de NØBEL, Kevin

[ This was written in April of 2007, and appeared in "The Repeater", our local repeater club newsletter. ]

After a lot of vacillation, I finally decided what antenna to put up.

After I started to realize that the "windom" antennas that are advertised and sold are not really windoms, I decided to put up an off center fed dipole.

The popular "windoms" that are out there all have about 1/3-2/3 ratio off center feeds, and radiating vertical sections (a-la g5rv).  Some radiating sections are open wire and some are even coax.  Then on top of that they call out specific lengths of coax to the shack. 

Then I finally started getting some real information.  An off center fed (OCF) dipole fed at the 1/3-2/3 point is resonant based on total length, and instead of a 75 ohm feed point they have anywhere from 200 to 600 ohms.  The interesting thing is that it is multi banded with no traps.   An 80m length of wire will do 80, 40, 20, and 10 (not 15) (and 6 too), all with nearly the same feedpoint impedance.

 

So, I started modeling one.  Initially nothing worked.  Then I remembered how important segmenting was to the modeling programs.  Once I did that, things started working.  Further research showed that people were putting in a 100pf cap in series with the longer leg at the feed point, to make 80 have a better swr.   I put that in and things worked in the model better as well.  Doing even more research I found that people were putting another OCF dipole for 15m in parallel with the main one.   Now all the sudden you can have 15 as well, with minimal interference with the other bands. 

I spent about 24 hours playing with the models, modeling where my antenna would be, and trying to fit it all in place.  One leg is 90ft and the other is 45 ft.  I couldn't find my 100' tape reel, so I couldn't measure all the points in the lot that I needed to, so I made some guesses.

What I end up with is an OCF going down and away to the north from the tower.  The apex angles are about 30 degrees down, but in addition the inside angle between the wires (as one goes north east and one goes north west) is about 90 degrees.  What I thought should happen did, whenever you bend a dipole at the apex, you lower the impedance.   So instead of the impedance being on the 400-600 side, its on the 100-200 side, with 80 being the lowest impedance.

Most OCF's are fed with a 4:1 balun..  50 -> 200.. about right for me after modeling.  I also found some OCF's feed with 6:1 and even 9:1.  But my modeling showed 200 should be good for me.

So, "somewhere" in the house or garage was a 4:1 balun that I bought years ago.  I also haven't seen my 2m and 440 mirage amps nor my 1296 transverter in a couple years.

Since I couldn't find a 4:1 balun, and its Sunday, I figured I'd see what it takes to make one.  During this research I see that there are two kinds of 4:1 baluns (and I guess all others too).   I knew that the traditional 1:1 balun with ferrite cores or the coiled up coax was a current balun.  The other kind of balun which used a torroid core was a voltage balun.  In the 4:1 type, a voltage balun has two windings, connected appropriately.  A 4:1 current balun can be on one core with 4 windings (two windings in one direction on one half, and two windings the other direction on the other half). It can also be made on two cores, with 2 windings each.  For me the current balun seemed better.  Most reports talked about the current version having more isolation and a larger impedance handling range.  Within that choice two core model seemed to me to be better, in power handling/heat dissipation at least.  So I dug up some cores of unknown properties (probably from an old triplite isobar) decided on some wire to use, and wound up my 4:1 current balun.  So after all this research, why would I use unknown cores?  Because its what I had.... remember its Sunday.

Now… how do I test the properties of a balun made of questionable materials?

Enter the antenna analyzer.  Ok, but I always questioned the absolute readings given by this meter because it aways seemed worse than other measurements showed.  However with my experience with the dx77, after using it with the radio, it seems that the analyzer was now showing better results than other measurements showed.  I wasn’t too worried at the time because all I was interested in were the swr nulls.

But now, I need accurate readings of impedance, and swr, to determine if I just made a collection of wires wound on some cores, or a balun.  Somewhere in my research of the previous few days, I ran across an article where someone recalibrated (and on one case upgraded parts) of their MFJ-259.  I found the article again, but it was referring to an MFJ-259B, which is really a different animal.   I then went to MFJ’s site, and found – a calibration procedure – for the 259!  I went looking for schematics, but I could only find stuff on the 259B.  The procedure I found was obviously for MFJ’s inside tech people, because it even pointed out problems with the high end of the impedance meter being inaccurate, so “make it work at 50 ohms, since most hams use this, and won’t notice problems in other areas”.  

So of course they use the old standard 50 ohm and 100 ohm resistor for testing.  100 ohm is of course 2:1 swr, and they use 50 ohm to set the impedance meter.   The two items I was looking for in a calibration procedure was simply what pots do I turn?   They were there.   R17 adjusts the SWR meter, and R31 adjusts the impedance meter.  After messing around with the 50 ohm, 100 ohm (and for good measure I used a 25 and a 200, to see how well the meter worked) I ended up with a reasonable result.   100 ohm measured 110 or 120, and 200 measured 250, basically just like the calibration document stated.  The 25 ohm reading was probably 23.   The SWR was quite accurate, even with the 25ohm which of course read 2:1.   Most of my testing of the meter was done at 3 and 7 mhz.  I did find more inaccuracies came into play above 26mhz, mostly in the impedance measurement, the swr seemed to work well all over.

Ok. So now I have a meter I can trust, so did I make a balun?  I needed a jumper from the meter to the balun.  Of course I couldn’t find one around the bench area.  Couldn’t find one around the radios.  I did find the jumper which I and removed the ends from to make my run out to the dx77.  Out to the garage again, looking for a jumper.   One box… nope… another box found a jumper, one end pl259 one end N.   Ok all my adapters are up at the repeater site in my adapter box.  Oh what is this box over here under some other stuff…. There’s a jumper sticking out, and as I dig inside of it to see what else is here for future reference… hmmm 2 mirage amps….. a 1296 transverter…. a coax coiled up ready for a dipole…  and what do you know… a 4:1 balun, of course.

The balun is a spiro-something “Pro-balun” 4:1, something I had picked up at the candy store 15 years ago.  While I was doing my research on baluns, I had seen this brand and model and found that this was a voltage balun.   So lets test it to see if it works!

So how would one use an antennal analyzer to test a 4:1 balun?   The 4:1 part is easy.  The balanced to unbalanced part I have to leave as faith.   So I hook up the my homebrew balun, and get my trusty 200ohm resistor.  I get an swr of 1:1 on 7mhz, with the impedance reading 50…. So, sometimes theory does match up to practice.   I then tune around and find that the swr and the impedance start moving around below 5mhz to where the balun is really not working,   and above 25 mhz its poor as well.   Based on my research on baluns I have some ideas as to why it didn’t work so well.

Well lets see how a “commercial” “PRO” balun works.   I connect that up in the same way and sweep it.  Interestingly its poor until about 2.5Mhz and above 30Mhz its poor as well.  So obviously the design criteria for this commercial balun is 80-10.

So why is a “broadband” impedance transformer not as broad banded as one would like?  When I was reading about baluns I found one persons description of balun design included a formula specifying that the minimum inductance of the windings the balun should relate to the maximum impedance that the balun would be seeing.  Basically, the inductive reactance, plus some safety factor should be the impedance which the balun sees.   Another article I read reminded the reader to consider stray capacitance when dealing with higher frequencies.

Putting this all together, my home brew balun wound on unknown cores which were small and therefore had only 7 turns for each pair of wires, probably didn’t have enough inductance to work at the lower frequencies.   In addition, the small cores meant that the windings were close, (packed in fact), which would cause more stray capacitance between windings.

However, when sweeping the commercial product, it became apparent to me to be sure to check the operating range of a balun, to be sure that it would work on the frequencies you want to use.   If I wanted to use this commercial balun, on 160m for example, based on my sweep and its poor performance below 2.5Mhz, it would not have worked.

So, I have a balun that will work (the commercial one), even though I’ll probably replace it with another one when I get the chance.   So I look for wire.  I need 45’ on one leg and 90’ on the other.   I find an old Radio Shack wire antenna… hmm the package says 75’.  I’ve always used bare wire in the past.   But that old field day 80/40 dipole we’ve used forever is insulated.  So I figured I’ll just go get some wire.   After a trip to Home Depot,  and after paying a ridiculous amount of money for a reel of 500’ of 14ga stranded electrical wire, I was ready to go.   It was of course raining by this time.

Rain always improves an antenna installation, so I continued, undaunted, if a little wet. Now one end of this dipole has to be anchored to an evergreen tree, as high as I can get it.  A little twine with a wrench tied to the end of it, and after a little exercise, I got it up in a decent place.  I decided to do something simple and wrap the twine 3/4 of the way around the tree and tie the twine to a stake on the ground.  This would let me drop the antenna more easily as I adjusted it.   With that done, I went back down to the workbench and prepared the balun with the 100pf cap, figured out how I would connect it to and support the wires.

I measured out 48 and 96 foot sections (the front facing of the garage is 24 feet wide) of wire to start off with.  I wasn’t going to mess with the 15m piece until after the main part worked.  I figured I could prune as needed.   I wasn’t sure what the insulation would do to the length, although a quick check of the internet showed most people claim the insulation caused you to end up with a dipole shorter than the normal calculation.

Its still raining, so I figured this would be the best antenna yet.   I attach the short wire to the balun, and go up the tower, and get it attached.   At the same time I decided to remove the my 20 year old discone (dummy load) antenna, since this was the coax run I was going to use.  It gave me great pleasure to drop that antenna to the ground and watch it go splat.   I went up again with the 96’ wire and connected that, and then connected the coax, which was miraculously exactly the correct length.  I then stretched out both ends of the antenna.  The short end going to the evergreen tree went up quite nicely and the insulator was about 3 feet from the tree.   So far so good.   I went to the long wire.   Here, I had difficulty getting the length, missing the satellite dish, missing another tree and attaching to the fence.   For the moment I had about 10 feet of extra wire weaved along the fence.   Well it’s a wooden fence, and its insulated wire… lets see what happens.

Ok… its raining… I did about 4 free climbs up and down the tower, my antenna is weaved through the fence at one end.  Interestingly enough, I feel pretty good, except for a little bit of sweating.   My wife is just shaking her head, now having proof, rather than suspecting, that I’m crazy.

Now I put my newly calibrated antenna analyzer on the coax…. And sweep the bands… And I get nothing, nada, bupkis.  Swr is infinite everywhere.  The impedance however does have some interesting transitions from infinite to zero and back again a few times through the HF bands.   Some of the transitions correspond to the predicted resonances that the antenna models showed.  My first thought was that I had the wrong coax run.  These infinite/low impedance transitions could be showing me that I had an open or shorted coax run.   So I re-inventory my coax runs.   I only have 4 runs.   One to the dx77, one to the old discone dummy load (now the new dipole), one to the Diamond 17’ 2m/440 vertical, and one to the satellite antennas.   I had identified each run a couple weeks ago.  And used the ICOM 820 to verify I had the right antenna for the satellites.   I had the right run for the DX77, and the 2m/440 vertical had been attached to my stack of 2m/220/440 radios forever.   This was the only run left, so it had to be it.. right?  

Ok, I’m sure the coax is right… so lets hack off the ends of the dipole (remember 96’ instead of 90 and 48’ instead of 45) to their proper length.  I figured this would also improve the fence wire weave problem.

So I did all that, and climbed up the tower once more to make sure that nothing had come apart.   Its still raining, a little windy, temperature is falling, and now, its getting dark.   With all that going for it, this antenna should be nearly perfect!  I figured it would be pulling in dx even when the bands were closed!

Back in the house, plug in the analyzer…. Same result infinite swr everywhere.  Not a null to be found.   Ok, it must be the coax.   Guess where my dummy loads are?   Up at the repeater site.   So I find an SO-239, solder on one of those 50 ohm resistors I used to check the analyzer, and climbed up again…. windier now…  maybe the dummy load will pull in the dx.

The analyzer still reads infinite swr… ok its either a bad coax, or the wrong coax.   So I disconnected the coax from the back of the 820, and plugged it into the analyzer…. And what do you know…. Its all flat 1:1 swr everywhere.   I guess that’d be the dummy load huh?... This coax mixup is a lesson in challenging your assumptions.   When I was on the tower earlier in the day, I had seen a coax cable tangled in the old discone.  It had an N connector on it, and also had some twisted wire taped to it.  It ran up to the top of the tower.  It was the UHF satellite antenna feed line.  The line that was supposed to be running from the mast to the back end of the satellite antenna and the polarity switch box.  Aparently over the years, the N connector had unscrewed itself.  After that I dismissed the issue as something to fix when I reworked the satellite antennas.

Well, when I identified the coax runs those couple weeks ago, I had used the idea that whatever coax worked best for the 820 would be the satellite antennas.   However what was the case instead is that the 820 like the discone rather than the satellite antennas, due to the bad shape (and disconnected coax found today), that they were in.

Ok, so now I have the right coax.    One more run up the tower… darker now, windier, but it stopped raining (damn).   Back to the house, analyzer connected (to the right coax), and I actually see swr nulls!... and even in the right places!

A little bit of study with the meter shows it works on the very high end of 80 and about the middle of 40 and the lower end of 20.   So, lets see if the radio like it.

The 736 showed low swr in the same places the meter did.  I didn’t check absolute values, but its in the ballpark.  I did some testing and low power transmitting, and the first thing I noticed was the transmit button seemed to “stick” on 80M.  I had to turn off the rig to stop it.  I finally figured out that if I removed the usb cable from my usb/sound/serial radio control box, the tx sticking would stop…. Can you say “RF”?  

The TV right next to the radio, which never exhibited problems on the DX77, had some beat bars on it (lots of beat bars, but all in all low amplitude)   Ok lets try 40m…. no problem with the tx button, but more beat bars, and a black and white picture.   This TV is connected to the dish receiver via video cables, so the tuner isn’t being messed up, the the electronics are being overloaded with RF.   Hmm, what about 20m?   The tv picture wiped out entirely.

The way the wire antenna is configured, the house (and the second floor where I the radios are), are within 10 or 15 feet of the wires.  Both wires go over the house.   At this point I’m thinking that the problem is simply antenna radiated RF mostly.

This is where I left things.  I was late enough by now, that I decided to monitor the various to the geriatric nets on 75 to see how the receiving worked on 80.   Which I have to report worked very well.  Noise was S8, which I didn’t expect, since its pretty quiet on the vertical on 40, with maybe S2 noise with an open band.

I’m happy that the antenna worked as it should have.   Again, theory and practice have some intersection.  I haven’t tried to tweak it yet. I need to move the low swr frequencies down some on 80 and up some on 20.  I’ll work with the model to figure out which ends to cut (or lengthen).

Also since the satellite antennas are in worse shape that I thought, and since there are no high orbit satellites up there, I’m going to move the apex of the dipole all the way to the top of the tower, maybe even remove the satellite antennas and get it a couple feet higher onto the mast.   This should help me get the longer wire clearer of a tree, and help if I have to lengthen it.  Also higher means farther from the house roof, and rooms, so maybe the RF will diminish.   Lastly I think I want to put some cores on the coax feed since its probable that the feed is picking up lots of RF from the antenna.   Since I’m using a voltage balun for now, there also might be some current on the coax (maybe I’ll make up a coax choke for that).   In addition to cut the amount of  RF that the feed is getting, I’ll probably put some cores on the coax where it comes into the roof.   In a week or so, I’ll probably be covering the entire house in aluminum foil…. Or maybe not.