could be tolerated that would result in insignificant effect on the radiation pattern of the antenna. I made the decision that if the shield current was at least 20 dB less than that flowing onto terminal B, the effect of the shield current on the radiation pattern would be negligible. I also considered what I believe is the worst case scenario seen by the amateur community: An 80m dipole resonant at 3.750 MHz operated at 3.50 MHz. From extensive measurements I've made on my 80m dipole, 125' horizontal at 40 feet above ground, the impedance at resonance is 64.8 + j0 ohms, while at 3.50 MHz the impedance is 53.1-j122.7 ohms (SWR = 7.54:1 on a 50-ohn line). The impedance appearing at each terminal, A or B, with respect to ground is half the impedance appearing between the two terminals. Thus the impedance appearing at terminal B is 26.66 - j61.35 ohms, for an impedance Z = 66.9 ohms. In the graph of impedance vs frequency pertaining to the W2DU balun, the balun impedance in series with the shield at 3.50 MHz is seen to be approximately 1050 ohms. The ratio of 1050 to 66.9 is 15.7:1, which amounts to 23.9 dB. Thus the shield current flowing in the worst case is 23.9 dB less than the current flowing onto terminal B of the antenna. Consequently, from the standpoint of the effect on the radiation pattern of the antenna, the W2DU balun is performing like a practical balun, should be considered as a balun and therefore should be called a balun.
 
I hope this clarifies the semantic discussion on whether the W2DU balun is really a balun.
 
Walt

[K1JJ]

Hi Walt,

Interesting about the debate.

I agree the W2DU is a balun.  It's simply a "type" of balun. There are many types of baluns out there.

Just like a Ford is a car and a Chevy is also a car. And a Chrysler is a car. There's many types (or even brands) of cars.

Maybe we would call the W2DU a "choke" balun, or "choke bead" balun?   And a 4:1 balun using 1/4 wave length of coax a "coaxial' balun.  And a 1:1 balun that uses a primary and secondary of wire wound on a toroid, a "toroidal transfomer" balun?    And a hank of coax turns at the feedpoint a "coaxial choke" balun?


Anyway, as I said there's many styles of baluns and yours does the same job as the others, so it's a balun as a general class of categorization. It creates an environment that goes from unbalanced to balanced with 20db or greater of isolation. That's a balun to me.

One of my favorite techniques is a combination of these techniques. For a 40M Yagi I'll use about 7 turns of coax wound on a 4" PVC pipe and some slip-on type 43 ferrite beads slipped on the coax at the feedpoint.  That will give somewhere up around 1K or more, I'll bet.  Good for a 50 ohms or less Yagi input.

Take care and don't let 'em wear ya down.... 

Tom, K1JJ

[KD6VXI]

One of my favorite techniques is a combination - for a 40M Yagi I'll use about 7 turns of coax wound on a 4" PVC pipe and some slip-on type 43 ferrite beads slipped on the coax at the feedpoint.  That will give somewhere up around 1K or more, I'll bet.  Good for a 50 ohms or less Yagi input.

Couldn't coil the hardline, but here's the W2DU on the mobile.  Next time I'm able, I'm going to measure the inductance we got.  Didn't have a chance to measure it at the time, BUT it DID cure a lot of RF in the mobile, so much so that the rack gear was unusable.

--Shane

[W2DU]

I finished a rather long response to Tom, and when I pressed 'post' I received the message that while I was typing another msg was coming in, that of Shane, and suggested that I review my msg. I did that, but my msg to Tom disappeared and I can't find it. Can anyone show me how to retrieve it if it's retrievable?

Walt

[W1AEX]

Hi Walt,

Sometimes it can be "recovered" by using the "back" button of your browser, as long as you have not closed the browser. Not sure what can be done beyond that!

Rob

[K1JJ]

Walt,

If you see another message coming in and want to post regardless, hit the post button a second time and it will go through.

If you want to read the new one before posting, you might copy your current message into memory just in case it gets lost. That has happened to me too.  As Rob says, the back button can usually get your message back if you haven't done too much surfing in the meantime.

Finally, there's probably a streamlined way to do it properly, if one of the moderators chime in.


T

[W2DU]

Thanks for the info Tom, I didn't know that hitting the 'post' button twice would ensure its being posted. Next time I will copy before hitting the 'post' button. I've had this problem twice before, and like the story goes, three times and yer out! I'm old, but not too old to learn!

Tomorrow I'll try to remember what I said in the msg that went astray and repost it--that is if I can remember what I said--that is if I can remember what I said--that is if I can remember what I said. Did I say that? I don't remember. Duh!

Walt

[ke7trp]

If you are going to type a long post.  Just type IT in microsoft word.. Then spell check it. THen you can cut and paste it from word to the forum and hit post.  THis way, You will never loose your work.  I enjoy your long posts!

Clark

[Ed-VA3ES]

Walt, you're 90, and can be fiorgiven if you are loosing your short term memory!  I'm 58 and it happens all the time to me!! 

BTW - did I mention that I have  poor short term memory?


 

Whu....

Tomorrow I'll try to remember what I said in the msg that went astray and repost it--that is if I can remember what I said--that is if I can remember what I said--that is if I can remember what I said. Did I say that? I don't remember. Duh!
Walt

[W2DU]

Thanks Ed, for forgiving me my short-term memory. Memory? What's that? Oh yeah, I remember now, I was going to reply to Tom. Guess I still will.

So Tom, I'm pleased that you, along with several hundred other hams, agree that the W2DU balun is a balun. I say hundreds because the Wireman has mfgd and sold that many to satisfied customers. There have been reports of W2DU baluns overheating, but those reports were from hams who had not purchased the baluns from the Wireman, because he has not had any reports of heating problems.

On the other hand, some of the other W2DU baluns are distinctly inferior to those from the Wireman, while his are of excellent manufacture. I know of one mfgr who short changed the buyer by stinting on the number of beads, allowing more shield current, thus overheating the few beads that were there.

At the time I developed the ferrite-bead balun I was certain that it wasn't patentable, but I've since discovered that it could have been. Of course, it's now too late. I if had been patentable I would have had some control over its manufacture.

And yer right, Tom, it is a 'choke' balun, no question. Lew McCoy said it was a choke, but wouldn't admit to its being a balun. Oh well.

Interesting that you used seven turns of coax around a 4" PVC in addition to #43-mix ferrites around the coax, about which I have a few comments. Did you check the PVC to determine whether it contributes to the loss factor? Perhaps you're already aware that some types of PVC are more lossy than others. If so you know that the test is to place a sample of the PVC in the microwave, and if it doesn't get hot it's made of the less lossy-type material. If you wound the coax around a 4" PVC it would seem you used a coax smaller than RG-8. In one way that's good, because the turn-to-turn capacitance would be less. Some people are not carefull to minimize that capacitance, and result in having an inductance near self-resonance, which condition seriously reduces the series impedance of the coil, thus detracting from the effectiveness of the balun.

Extrapolating from the graph of frequency vs impedance that appears in my article from QST, March 1983, the #43 mix ferrite bead of the size that fits over RG-303 coax is approximately 11 ohms/bead. It would be interesting to know how much impedance was obtained from the coiled coax vs the number of beads you used. It should be noted, however, that the balun stress on 40m is much less than on 80m if you use the dipole across the entire band when resonant at the center frequency.

Incidentally, my QST article mentioned above is available for downloading from my web page at www.w2du.com, Chapter 21. In addition, I have made an addition to Chapter 21 that will appear in Reflections 3, which is now being readied for publication at the end of this Summer. The addition is Section 21.11, which I'll try to add to this post. If some of the greek characters don't come through I'll have to repost with the material changed from Word to PDF as an attachment, so here goes:


Sec 21.11 The Labor Pains and Birth of the W2DU Ferrite-Bead Current Balun

    It is interesting to know that the original design of the W2DU ferrite bead balun is a spin-off from the development of antenna systems designed for spacecraft built by RCA’s Astro-Electronic Division, specifically the antenna system developed by the author for the World’s first weather spacecraft, TIROS 1, and its successors.
     Because the radiation patterns of the antenna system are vital to the successful operation of the spacecraft, many measurements of the patterns were required during the development stage to discover the correct physical and electrical properties of the antenna that would produce a satisfactory pattern while attached to the spacecraft.
     To obtain realistic radiation patterns of the antenna it must be mounted on the spacecraft during the measurements, because the spacecraft is in the radiation field of the antenna, and therefore distorts, or modifies the shape of the field that would be radiated from the antenna if it were operating in free space, i.e., without being mounted on the spacecraft.
     Consequently, to measure the patterns from all angles around the entire radiation sphere of the antenna, the spacecraft was mounted on a revolving pedestal having a freedom of motion that allows the spacecraft to be rotated on two separate quadrature-related axes, i.e., 90° between the axes.
     In the measurement setup used at RCA’s antenna test range, the entire spacecraft with its antenna was immersed in an electromagnetic field radiated from a ground-mounted log periodic antenna aimed at the spacecraft. The field comprised an RF signal modulated by a 1 KHz audio tone received by the spacecraft antenna. The received audio signal tone was conducted down the pedestal through a coaxial downlead cable and routed to the control room where it was recorded during the rotation of the spacecraft, recording the RF signal level.
     However, the vertically-oriented coax downlead distorted the radiation pattern when measuring either the vertical or circular polarization of the spacecraft antenna, but not when measuring with horizontal polarization. It was thus obvious that because the downlead was also immersed in the RF field, re-radiation from the downlead as a second source of RF, was distorting the field illuminating the spacecraft antenna and thus distorting its radiation pattern. Proof that the downlead was the culprit was obtained when manually moving the downlead three or four inches in any direction caused a variation in pattern level greater than 3 dB in any antenna orientation where the pattern level was a few dB below the maximum. Unacceptable.
     Consequently, to obtain true radiation patterns from the spacecraft antenna it was necessary to eliminate radiation from the downlead. I knew that a conductor of length lamda/4 or less could not sustain an RF current, but at the VHF/UHF frequencies involved in the measurements the downlead was several wavelengths long. One way to solve the problem would be to break the downlead into individual lengths, lamda/4 or shorter, and connect them with resistors that would effectively impede the RF, but allow the audio to travel. But the mechanical construction for this solution seemed impractical.
     During this era a new method of restricting flow of RF current on conductors was coming into vogue with the use of ferrites. Ferrite beads placed around a conductor allowed DC to flow, but restricted the flow of RF. I contemplated what would happen if I were to place an appropriate bead around the downlead at every lamda/4 point along the coax at the measurement frequency. This arrangement would effectively break the coax into lambda/4 RF sections electrically, but leave it intact physically. So I experimented with No. 43 bead material placed around the RG-58 cable at lambda/4 intervals, and voila’—no more radiation from the downlead and thus no more distortion of the radiation pattern from movement of the downlead. Accurate radiation patterns from the spacecraft antennas at last!
     Several months later, while listening to on-the-air discussions of problems that occurred when using wire-wound voltage baluns constructed around a ferrite core, a light bulb turned on in my mind. My immediate thought was, if beads impeded current flow on the downlead in the radiation pattern measurement setup, why wouldn’t it also impede common mode current flow on the outside of the coax feed line resulting from the balanced input of the antenna terminating the unbalanced coax? Why not indeed! I knew current was flowing on the outside of my feedline, because when measuring the impedance at the input terminals of the line using the General Radio GR-1606-A RF impedance bridge, the indicating null would disappear while running my fingers along the line. I knew from those symptoms that the common mode current on the feedline was also destroying the accuracy of the impedance measurements.
     I then reasoned that a bead resistance of at least ten times the impedance looking into one half of the dipole should reduce the current flow on the outside of the coax shield to one-hundredth of the power delivered to the dipole half connected to the shield, an insignificant amount.
     After researching the various terminal impedances that would be encountered with dipoles throughout the HF bands the worst case situation appeared to occur on the 75-80 m band, when operating at the low end of the band at 3.50 MHz with the antenna resonant at 3.75 MHz. At 3.778 MHz the terminal impedance of my nearly resonant 125-ft dipole was 64.83 + j0.18 ohms, for a 1.28 SWR. However, at 3.50 MHz the terminal impedance was 53.17 – j144.35 ohms, for a 9.6 SWR. (These values can be seen in Table 15-5, Page 15-14, and from Fig 15-1 on Page 15-15.) The magnitude of this impedance is 153.8 ohms at 69.8°. However, this is the total input terminal impedance of the dipole, while the dipole half fed by the shield, or outer conductor of the coax is only one half of this value, or 76.9 ohms. Thus a bead resistance of approximately 800 ohms should provide adequate reduction of common mode current on the outside of the feedline in this worst-case situation.
     After studying the specifications of several ferrite beads I ordered 300 No. 73 beads from The Wireman for experimentation. A brief report of some of the experimental data that led to the design of the commercial version of the W2DU balun are shown in Sec 21.6, Page 21-7. Additional data can be seen in Fig 21-3, Page 21-8, which shows the impedance, resistance, and reactance of 50 No. 73 beads versus frequency. Observe that the bead impedance at 3.50 MHz is slightly greater than 1000 ohms, amply sufficient to reduce the common mode current on the feedline of the 80-m dipole to insignificance.
     In addition to showing the impedance plot of my 80-m dipole from 3.45 to 4.075 MHz, the graph of Fig 15-1 also provides evidence that the common mode current on two different coaxial feedlines is insignificant when the 50-bead W2DU balun is inserted between the feedline and the antenna. As mentioned earlier, common mode current on the outer surface of the coax will destroy the accuracy of any measurement of impedance at the input of the coax. When measuring the terminal impedance of an antenna by measuring the impedance at the input of its calibrated feedline, the same impedance reading will be obtained regardless of the length of the feedline as long as its calibration is accurate. However, if a common mode current exists on the outer surface of the coax, different input terminal impedances will result if the impedance is measured using coax of different lengths, even though both are accurately calibrated. The greatest difference will prevail if the difference in lengths is lamda/4.
     Observe that in Fig 15-1 the solid lines represent measurements of the resistance, reactance, and SWR of my 80-m dipole made with a lamda/4 length feedline, while the dashed lines represent measurements made with a lamda/2 length feedline. Observe also that the difference between the solid and dashed lines is almost non-existent, indicating insignificant errors in measurement, and showing negligible common mode current flowing on the lines, thus proving the effectiveness of the W2DU balun in eliminating the common mode current. Note that the dipole data measured with both the lamda/4 and lamda/2 length feedlines plotted in Fig 15-1 appears in Tables 15-4 and 15-5.
     Thus endeth the story of the evolution of the W2DU ferrite bead balun.

The Tables and Figs 15...are available on my web page from Chapter 15.

Walt
 

[Steve - WB3HUZ]

Might be some answers here.

http://amfone.net/Amforum/index.php?topic=7694.0

[K1JJ]

HA!   I just copied that link and was gonna post it here... and see you just posted it...
http://amfone.net/Amforum/index.php?topic=7694.0

It shows for 40M I will need more like 12 turns on a 4.25" form to give about 560 ohms. The Yagis are about 50 ohms input, so it will be over X10.  I'll add some beads and it shud be FB.

Walt:  Very nice summary as usual.  I have been using ABS black 4" pipe for years now as the form to wind my coax choke  baluns and placed at the feedpoints of both Yagis and dipoles.  I use the pipe as the center insulator too.

Anyway, I had recently mounted some of that white PVC 4" sewer pipe (schedule 20) as the forms for the 40M Yagis. I never tested it in the microwave oven, so just tried it. I found it does get very warm to hot after about 60 seconds.  I then tried my old standby, the black ABS pipe and found it got only luke warm after 90 seconds.  I then tried some ceramic dishes and found they too, got warm after 60 seconds. Even the stock, glass microwave plate that sits in the microwave gets warm after 60 seconds. So, guess most materials do.  I should try some Plexiglas... Lords knows I have enuff of it.

Then again, there is a large difference between 7mhz and microwave, so maybe it's not significant.

So, bottom line is I will replace the PVC pipe with ABS pipe. Thanks for the heads up on that, Walt. Thankfully the Yagis are still on the ground under construction.

BTW, you alluded to the RG-213 coax being wound too tight as a 4.25" coil - is that right?   I've wound most of my coaxial  baluns that diameter for years at substantial power and frequency and have not seen any inner conductor migration problems, or coax failures. I could look up the specs on Google, but what is YOUR opinion on the limitations of tight coils for RG-213 or RG-11A ?(non-foam, using PE)   I'm assuming 1500w.

Later -

Tom, K1JJ

[KD6VXI]

Might be some answers here.

http://amfone.net/Amforum/index.php?topic=7694.0

And a few from K1TTT

http://www.k1ttt.net/technote/airbalun.html

--Shane

[W2DU]

Steve, your reference to the measurements made by WA2SQR is very enlightening, indeed. This data is the first detailed info concerning coax baluns that I've seen. The results are more encouraging that I had thought. Great data!

However, with so many different ferrite mixes abounding, I wish WA2SQR's data had included which mix he used in the measurements. The mix for one area of frequencies is not optimum for other frequencies. For example, when I did the development of the W2DU balun in 1982 the only mixes that were applicable to HF and VHF were #73 and #43, respectively. Since then some people have touted #77 as the way to go, but my measurements placed that mix as optimum for 2 to 3 MHz, and going downhill rapidly on frequencies higher than those.

On the other hand, I've learned recently that mix 61 (or is it 62?) far exceeds the capability of #73. When I return to my RF lab in Florida in Oct I plan to acquire some of this mix and measure them. If they do a better job than #73, as used in the HF W2DU balun, I'll have the Wireman change the beads to the better mix.

And Tom, the reason I was concerned about the 4.25" diameter PVC is that I thought at that diameter there would be sufficient force on the inner conductor to push through the dielectric onto the shield, causing shorting. According to the results of WA2SQR's experiments my thought was wrong. So carry on with the 4.25" and forget about my concern, it's gone!

Walt

[Steve - WB3HUZ]

Some more info on Z measurements

100 ferrite beads of Fair-Rite Products mix 43, with approx dimensions of  o.d. 0.562 in., i.d. 0.250 in.,
length 1.125 in. (per bead) on 10 ft.(!) of RG-142 (0.195-inch o.d., PTFE dielectric) coax. 

K3LR measured the impedance of one of these chokes to be:

Freq. (MHz)      |Z| (Ω)
1.8                    1152
3.7                    3483
7.1                    4115
14.2                  1783
21.2                  1280
28.5                  1234


Source: "Common-Mode Chokes", Chuck Counselman, W1HIS, 2006

[KD6VXI]

And Tom, the reason I was concerned about the 4.25" diameter PVC is that I thought at that diameter there would be sufficient force on the inner conductor to push through the dielectric onto the shield, causing shorting. According to the results of WA2SQR's experiments my thought was wrong. So carry on with the 4.25" and forget about my concern, it's gone!

Walt

I know of more than a few guys that ended up using hardline from the shack to the top of the tower.  At the top, they used a "superflex" type (LMR600 and up).  With the 'superflex' type of coax, I think your safe.  If you use the standard LMR type stuff, you might have coax center conductor migration, depending on the dielectric and the atmosphere.

I can say that some 'unknown' (Tram) branded RG213 in a 4 inch dia. choke did migrate, when I pulled it apart.  I went to a 6 inch and changed to superflex coax, and 'based' my model on the info on the k1ttt I posted above (which is a mirror of the data from huz, after I looked at both).  The heat would get > 110 degrees in the summer, and single digits in the winter.  See www.n6nb.com for a description of the location, my old QTH was 500' below his, and a couple miles away, as the crow flew.

You have a valid concern, Walt, and one that has been brought up for years by the PROponents of the W2DU style baluns.  The Proponents of the coaxial baluns bring up the fact that you can overheat the ferrite, so it's kind of a wash.  I've used both, in the past, on the same installation.  The type of weather it's exposed to can make a big diff...  That SAME 4 inch balun had been used in mild Santa Cruz, Ca weather (RARELY above 95 degrees) for a couple years, and no probs...  My SWR changed from the beginning of summer to the end, after moving, prompting my looking into it.

The other types of ferrite you mention (the 60 series mix) is what the "Battle Box" CB manufacturers are using now.  LOTS more L for a given amount of turns, from what I've been told, and it's supposedly better across a bigger freq range. 

hehe  I started this thread / mess   I'll tell BDW Hi next time I'm on that board for ya.

--Shane

[K1JJ]

Some more info on Z measurements

100 ferrite beads of Fair-Rite Products mix 43, with approx dimensions of  o.d. 0.562 in., i.d. 0.250 in.,
length 1.125 in. (per bead) on 10 ft.(!) of RG-142 (0.195-inch o.d., PTFE dielectric) coax. 

K3LR measured the impedance of one of these chokes to be:

Freq. (MHz)      |Z| (Ω)
1.8                    1152
3.7                    3483
7.1                    4115
14.2                  1783
21.2                  1280
28.5                  1234


Source: "Common-Mode Chokes", Chuck Counselman, W1HIS, 2006

100 cores = $330, depending on vendor and qty break.

I just bought qty 20  "FB-102 1/2" ID type 43 ferrite beads" from Palamar for $66 plus shipping. ($3.30 each)  I figgered 7 per 40M Yagi (three Yagis) plus the coax turns.  The example says about  41 ohms per bead, so that's about 287 extra ohms.  So in my case, I wud need the extra coax turns. 

At the price of ferrites, it doesn't appear an efficient way to go to get 500 ohms on some bands, cost-wise .. compared to rolling up some coax turns, that is.  $50 for 75M,  $40 for 40M, etc.  A long string adds some weight too, though probably a wash when compared against  added  coax turns.

T