SIGMA IV Details

[Shockwave]

The next logical question has been asked many times. If the 5/8 wave ground plane has 1/8 wave length of inverted currents, how can it work better then a 1/2 wave for many people? Would this mean that another 1/8 wave of in phase wavelength is wasted on the radiator to cancel these affects? Leaving you with less in phase radiation then a half wave?

The short answer is not all the time. Antennas are complex devices even though their components are simple. There is no black and white here. It's not like you're going to fall off the edge of the ocean as soon as you pass 1/2 wavelength. Although, two very distinct things begin to happen as soon as you do pass 1/2 wavelength.

This is the point where the first secondary lobe will begin to form that is not a result of ground reflections but the result of inverted currents. That higher angle secondary lobe also creates a null between it and the lower primary lobe. Here is where the balancing act to fit your target area begins.

As that null just begins to form, it compresses the primary lobe in such a way that it can be beneficial over useful radiation angles. The trick is getting the length that gives the most useful compression to the primary lobe before shifting too much energy into the wasted secondary lobe.

 

[Marconi]

Shockwave there is a lot to try and digest here, and I'll try that later. I have several points where I think you and Homer misunderstood my words or I said what I meant badly.

I believe you are sincere in your words, but I see things differently in my models sometimes, and I only wish to try and find out why or the truth.

I will talk about one issue we might can talk about later, but I originally made my Sigma4 model to specs, because I had the actual antenna to measure. When I finished, I didn't see all the pitfalls that other's said I would, like Eznec cannot model this antenna without showing very high maximum angle of radiation, or the program could not recognize the radials. I referred to you as making some of these claims, but I could be wrong.

To single out anybody is not or was not my point in modeling the Sigma4 or trying to talk about it here, but it seemed logical to refer to the folks that have already joined in before. This New Vector model provoked me to do this work. I would just as soon not mention names, but that is not very descriptive when I'm trying to compare notes as it were.

I'll cut to the chase.

When I recently modeled my New Vector to Bob's dimensions posted a while back, I noticed the Vector was somehow different from my Sigma4, besides the obvious. It showed better gain at the same angle, but the patterns were not similar, and the Vector showed a much higher, high angle lobe even while showing better gain. This seemed strange to me, but I was sure the longer radials on the Vector made the difference, in spite of the information we have about adding the hoop and its effect on resonance of the 1/4 wave cone. Before I started this, I did some models with just the a 3/4 wave radiator and the 107" radials at different angles, starting with horizontal radials. I think I saw much of what you described above. I even made my version of your wide angled model and I posted it too, but I didn't get one response.

I studied and compared my S4 to this new V4, and I noticed that the phasing in these two were totally dissimilar, as well as the patterns they produced. I kept close tabs on the settings to try and insure I was comparing these models fairly, and I used Cebik's cautions for Eznec modeling ideas as best I could and as I understood them.

I also modeled my S4 using 107" radials, and in that case the S4 with the longer radials and the New Vector showed similar results in pattern, and radial phase...relative to the radiator. Both showed a rise in the high angle lobe area, and I never was able to understand why.

Then I got the notion to model what I thought would be your idea of your .82 wave Dominator in CB. To my surprise, with its 107" radials, I saw similar pattern results to my S4 to specs model. Of course the longer radials in your model, I suspect, allowed a nice advantage in gain with your design, at 4.73 dbi @ 8* degrees vs. 4.46 dbi @ 8* degrees, even while the S4 showed a better match by about 50%. These antennas of course do not have matching devices modeled. I have not adjusted the antenna gain for Average Gain values yet, but I noted the values on the models.

The big surprise for me is the pattern I see with your model and my S4, compared to my S4 with 107" radials and my New Vector model. In addition both my S4 and your model, show the currents in the radials and the radiator are in phase, while the other two are not and I do not understand that as yet either.

All I've ever seen before doing this New Vector with its out of phase condition, was my S4 showing where almost the whole antenna is in phase and radiating, just like my idea for your .82 model.

This is why this all popped up on the scene again, no other reason, well maybe with the exception we have a new buddy, kt4ye Bill posting this thread.

You might take a look at these models, and I think you will see what I see. You might also see the failings in the effort as well, but I was impressed with your idea for dimensions, but I'm also impressed with the results I see with my original Sigma 4, particularly in the area of the pattens.

View attachment Shockwave's .82 Vector.pdf

View attachment Sigma 4 Per Manual w mast 36'.pdf

 

[Marconi]

Shockwave writes:

If they were parallel to the vertical radiator and this were a balanced antenna with equal currents, the cancellation rule would apply. Neither condition has been met.

Marconi writes:

So Shockwave, are you suggesting that because the slanted up radials are not fully parallel to the radiator and the current magnitudes are not perfectly equal…that the rule for cancellation does not apply?

I realize this cone is not fully parallel like coax line, but I thought cancellation could occur in all manner of magnitudes and degrees to some extent, and that cancellation/radiation were NOT like a light switch.

If you’re right, then how do we get cancellation inside of the cone? Isn’t that basically what you guys are suggesting is going on with the Sigma/Vector design, and that CMC’s are radiating from the hoop down on the radials?

Stop thinking I'm holding out on you Marconi. The only tabulated logs I have for this design are what was required by the FCC and Industry Canada and they are for the total relative field, not individual element currents. Have you considered searching online for information regarding CST to help you interpret the information it provides? I think you'll find the old analogy "a picture is worth a thousand words" fits well here. The very questions you are asking about the currents and phase on individual elements is all there. Homer is able to see blue and yellow radiation emitting from the outside of the cone. Just compare that to the chart on the right for phase and magnitude as related to color.

No SW, this is the first time I’ve asked you for more information on the current data. I think you’re confusing me with another member who used to hound you for some more information on some issues in discussion.

If you can't support the data that your graphic is based on, then you can’t. End of story. Then we either see what we believe, or believe what we see in the graphic, which for me is based strictly on faith in advertising.

I asked the question of all the guys about what they saw in the graphic. I wanted to know how they described what they saw, and thus far…no takers.

The graphic is a novel idea, and believe it or not, I think I see close to what you and other’s claim to see, maybe with the exception of where and what exactly the current phase is at any point in the antenna. It is obvious that the colors are indicating what is supposed to be + and – phase, but that is not a picture of the antenna.

For me, looking at amp/m in the Eznec currents log, the numbers in the graphic don’t seem to jive with what I’m use to seeing. They look way too small on the lower end and way too big on the upper end of the scale, and this scale looks to be severely attenuated.

The color depictions looks very similar to the those I see with Eznec red lines, but I get the chance to see the sign of the phase by checking the currents log, and I can see how the currents are working at the segment level and exactly where they might reverse. If I understand this correctly…that can be quite revealing.

I'm not sure where you're getting this from Marconi. I've said many times that it is the cone on this design that PREVENTS the high angle lobe from becoming dominant on this 3/4 wave. I've also urged you to model a standard 3/4 wave with anything other then radials that sweep upwards so you can clearly see all other 3/4 wave ground planes have this lobe and sweeping the radials upwards puts more energy on the horizon.

Yes SW, like I already said, I did the ¾ wave thing. That is what got me started on this Vector project and when I saw that the Vector did not seem to follow the idea you note above. This is when I started to question what was going on.

Up until I saw what longer radials could do, and what happened with this Vector model, I was one of the cheerleaders in the Sigma4 camp. I’m just looking for answers, and I don’t really care what side of the scale, the real answers fall on.

I realize you can’t and Bob won’t take the view I currently see…as worthy of consideration. You can’t, because you have a dog in this hunt. The only difference is…I can change my thinking at will. Hopefully both you and Bob will accept that I say this respectfully, and do so just because I think I see a fly in the ointment with what has previously been said on this subject, and in light of what I now see with this New Vector model. Again, besides the obvious, I can’t explain why the NV4000 pattern is so different from my S4. If you have modeled these antennas, have you ever noticed a pattern difference between your Vector and your Dominator version?

Model what I said and you will see what I said. You can already see the strong secondary lobe around 40 degrees in your model while the lower 8 degree lobe is dominant. Sweep those radials away from the vertical radiator and watch the power shift to the upper lobe.

I did that already, and I referred to raising the radials upward as pressing down on the high angled lobe. However on finishing the Vector model, I now see that adding the length to the radials on this Vector seems to cause this idea to fail. Additionally this model shows the radials out of phase with the radiator. SW, this revelation is all new to me.

I've shared everything I have access to in CST Marconi. I didn't spend the $2,500 that CST costs, that was provided to me by Sirio and it contains the info regarding phase and current. When people said the one still photo wasn't enough, I asked Sirio for the entire file containing the currents through one complete cycle of drive. I'm somewhat tired of having to always come up with more evidence. My problem with EZNEC+ is unfortunately it supports the 1/2 wave J-Pole theory by missing the fact that effective radiation is taking place on the cone. I can prove 100% beyond any doubt that EZNEC is miscalculating the radiation wavelength of this antenna by 1/4 wavelength or 90 degrees too short. I've explained to Roy Lewallen himself how I've spotted this 90 degree miscalculation and he is requesting more information from me regarding my claims on these phase angles.

I intend to build him a model that is modified in such a way that it will depend on his program reporting the correct radiation wavelength for it to line up with a real working prototype. I have a feeling I will have to send him the modified prototype before he believes me over the program. I'm prepared to do that because adding just 4 wires to the EZNEC model and 4 wires to a working prototype are going to demonstrate a clear physical deference in length of more then 8 feet on this band. That's a problem when the model is off a 1/4 wave and I'd love for Roy to fix it so I can use the less expensive program and not have to buy CST. I already wasted enough money having to refile inaccurate patent documents on improvements for the design that were calculated off the EZNEC+ model rather then the working prototype that couldn't be tested until last spring. That was a disappointment.

Are you saying that CST does not manifest the same results as you see when modeling your antenna in Eznec. How did you figure that out?

Good luck in working with Roy, I find that he is dutifully concerned with how his software works, maybe to the exclusion of other customer service.

Detune the antenna, especially that cone away from an electrical 1/4 wavelength and watch the common mode current problems creep up in the field.

I do take notice of CMC in my models. This is why I often model the mast, so that the CMC’s will be exhibited. I think most of these particular models appear to be mostly CMC free, at least to the level of being detected just using our radio.

Thanks for your comments, this helps me understand better.

 

[Marconi]

Shockwave writes:

Don't be confused into thinking that "one side" of the antenna is in an opposite phase as the other in CST. The difference in color between sides is so you can contrast the phase change along the length. There is no change from side to side. This is a difference in the visual interface much like EZNEC only shows current on one side of the antenna with it's single line. That doesn't mean there is only current on one side of the antenna. It's just one way of displaying the change in phase along the length of the radiator.

Marconi writes:

I think I get that idea. I probably made some confusing comments.

The visual model was not created or altered to display some desired goal either. CST inputs information much the same way as EZNEC does. You build the model to match the antenna, it provides many different areas of information. The posted model is just one thing it does. Showing not just the currents on the individual radiators, but more importantly how they propagate away from the near field and in the case of the Sigma, they combine.

I understand that SW. It is just like my models, I build em’ and they turn out looking Ok to good, or they look terrible. If they turn out terrible, I check em’ tooth and toe nail. If they turn out good, I do the same. They are what they are.

With the 5/8 wave ground plane you posted in CST, it's exactly the phase angle I would expect. Anytime you have an end fed single inline radiator longer then 1/2 wave or 180 degrees of the signwave, expect phase inversion to occur at the base. The length of this inversion will continue in the base to whatever length you add to the top of the half wave all the way until you reach full wavelength where it begins to come back into a constructive phase.

I just posted that image because I was on the Gain Master site looking for the name of the software that Sirio uses. I thought it might demonstrate another look at the graphic I was trying to discuss. My idea was to call attention to a model that I think most know has cancellation, the 5/8 wave vs. your antenna graphic.

 

[Marconi]

Shockwave writes:

The next logical question has been asked many times. If the 5/8 wave ground plane has 1/8 wave length of inverted currents, how can it work better then a 1/2 wave for many people? Would this mean that another 1/8 wave of in phase wavelength is wasted on the radiator to cancel these affects? Leaving you with less in phase radiation then a half wave?

Marconi writes:

I believe this idea is one that our Needle Bender holds to be true, but I do not.

I won’t argue the point however, because I find the idea for the physical length manifested for the bottom 1/8 wave length on a 5/8 wave radiator, does not jive with the length we might figure mathematically for this 1/8 wavelength. Example attached: shows that the inverse point occurs near the middle of the 7th segment in the model. The model shows a segment length of .97' x 7 = 6.8’ while 36 / 8 = 4.5’ feet. I just don’t wish to argue the point trying to explain, and I don’t believe I could if I tried.

View attachment .625 wave cancellation idea.pdf

I believe the only advantage that the 5/8 wave has over the end fed ½ wave is the slight height advantage noted if the antennas are compared at the same feed point height. There is also a height advantage for the maximum TOA that might be noted with the 5/8 wave at some particular heights. Otherwise, I see about the same results from both styled antennas if the height advantage is mitigated out of the comparison.

I recently did a model test of this idea using several different length vertical antennas that showed a nice advantage for the longer radiators, but when I laid all the antennas down horizontal, taking away the height advantage, all the antennas produced gain and angle figures very similar. It was almost unbelievable.

The short answer is not all the time. Antennas are complex devices even though their components are simple. There is no black and white here. It's not like you're going to fall off the edge of the ocean as soon as you pass 1/2 wavelength. Although, two very distinct things begin to happen as soon as you do pass 1/2 wavelength.

That is a similar point I was trying to make to you in my previous post regarding, your words, the fact that the Vector does not have perfectly parallel elements and thus the cancellation rule does not apply.

This is the point where the first secondary lobe will begin to form that is not a result of ground reflections but the result of inverted currents. That higher angle secondary lobe also creates a null between it and the lower primary lobe. Here is where the balancing act to fit your target area begins.

As that null just begins to form, it compresses the primary lobe in such a way that it can be beneficial over useful radiation angles. The trick is getting the length that gives the most useful compression to the primary lobe before shifting too much energy into the wasted secondary lobe.

I don’t figure too much in worrying about currents radiating in the higher end of the 90* degree range our antennas work in or the nulls either unless they occur between the horizon and about 30* degrees above the horizon. Your work I guess is different, trying to satisfy FCC rules.

 

[Marconi]

Homer writes:

I was going to supply the transmission line radiation terminology, but Bob beat me to it.
I agree with the notion that the cone is radiating due to a coaxial effect.

Common Mode Currents:

"Common mode currents can flow in all sorts of systems, not just antenna feed lines. . . . In an ideal world, RF flows down the outer surface of the center conductor of the coax cable cable, and returns on the inner surface of the coax shield. When there is an imbalance in the antenna (for what ever reason), current will flow on the outside of the coax shield. This may not seem possible, but it is important to remember, unlike DC, RF current doesn't flow through the conductors, it flows on the surface of the conductors. The current which flows on the outer surface of the shield is called common mode current. In other words, it is the unbalanced current not returned within the coaxial cable.
This leads to a very important question. If the current isn't returned in the cable, where does it go? The answer is, it radiates! In fact, the amount of radiation from the coax cable is proportional to the common-mode current on that cable. This leads to another question. What causes common mode currents? Well, hopefully this article will answer that question sufficiently."

I think this applies to the Sigma/Vector antenna system. We all know among other functions, the purpose of a GP radials system is to decouple the vertical, add an element of balance to the antenna, and serve to restrict common mode currents. I do not think these radials/cone are actually ground planes functionally, but due to how they are turned up to form a sleeve, are exactly what has been said of them - a coaxial sleeve that utilizes CMC to assist in radiating while capturing the opposing phase currents on the antennas lowest 1/4ƛ within itself.

Marconi writes:

I agree that is what we see going on in the graphic, and that has been the words of wisdom from on high for some time. And, until I did this New Vector model and saw it was so different, besides the obvious, I thought I would check it out as best I could and I notice that my Sigma4 model has in-phase currents throughout most of the antenna and for sure between the radiator and the radials. Is it right, I don’t know, but that is what I saw upon completing my first successful and accurate model of the Sigma 4?

If that was the case, then we don’t need any phase shifting or CMC’s for the antenna to radiate constructively, I don’t think. Now on the other hand I might agree with you on my New Vector model that shows the radials out of phase with the bottom of the radiator that is surrounded by the radials. I'll even concede that it looks and sounds reasonable that CMC could be produced and spill over, for lack of a better description, onto the radials...like we’ve learned to understand about feed lines and CMC’s. However, there is a problem with the condition at the top of the hoop, and Shockwave as said it himself, “…there is a lot of voltage at the top of this hoop, so don’t place the radial supports too close.”

Here is my point specifically. Isn’t this high voltage, high impedance point exactly what we hope to create when we install a coaxial choke at the feed point of an antenna, trying to stop common mode currents? If this is so, how do we explain this idea of CMC flowing down on the radials, and combining constructively with the ½ wave element on top of this transmission line cone setup?

So, your CMC's comments noted above, should had included what stops or mitigates CMC's too. If I'm wrong let me know, but the coaxial cone in this case has one flaw and that is the fly in the ointment. The top of these 1/4 wave elements, and the attached hoop, shows very little current, because this area is at a voltage node, so how are such CMC's produced under such a condition?

Marconi, I realize that you are looking to EZnec to supply some proof of sufficient radiation to assist in the remarkable performance some claim for this antenna. Why it seems lacking to you may be owed to EZnec not correctly supplying enough information on this antenna type. I do not know the program, so I can not say.

Well I know a little about Eznec, but I’m like you, I can’t answer if Eznec is lacking with its limitations. The implication is in the word, so we at least have to consider the cautions and the rules. So, is picking a side to agree or disagree, factual? I say at the very least we have to try and keep an open mind...if it really matters.

We read the manual, and hope we understand it. We hear stories, and some come from folks that know very little about the program and will admit it. We all have read the remarks written right here on WWDX, telling us that 12 or more professional modeler's have stated that Eznec cannot model the Sigma 4, because it does not recognize the radials, and the results in trying would result in the model producing a maximum TOA in the area around 40*- 45* degrees. That is not what I see with my models and I'm surely no expert,

One thing does occur, however, the antenna behaves in the manner that supports the theory behind it.

Homer it is fine to believe that, but what if some evidence was produced that refuted that idea in just one situation, would that be reason to consider or reconsider?

As for your question of why EZnec doesn't show the same currents that are shown by the red and blue being on the opposite sides of the vertical above and below the cone ring, it does. At least that's what I see when I look at the red current line in the EZnec models bowing out in the upper part of the antenna, crossing over the vertical near the ring, and bowing away from the vertical within the cone on the opposite side. Seems remarkably similar to me . . .

If you will look back at my less that clear remarks at that point you will note that I was talking about the red only. I said nothing about the blue. I was just trying to state a point of view about what I was seeing, and I could have said it better.

I think when you can not see something to show you what is happening, then the obvious course is to reason why/what would produce that kind of phenomenon within the known rules of that science, in this case, antenna science. What would cause it, why it happens, and how is it done. In the case of the Sigma/Vector we already have the working antennas, so reverse engineering through analysis should provide answers to a reasoning mind. We have spent lots of time doing this and so far unless the mind is closed we keep looping back to the same conclusions.

You hit it pretty much on the head Homer. I saw something that didn’t jive, based on what I thought I understood from before, so I presented this information for consideration, but all I got was the same old story. It is a stretch for me to expect most to look over all this volume of work for this one antenna, but that is the way the model came out, with lot of data.

I wasn’t successful in developing any interest in a new way of looking at this design, so like I told Shockwave earlier, end of story…unless maybe you get curious as to how we can develop constructive CMC radiation at a high impedance voltage node point. I've always heard that such a situation was a death nail for CMC, if the choking idea used really worked, or am I corn’fused on this too?

I’m trying to keep this simple, and avoid the complexity of a complex idea, but this is about as simple as I can make the words.

Thanks for your comments though, I was just looking for opinions from you guys.

 

[Shockwave]

Are you saying that CST does not manifest the same results as you see when modeling your antenna in Eznec. How did you figure that out?

I'm short on time right now but I'll comment on this important issue. That's exactly what I'm saying. Build a collinear version of the antenna in EZNEC and it will show peak gain at a low angle using a 180 degree or half wave phase delay because EZNEC does not see the constructive radiation in the cone that it will oppose with a half wave delay.

It's clear to me that CST is able to demonstrate this cone radiation in the model. With the cone radiating 90 degrees, we would only need a 90 degree phase shift between sections. That is what will work if you build a prototype. The fact that EZNEC wants to invert the phase a full 180 degrees can only mean it missed the 1/4 wave of radiation on the cone.

Of course Roy assured me EZNEC cannot miss any currents. I'm not the least bit confident in that statement when the working prototype has a electrical length that is 8 feet different than the EZNEC model shows.

 

[Marconi]

Shockwave writes:

I'm short on time right now but I'll comment on this important issue. That's exactly what I'm saying. Build a collinear version of the antenna in EZNEC and it will show peak gain at a low angle using a 180 degree or half wave phase delay because EZNEC does not see the constructive radiation in the cone that it will oppose with a half wave delay.

Marconi writes:

I may have already done this idea for Needle Bender, and I could not get it to work just using a Starduster at the base of a EF 1/2 wave radiator, and a 10T x 4" x 1/2" phasing coil in between the antennas. NB was sure it would work so maybe this is the problem I was running into. On the other hand, I took the SD down and replaced it with an end fed 1/2 wave and the model responded nicely as on might expect with a collinear 1/2 wave setup and a modest ground plane below.

It's clear to me that CST is able to demonstrate this cone radiation in the model. With the cone radiating 90 degrees, we would only need a 90 degree phase shift between sections. That is what will work if you build a prototype. The fact that EZNEC wants to invert the phase a full 180 degrees can only mean it missed the 1/4 wave of radiation on the cone.

Of course Roy assured me EZNEC cannot miss any currents. I'm not the least bit confident in that statement when the working prototype has a electrical length that is 8 feet different than the EZNEC model shows.

The rest of your remarks are above my paygrade.

 

[HomerBB]

Well, an opinion is what yo got from me.

I don't see a high voltage node at the top of the cone on you model(the one you sent me), but as I've clearly said, I am handicapped with not knowing the modeling program. Where I do see it is on the vertical.

Of course, something are different with this cone set up than with a coax, it is attached to the bottom of the vertical radiator, so current on the cone could be supplied from there and not from the top if a high voltage condition is present at the ring.

SW is also saying, it seems, that the cone adds its value to the TOA by preventing the lower 1/4 wave of the antenna from driving the 40° lobe making it possible for the higher gain of the longer antenna to direct toward the horizon.

Now I see a part of dilemma, can we have it both ways or not?

1. Does the bottom cone radiate at all, or sufficiently enough to add gain? or

2. is it simply shielding the negative aspects of the lower 1/4ƛ that drive the 40° secondary so high?

Because you seem to lack evidence of the 1st, the second seems to have greater plausibility, so you want to see some evidence of what is actually going on within the data. Fair enough.

I have not intended to not answer any questions. Some I can not answer. Others I have to the best of my ability.

As for what I think I see in the CST graphic I answered fully. I admitted I don't have any way of knowing what is going on with the graphic except what I trust SW has said of it, and that if I have understood the EZnec model based on what has been explained to me regarding it, then I saw the what appears to be the same demonstrated in the CST model except in a more graphic way.

Trusting any of the models has to go back to two things:

1. that the one modeling with the program has used it competently, and

2. that the maker (programmer) of the software managed to build in the sufficiency of attributes to capture fully the workings of the antenna.

Like you, I am waiting for the outcome of the second question that has arrisen throughout all the discussions of this antenna.

One thing I would like clarification on, not due to your lack of already supplying it, but to my tendency to lose focus of one point or the other in the complexity of this discussion.

Can you once again demonstrate the problem you have expressed right here?

I did that already, and I referred to raising the radials upward as pressing down on the high angled lobe. However on finishing the Vector model, I now see that adding the length to the radials on this Vector seems to cause this idea to fail. Additionally this model shows the radials out of phase with the radiator. SW, this revelation is all new to me.

I know this is prominent part of this thread, yet I am not following it clearly, more than likely due to my work/sleep pattern currently providing gaps in the flow for me, which results in a start and stop kind of thinking.

As for this:

. So, is picking a side to agree or disagree, factual? I say at the very least we have to try and keep an open mind if it matters.

It is why I am following the thread.

When it comes to using an antennaa, I figure build it and fly it, or buy and fly, and if it works for you, be happy. Understanding an antenna is another stoy, and that is why I bother to read these threads.

Another point you make that my experiences have convinced me of:

I believe the only advantage that the 5/8 wave has over the end fed ½ wave is the slight height advantage noted if the antennas are compared at the same feed point height. There is also a height advantage for the maximum TOA that might be noted with the 5/8 wave at some particular heights. Otherwise, I see about the same results from both styled antennas if the height advantage is mitigated out of the comparison.

I recently did a model test of this idea using several different length vertical antennas that showed a nice advantage for the longer radiators, but when I laid all the antennas down horizontal, taking away the height advantage, all the antennas produced gain and angle figures very similar. It was almost unbelievable.


Arguments have yet to convince me otherwise where omnis are concerned. It isn't a popular viewpoint. Why bother with a longer antenna? Because folks are most often limited in the height they can get the bottom of an antenna, so it matters.

What I will likely pursue is an understanding of what height actually maximizes which antenna, or maybe I'll just lay them all on their sides

 

[bob85]

eddie,
one of these two models you posted don't seem right, afaik its not possible for the lower 1/4wave of a straight monopole to be in phase with the upper 1/2wave as depicted in the sigma model,
halfwave down from the tip the phase shifts to the opposite sign as depicted in the vector model, the only way i see the two antennas having opposite phase lower 1/4wave is if you had the phase turned off in current view on the sigma model,
if that is not whats causing your confusion then im as baffled as you are with the sigma model,

Shockwave's .82 Vector.pdf

Sigma 4 Per Manual w mast 36'.pdf

 

[HomerBB]

Uh-Oh

 

[bob85]

i see it now, the current log looks about the same, i missed the hand written note "phase off" on the sigma model

 

[Shockwave]

When I first looked at the CST model about 2 years ago it was clear to me that the currents on the cone were out of phase with the vertical inside it. This is why the radiation from the cone is desirable at the base while the shielded radiator in the center that is out phase, doe's not radiate. I also agree that there is a cancellation affect going on between these two elements. Just that field tests and CST have revealed to me that there is far from complete cancellation here.

All you have to do is model the Sigma in EZNEC using the free space option and you'll quickly discover the model has virtually the same gain as a half wave dipole once the losses are minimized. That tells me EZNEC is only seeing a half wave antenna. Half wave antennas need a 180 degree phase shift to convert them to collinear. EZNEC shows peak gain with the 180 phase shifted collinear, further confirming it's only recognizing a half wave antenna.

There are three significant indicators pointing to flaws in the EZNEC model that make it inaccurate with the Sigma design.

1) Working collinear prototypes of this design require 90 degree phase delay sections. This is quit different then the 180 degrees a 1/2 antenna requires. This is 8 feet of electrical length difference between the EZNEC model and what works in the field. For 90 degrees to work, there must be 3/4 wavelength of constructive radiation taking place below the phase delay.

2) EZNEC only reports about unity gain for the Sigma in free space. CST reports closer to 2dbd and I see about 3dbd when I replace standard 1/2 wave side mounted dipoles with the top mounted design due to elimination of the tower distortion in the pattern.

3) CST shows a second current node radiating from the cone that is in phase with the top 1/2 wave. One can debate about the magnitude of this current but it's not that difficult to see the colors of the top radiator and bottom cone combine together. Two separate current nodes forming one larger pattern as they expand outward.

Marconi, I also wanted to mention I have no clue how you got your model to show current on the cone and the vertical radiator inside it are in the same phase if you had the phase option turned on. Changing the radial length by 10 or 15% doesn't even come close to changing the entire phase.

This design will always have opposing phase between the cone and vertical radiator inside it. Changing the length will change the overlap between how the two currents line up at the loop but not the overall phase. The model is somehow flawed in a way I've not seen before. It appears the currents on the vertical are correct but the cone is inverted.

 

[bob85]

here are some words from w8ji that are backed up by other sources that may shed some light,

"In order to be balanced, a balanced transmission line must have both equal and opposite voltages at any particular point along the line as well as equal and exactly opposite currents at any particular point along the line. If the voltage is not equal and opposite, current cannot remain equal and opposite along the balanced transmission line. This will result in a very large increase in feedline radiation because the imperfection causes common mode currents.


In order to be unbalanced, an unbalanced transmission line must have equal and exactly opposite currents entering and leaving at every point along the line. The voltage gradient laterally along the outside of the transmission line has to be zero. If either the lateral voltage gradient is not zero, or currents entering the line are not equal and opposite on the shield and center, current will not remain zero on the outside of the shield. This will result in common-mode shield currents and feedline radiation."


i don't see the sigma design meeting any of these requirements for a none radiating transmission line, what do you guys think????

if cebik was correct in saying a form of co-linear was possible i don't know of another mechanism that could cause radiation from the sleeve to be in phase with the upper 1/2wave,

at this point in time i think the cst plot shows mr cebik's none apparent co-linear and that its in phase common mode radiation coming from the sleeve that is creating the effect,

i don't know if eznec uses the radiation in its calculations but its not shown in the current log or displayed in current view like i think it is displayed in the cst animation.

im open minded to new ideas, if anybody has an alternative view on the issue.

 

[Shockwave]

You know I agree on these points Bob. There are far to many discrepancies relating to phase and gain between the EZNEC model and the field tests for me to have any faith in much of what the program is reporting on this design. Sure it's easy for those with an opposing view to think I'm just not happy with the results the programs shows.

Would you be happy if EZNEC models produced element lengths that were off by double in the field tests? Would you be happy if EZNEC could not identify a trace of constructive radiation from the cone when phase, gain, CST and field tests all support its existence?

This is the majority of Roy's response when I raised the issue: "EZNEC is not capable of "overlooking" a current node. It calculates the self and mutual impedances of all segments in the model. From that and the specified sources, it calculates and reports the current in each segment. It does not "overlook" any currents." Going on to say the program can't be wrong and the error must be with my model or tests.

Even though he was sent both the EZNEC model and CST model he failed to comment on the secondary cone current in CST and focused in on finding errors in my phasing network length or where I'm claiming the angle of peak gain is occurring with the phase delay used. Give me a break. When you see a $2500 piece of software demonstrate a current node in a design your program has never been able to display, doesn't that make you think there could be a problem? If you question the length of the phasing network in your program, how hard is it for the designer to model a 3 wire folded phasing stub and measure its overall length? If you question the radiation angle the model is targeting, open the total field plot and play smart rather then stupid.

Is it just me or does it seem like many people in business today have a difficult time addressing issues with their products? You have to drag them by the hand step by step, resisting the entire time if you can even get someone to listen. Either way it has taken me a while to compile enough evidence to get to the point where I was willing to throw EZNEC under the bus with this design. I could have debated the issue with Roy some more before posting this but I doubt it would have changed the outcome.