Base New Vector 4K radiator at 0.75 - 0.82 - 0.875 wavelength.

[Marconi]

I think it's all sorted out.
The only odd thought I have is that I considered these antennas to be end fed by design.

Henry suggested the S4/NV4K could be fed at the base of the radiator and that much of the base element provided a near 50-ohm match point as well. I posted this earlier in this thread. See my post #7 above.

Attached below are 4 models. This time I used taper for the model's radiator. I did this so as to help with accuracy for the input of the dimensions that Bob85 posted some time back. Later I will try and tune these same models and compare any differences noted for using taper and correcting for Average Gain.

Attached PDF files content
.
1. is in Free Space with a gamma match.

On these free space models, I included an image that shows the Average Gain results for the models used. I just set these models to specs and did nothing to control the AGT results or alter the wires in any way.

I also added the wires data so the diameter and the number of segments used...can be easily compared. Only wires 1-17 make up the radiator and radials. All the rest of the wire are the radial hoop and the Top Hat wires.

My previous models might have has some minor differences among the wires. That said, I doubt it made any difference if there were any minor differences.

I owe you this Homer, so this time I went to the trouble to make sure there are no wire differences among the models.

2. is over Real Earth with a gamma match.

3. is in Free Space and is End Fed.

4. is over Real Earth and is End Fed.

I made notes correcting the overstated gain for the Real Earth models. I used the AGT correction factor and noted the math used to correct the gain on the image for the antenna RF patterns.

The End Fed mode still shows more gain. So Henry is right. hehe

Note: when I fixed the AGT results for the end fed model earlier...I got the Average Gain results to show perfect = "1" but the match would not improve. I'm thinking, I might have a similar problem when I try and fix the AGT results for the End Fed real Earth model. Stay tuned!

 

[Marconi]

Homer, just as I thought...nothing I do will fix the match for the End Fed model without adding some type of matching device. The AGT result is perfect, but the match shows 1.820 SWR with resistance = 27.92 Ohms, while reactance = - 4.635 OHM's...very close to resonance.

If you build an End Fed for this 3/4 wave with the radiator at or close to 324" inches long my model shows the match will be an issue to solve, or you will have to use a tuner. That should not be an issue, however, the match should not need very much transformation, and that is good.

Whatever you do Homer...I don't think the End Fed antenna will show a very good match if you try and direct feed it...like I think Henry was suggesting. Maybe he will stop by and have some better news on this idea.

 

[bob85]

Eddie,
Have you tried increasing the radial spacing to see of you get a better match,

if the end impedance of a 3/4wave ground-plane with drooping radials is higher than 50ohms and sweeping the radials up towards the radiator lowers impedance,
& makes the radiator electrically shorter increasing resonant frequency,

is there not a spacing that gives us 50ohms and a length of radials & radiator that gives resonance where we want it ?.

 

[Marconi]

Eddie,
Have you tried increasing the radial spacing to see of you get a better match,

Bob, I was trying to maintain specifications as best I knew them. I used the dimensions that you posted sometime back. No! I did not consider to change the dimensions except to try the change in the feed point that Henry told us would work. I can't see the technical side of antennas in my mind's eye like you can. So, I just stumble around until I find something that seems to work using Eznec. Since I can't work with real antennas anymore...all I have is modelling.

A few of us got to talking about Henry's idea for the Sigma4 he talked about in his report. He claimed a possible advantage by direct End Feeding this design, so I decided to model the idea using an NV4K model...just because I had recently finished working on that model.

Homer's gave us his opinion that the Sigma4/Vector were end fed by design. However, my models that were end fed did not respond well to the control efforts I've been using to better control the Average Gain Test results.

I see a possible glitch in my model trying to end feed this design directly with coax. So, I'll just take your word for it Bob...that it can be done.

Anyway, I doubt I could make a simple enough explanation to make my findings clear. I would end up maybe showing a bunch of models, likely in vain, and end up making everyone's eyes glaze over with a lot of words.

Maybe Homer will actually make such a model, so I'll just wait to hear the good news.

 

[Marconi]

is there not a spacing that gives us 50ohms and a length of radials & radiator that gives resonance where we want it ?.

I don't know Bob?

My not knowing the answer, is why I suggested to both Homer and Booty Monster to set their homebrew Vectors up close to specs first, get the match under control, and then have at modifying the antenna with some of the ideas floating around to increase the performance. Just watch out for ideas and claims that are "Shockwaved" or "Vortexed."

 

[HomerBB]

Marconi, what occurs with the match if the radials are isolated from the vertical on the end fed?
In my experience with other antenna types a high SWR occurs when there is direct shorting between the the coaxial conductors. Without the gamma to provide a matching device this antenna is just shorted... maybe?
Did we try the end fed with a cap hat?
Posting these questions surely reveals my ignorance.

 

[bob85]

I don't think you can direct feed it with 50ohm coax & have good vswr without altering the antenna Eddie, impedance is lower than 50ohm when the radials are so close to the radiator,

i think you could obtain good vswr with the right spacing & element lengths,

 

[Marconi]

I don't think you can direct feed it with 50ohm coax & have good vswr without altering the antenna Eddie, impedance is lower than 50ohm when the radials are so close to the radiator,

i think you could obtain good vswr with the right spacing & element lengths,

I don't think we can direct feed the Sigma 4 design either...unless the design is changed. Do you think I misunderstood Henry's point when he wrote in his "Room for Improvements" section #6, page #39, that the S4 could be direct feed and eliminate the losses associated with the gamma match.

I did the models below in Free Space to start, so I could track the Average Gain Test results in each iteration to a fair match.

For convenience, in the models below I used a 3/4 wave 333" x 0.625" radiator with 4 x 0.625" x 107" inch with radials slanted up at 45* degrees above 90* degrees at the horizon. The radials have no hoop. I used FS to start. I got to an acceptable AGT result of 0.963 and stopped. This value is acceptable but not perfect.

I was tracking the radiator length, radial length, radial angle above 90* degrees, AGT, SWR, Gain and Angle, plus running the SWR bandwidth to check resonance. Bob, this was like I imagined playing 3 chessboards at once was like. So, I stopped here and set the model over real Earth.

I also did an overlay for comparison to my most recent Sigma 4 over real Earth compared to this End Fed model. For comparison in words only with no images, Henry's Sigma4 model shows us an image on page #39 with gain of 4.16 dbi at 8* degrees. My Sigma4 model shows 3.96 dbi at 8* degrees. I can't explain the difference but I accept it as close. To be fair, I admit the patterns do show a slight difference...probably due to construction differences. Except for my using taper in these models and my once seeing Henry posting a S4 with a square hoop above the radials...I can't explain these differences.

 

[Marconi]

Marconi, what occurs with the match if the radials are isolated from the vertical on the end fed?

With Eznec, isolating the radials is not necessary on a gamma fed S4/NV4K. If you have an idea to do an end fed Vector, you will have to figure that out.

In my experience with other antenna types a high SWR occurs when there is direct shorting between the the coaxial conductors. Without the gamma to provide a matching device this antenna is just shorted... maybe?

Homer, I think basic antenna rules apply here.

Maybe consider something like you did on your plastic and aluminun foil duck tape AstroPlane?

A gamma is so effective and solves many of the problems that guy's in the know talk about regarding CB vertical ground plane antennas.

I've tried before to figure out the feed point area of some of your antennas like the Vector, but I'm always fooled on what was conductive and what was not. So, I gave up on trying to figure it out.

Did we try the end fed with a cap hat?

Try what? I don't model that little top hat on my Vector models anymore. About all it might do is ward-off some static and maybe lightening, but that is a big if. I also do it for convenience in modeling...when making changes in the radiator length.

Keep us posted on your progress with end feeding no top hat cuts down on things I try and watch.

 

[bob85]

maybe you did misunderstand what Henry said Eddie, or maybe i did,

Henry never indicated what would need adjusting in order to get a low vswr without the gamma so we can't think that he meant removing the gamma & adding an insulator at the base, with no other adjustments,

i presume Henry knows what would need adjusting,

looking at your model im guessing the wide radial angle is not suppressing the out of phase radiation from the lower 1/4wave as well as the original setup,

maybe moving the bottom end of the radials out further & the tips inward so you have a steeper angle with a similar average spacing would help.

 

[Marconi]

Do you think 5" inches wider at the bottom would be better?

 

[Marconi]

Bob, the first thing I did to make the bottom radial bracket wider to 5" inches offset...I discovered that the radials were not 112" inches as I intended in the final FS model I posted above.

Right now I figure I made the real Earth model base on this FS model...so it too probably has a little over 90" radials. This is what happens when I get too many irons if the fire.

I fixed the model and I see very little difference. I'm no more captivated by this idea now...than I was when I tried the end feed idea back when I read Henry's "Room for Improvement" ideas.

Maybe Henry has more to add to the good news.

Bob, I have an idea why the radials changed on me during the iterations, but if I told you...I'd have to kill you, it's the 007 code and M does not tolerate slackers spreading secrets.

 

[bob85]

have a look at how they make bottom fed j-poles Eddie, you may get a clue from those.

don't the segment ends need to be level as they would be if the radials were parallel with equal segment lengths, the wider the flare the less equal the segments become, could that be causing an issue?

 

[Marconi]

Well Bob, I knew I had made some mistakes and I couldn't put this End Fed model down without getting it fixed and tweaking it as best I could. I still can't get a really good match if I peak the gain in FS and over Real Earth...but I know now that Henry was right...Exnec will model and End Fed version of the S4/NV4K.

You will also see it is not that much different when I finally got control and maxed out the gain. At the end of my iterations I was making 1* degree changes in the radial angle and then shortening the radiator to remove the inductance added on radials being closer to the radiator.

After I got the maximum I tried to get the AGT = 1, but no luck. It is still very good at 0.987 and is showing the gain reported is understated by 0.06 dbi. So, the real Earth model is showing 4.16 + 0.06 = 4.22 dbi @ 8* degrees with correction error included...and that is not bad.

Later I'll make an overlay of my latest S4 and NV4K models and see how they look compared to this EF design.

Here are the FS and the Real Earth models. These radials really look close, but they are 14.1879" inches radius from the radiator center to center. The 5" base offset makes it look different and I did not get any geometry or segment errors because the base is so wide.

I'm out of energy for now. I hope I did not make some silly errors again.

 

[Marconi]

have a look at how they make bottom fed j-poles Eddie, you may get a clue from those.

don't the segment ends need to be level as they would be if the radials were parallel with equal segment lengths, the wider the flare the less equal the segments become, could that be causing an issue?

Good point Bob.

I used an idea the DB uses for the segments in this case. It helps me prevent making a lot of segment errors.

I set the segments for the radiator and the radials as close as I can to the same segment lengths. I also make the segments rather long Then I switch on the auto segmentation feature and it automatically sets any additional wires for the model to a similar ratio. The choice I used is noted as "conservative" auto segmentation and it typically does segment longer than I would otherwise use.

There is a caution in the manual that states the reason for auto segmentation and suggesting the feature is not always accurate. This is to say, that tweaking is still recommended as the rule when wires are close and parallel or near parallel and intersect at sharp angles. The 5" inch adjustment I added to the model at your recommendation help in this regard, but 5" inches was just a wild guess.

Thank you, Jesus.