Fat EFHW

[HomerBB]

Modeling is not among my skill sets. I don't see where they usually take in some data that might be relevant . . . but I don't know

One observation in real antenna comparison between the EFHW 1/2 wave, and the 1/4 wave GP, and the 5/8 wave GP is referenced to the ways that folks commonly seem to employ radials with each of them, perhaps because of the matching networks.

The 1/4 wave GP isolates the vertical from the radials completely and uses the separate feed line conductors on each part, very similar construction to a dipole.

The 5/8 wave GP isolates the radials from the vertical, but re-establishes a grounded connection between the radials and vertical radiator through the matching system, inductor coil, loop rings, or trombone, etc, all of them providing a positive ground between all parts of the antenna.

Yet, the EFHW often simply gets a set of wire radials attached to the bottom of the antenna without respect for the type of matching network employed at all.

"Hey, just attach three or four slanted/horizontal wires to the bottom of the antenna and make a GP out of it."

Perhaps consideration for HOW the radials are integrated into the system would make a difference in results?

Just asking . . .


When I experimented with an EFHW with and without GP I was careful to follow the same practice I used with the 5/8 GP - isolate the antenna halves from each other, provide positive ground potential through the matching network, and keep working to tune the radials lengths, tap point on the matching inductor, and amount of capacitance by adjusting the variable cap in the matching device until the resonance of the antenna, the SWR match, and the best response from the antenna seemed to coincide.

I believed then, and have seen nothing to persuade me otherwise, that I detected a performance difference in terms of potential gain improvement over the antenna without the radials. It was the best 1/2 wave I had ever used. Not having any way to prove this beyond my experience I simply filed it away as a possibility I trusted was possible to duplicate.

 

[The DB]

DB, I've been claiming the same for a long time and maybe to little avail.

Here are two of my EFHW models at 18' feet with and without radials. I did not do an overlay, but you can see the patterns are very close like yours are, and the angle did not change. There is little to no currents shown on these radials either, at least I can't see any. My currents log might show some however.

View attachment 13242

For me it is like the radials are not there, but I do see a noticeable reduction in the FP match, like the difference between 100> SWR to 30:1 SWR, so the radials are doing something.

Personally I would much rather deal with a natural match at 30:1 than one over 100:1 SWR.

There is so much I don't understand about modeling and antennas, but I get similar results to what your show here. Maybe someone can give us a link to an EFHW model that produces different results and proves the claims that radials are a benefit to an EFHW, and that adding radials also improves the TOA.

With how an end fed half wave works there really shouldn't be any currents on the radials to begin with, assuming you are actually close to a half wavelength. I can tell ya that, not from modeling experience (I don't have much, and far less than you Marconi) but from my studies on antenna theory. At the very least I think the model is accurately showing currents. If I apply an L matching network there are even fewer currents on the radials.

Adding in L matching networks, with the coil in series (the networks act as a low pass filter) changed the total gain some, they are closer together overall, very close now actually. It is conceivable that depending on how the matching was done and the height above ground that you could get a benefit to the EFHW with radials. The radials model shown below is actually slightly more efficient than the no radials model at a half wavelength above ground.

It is common for people to claim that radials effect the angle of radiation, but I have yet to see it in modeling, and based on antenna theory I disagree with the premise. The earth beyond the radials will always have far more of an effect on the angle of radiation than the radials themselves. The radials simply don't extend far enough to make a difference of more than a few small fractions of a percent.


The DB

 

[HomerBB]

This link provides a photo of the type matching network I am using. It is in a smaller box, but the same thing.

http://www.worldwidedx.com/cb-antennas/138558-end-fed-1-2-wave-antenna-10.html#post371870

 

[Captain Kilowatt]

I decided to model a half wave with no radials, then copied it and added 1/4 wavelength radials to the copy and compared the two, this is what I got...

Note: The red plot is from the model with radials, the blue plot is from the model without radials.

I don't see much difference so I raised the height of the models...

Still not much of a difference...

Looking at the models I noticed something...

I see next to no currents on the radials themselves. If I change the length of the vertical element either direction (longer or shorter) it isn't long before currents on the radials begin to rise, but near the half wavelength point of the vertical the quarter wavelength radials have next to no currents flowing on them. As there is next to no currents flowing on them it is almost like they aren't even there.

About the only benefit I see adding radials is the model with radials has a slightly closer natural match, but even at that, the mismatch is still extremely high as would be expected with a half wavelength end fed antenna...

So I ask again, why bother with the radials?


The DB

I must admit to being a little surprised at those plots. I would expect to see some difference although nothing drastic. Anything I have ever read or studied stated that even a half wave vertical would benefit from radials. Even in the broadcast industry any station using half wave antennas must still use the FCC standard of 120 radials at the base.

 

[The DB]

i like to see more real world test results. imho these tell ya how it SHOULD perform
in a perfect world. however thats not the case. real on air tests MIGHT differ
from this im curious if would help decouple the coax so its not acting like a antenna
[a issue with both the a99 and imax ]and maybe lower the floor noise level a bit.
another comparsion i did maco vs imax the maco recieve was much quieter
same with my sirio .im no antenna genious i only know what ive tried and seen
and its been my opinion that especially the a99 is a incomplete antenna without
radials,yet everything ive read tells me im wrong.i view the a99 as a below average
to average antenna at best.

Ah yes, opinion. I love opinions. My opinion on modeling is quite unique. I've said things for and against modeling, specifically the NEC2 engine, the only engine I have any direct experience with, on this very forum. My opinion has admittedly changed as I've done more and more modeling over the last year, but it still can be considered harsh at times. The key is understanding the limitations of the tool that you are using.

Every tool, including your favored real world tests has its limits. A single real world test is far more useless to you than a model, unless the real world conditions that test was done in exactly match where you plan on putting an antenna yourself. Good luck with that.

I've said some harsh things about modeling in the past that I stand by, but at least I have some experience to back it up, even if I am still only an amateur at it. The weight of the opinion of someone who apparently has no experience at something, and apparently no intention of getting any experience or even a basic level of understanding at that something, really doesn't mean much to me. Especially when that opinion is spouted as a fact that you have no way of backing up. Sorry, but you don't have enough real world experience with modeling to even be able to properly put it in perspective with the so called "real world results" that you desire.

You assuming modeling only models a perfect world is a perfect example of what I am talking about...

Moving on...

When it comes to decoupling with the half wavelength antennas modeled, I can simulate the outside of the coax in the models by simply adding another wire. I can even simulate differing lengths to show what happens as you change the length of the coax and have it attached to an earth if need be. I didn't do that above but I can... At worse this makes quite a bit of difference depending on the length of the feedline between the antenna and the radio and weather or not it is grounded at some point, but these differences are also easily minimized in the real world by adding in a well designed choke.

Modeling does not determine how quiet of a receive you will have to my knowledge, and I have never seen anyone make such a claim using modeling, so to the statements made on modeling it is irrelevant. That being said, you are not the only person to make such a claim, although that is environmental. I have seen claims by people who live in a country environment say that the same change had absolutely no effect on their noise floor. Variances in the real world strike again...

When it comes to the a99, I think it is a crap antenna as a whole. Its weakness is not, in my opinion, that it has no radials, but that its matching section is, well, crap. I don't know that it would even be possible to model the internal matching network but I would love to see it done if only as a measure of the losses present within it. Even adding radials to this antenna is not as easy as you might think as the base of the antenna is a more than a foot down from the matching network which internally has a coax run with its common mode currents, which effects the length the quarter wavelength radials need to be...


The DB

 

[2RT307]

This link provides a photo of the type matching network I am using. It is in a smaller box, but the same thing. end fed 1/2 wave

Link is not working on my end, Homer. May be the Germans telling me to get back to work, though.

73,
RT307

 

[The DB]

I must admit to being a little surprised at those plots. I would expect to see some difference although nothing drastic. Anything I have ever read or studied stated that even a half wave vertical would benefit from radials. Even in the broadcast industry any station using half wave antennas must still use the FCC standard of 120 radials at the base.

One thing that may make a difference is where the 120 .4 wavelength radials fits in is on ground mounted antennas. The models above are .5 and 2 wavelengths above the ground. I don't think it would significantly affect the pattern itself, but it should effect the efficiency of the antenna, giving the antenna with more radials more gain. I don't like modeling things that low using NEC2 as the the results from wires below .2 wavelengths high are not accurate. That is a known limit with the NEC2 engine. That was apparently fixed in the NEC4 engine, but they want more money for that than I'm willing to pay at the moment.


The DB

 

[Road Squawker]

Gee, isn't the term "EFHW 1/2 wave" kinda redundant?

 

[The DB]

One observation in real antenna comparison between the EFHW 1/2 wave, and the 1/4 wave GP, and the 5/8 wave GP is referenced to the ways that folks commonly seem to employ radials with each of them, perhaps because of the matching networks.

These changes are necessary are they not? These are three very different antennas and all have different feedpoint requirements.

The 1/4 wave GP isolates the vertical from the radials completely and uses the separate feed line conductors on each part, very similar construction to a dipole.

The 5/8 wave GP isolates the radials from the vertical, but re-establishes a grounded connection between the radials and vertical radiator through the matching system, inductor coil, loop rings, or trombone, etc, all of them providing a positive ground between all parts of the antenna.

Yet, the EFHW often simply gets a set of wire radials attached to the bottom of the antenna without respect for the type of matching network employed at all.

"Hey, just attach three or four slanted/horizontal wires to the bottom of the antenna and make a GP out of it."

Perhaps consideration for HOW the radials are integrated into the system would make a difference in results?

Just asking . . .

The 5/8 can go either way with respect to ground depending on how they matched it, and that is technically true for any antenna, but yes, all for the most part true. The thing about the 5/8 wavelength antenna, and this applies to the half wave antenna as well if it is slightly longer than a half wavelength on the frequency of operation, is that downward slanted radials will actually work against the antennas radiation pattern enhancing the high angle lobe.

When I experimented with an EFHW with and without GP I was careful to follow the same practice I used with the 5/8 GP - isolate the antenna halves from each other, provide positive ground potential through the matching network, and keep working to tune the radials lengths, tap point on the matching inductor, and amount of capacitance by adjusting the variable cap in the matching device until the resonance of the antenna, the SWR match, and the best response from the antenna seemed to coincide.

I believed then, and have seen nothing to persuade me otherwise, that I detected a performance difference in terms of potential gain improvement over the antenna without the radials. It was the best 1/2 wave I had ever used. Not having any way to prove this beyond my experience I simply filed it away as a possibility I trusted was possible to duplicate.

Just a quick question with your comparison between half wavelength antennas, did you use the same matching network design and components just tuned to the various different half wave antenna layouts?


The DB

 

[HomerBB]

Gee, isn't the term "EFHW 1/2 wave" kinda redundant?

Yep. I did it didn't I?

 

[HomerBB]

Link is not working on my end, Homer. May be the Germans telling me to get back to work, though.

73,
RT307

FIXED

 

[HomerBB]

These changes are necessary are they not? These are three very different antennas and all have different feedpoint requirements.

To some degree, yes, but the point was/is that we tend to precision with other antennas respecting where/how we locate radials, but tend to just assume tossing a set of radials at or near the base of a vertical half wave constitutes a well done ground plane. I wondered whether this was taking place within the modeling results, too.

The 5/8 can go either way with respect to ground depending on how they matched it, and that is technically true for any antenna, but yes, all for the most part true. The thing about the 5/8 wavelength antenna, and this applies to the half wave antenna as well if it is slightly longer than a half wavelength on the frequency of operation, is that downward slanted radials will actually work against the antennas radiation pattern enhancing the high angle lobe.

And so it seems that the placement and design of the radials is critical at least in some applications to antenna performance.

Just a quick question with your comparison between half wavelength antennas, did you use the same matching network design and components just tuned to the various different half wave antenna layouts?


The DB

Yep. The linked photo lies within that thread. I worked with the same 1/2 wave vertical carrying it through several nuances. My comparison was against my experience before and since that antenna here at my QTH.
I realize it is anecdotal reporting, but I went out on a limb and still think it was satisfactorily conclusive enough for me that others may see similar responses to the effort should they attempt to take the time. Much like your expressed opinion of the use of modeling.

http://www.worldwidedx.com/cb-antennas/138558-end-fed-1-2-wave-antenna.html

 

[HomerBB]

I like opinions, too. I like mine the best. here is one of them

It may be that modeling is fraught with more limitations than how close to the ground the antenna modeled can be trusted for accurate results of the model.
I like models, and find them interesting. Models may indicate potential for antenna designs, and offer insights into design flaws or benefits, but they have limits that are more far reaching than what I am willing to trust as infallible.
I think their real value is in use for comparing one modeled antenna setup against another modeled antenna setup.
I think when I work with a three dimensional real world antenna the best way to figure out how well it works is against other real world antennas.
Comparing models to real antenna results will never report the accuracy of environmental differences, as you have said.

Modeling software is, after all, software. It evolves with better coding, and understanding it is the primary purview of the geeky among us, so who can know it? Someday it will be simple enough to just put info of what you want to modelin a few blanks and get results. Until then, it temains on my suspicious of it list.

 

[The DB]

To some degree, yes, but the point was/is that we tend to precision with other antennas respecting where/how we locate radials, but tend to just assume tossing a set of radials at or near the base of a vertical half wave constitutes a well done ground plane. I wondered whether this was taking place within the modeling results, too.

And so it seems that the placement and design of the radials is critical at least in some applications to antenna performance.

Yep. The linked photo lies within that thread. I worked with the same 1/2 wave vertical carrying it through several nuances. My comparison was against my experience before and since that antenna here at my QTH.
I realize it is anecdotal reporting, but I went out on a limb and still think it was satisfactorily conclusive enough for me that others may see similar responses to the effort should they attempt to take the time. Much like your expressed opinion of the use of modeling.

http://www.worldwidedx.com/cb-antennas/138558-end-fed-1-2-wave-antenna.html

In the modeling I put the feedpoint as close to the bottom of the vertical element as possible. I suppose I can move the radials around a bit, and do like the a99/imax antennas do and put them further away. One thing that comes into play is these antennas all have a matching system, which should, in and of itself, tune the radials as well as the antenna, perhaps making the absolute length of the radials less critical on an antenna with such low radial currents to begin with?


The DB

 

[Captain Kilowatt]

One thing that may make a difference is where the 120 .4 wavelength radials fits in is on ground mounted antennas. The models above are .5 and 2 wavelengths above the ground. I don't think it would significantly affect the pattern itself, but it should effect the efficiency of the antenna, giving the antenna with more radials more gain. I don't like modeling things that low using NEC2 as the the results from wires below .2 wavelengths high are not accurate. That is a known limit with the NEC2 engine. That was apparently fixed in the NEC4 engine, but they want more money for that than I'm willing to pay at the moment.


The DB

You may be right. I cannot for the life of me get the broadcast biz out of my head. Almost without fail every time we talk about vertical antennas and groundplanes etc I immediately think of ground mounted antennas. old habits are hard to break.