ASTROPLANE best vertical antenna ever?

[bob85]

i have thought about what cebik meant when he told me it would be very difficult to get accurate results for the vector using eznec,
he did not say impossible or give me an idea of what magnitude of error we can expect or why

i don't use software so im just sharing my thoughts from what i have read about nec,

if i remember correctly nec samples in the middle of each segment, which means the sample is not at the very end of a conductor,

could that cause the model to not be showing the real difference in magnitude between the top of the sleeve and the corresponding point on the monopole?,

if it does work like that how much difference is there?
is it enough to stop eddie seeing the claimed free space gain in his models or not?

also in the real antenna there is significant difference in the diameter of the monopole to radial conductors,
how does nec handle that if you try to model the real antenna?


what i don't understand is when you lengthen the monopole to more than 1/2wave the corresponding point on the monopole becomes a lower impedance point in my minds eye,

the magnitude of differential should increase which should increase the magnitude of radiation from the sleeve, does nec show that to be true?
if not what am i missing?

when you add extra conductors to a j-pole to make a skeleton sleeve monopole does the bit of extra gain in the direction of the short leg of a wider spaced j-pole go away in an averaging type function or do we end up with the same peak gain but omnidirectional?

it seems to me that it makes sense to use as many small equal length segments in the radials and monopole as possible to minimise the
disparity between sampling points and the real end of the sleeve and its corresponding point on the monopole,

maybe two models of the same size vector, one with a low number of equal length segments and one with as many smaller segments as possible would confirm or refute my idea.

 

[Shockwave]

Donald I'm just not convinced that you are right yet, and I don't understand your point in proving your ideas based on my adding a 4 wire collinear 1/2 wave above a S4. I wish I could understand what you have in your minds eye.

If I did do another model like you describe and it didn't produce the results you have in mind you would just tell me I had it wrong. What is the future in doing that?

I work this way. If I have a model that I think can help prove my idea, then I post the model and then we have something to consider. Me just guessing exactly what you have in mind is not productive.

I get no joy in disagreeing with either you are Bob, but I do have my opinions. I'm waiting for someone to question, as wrong or error, the model ideas or words that I just posted. I knew you would not check it out.

Thanks for the offer though.

I'm disappointed to admit that I have lost much hope in getting you to understand or accept how the Sigma works. I've moved my focus beyond this point and am now looking to prove that the common software modeling programs are absolutely wrong and remove the possibility of some other issue. That perhaps to date, CST may be the ONLY software capable of modeling this design accurately.

One example is your noticing where the wire end and start points are orientated. I was willing to give your thought some merit if you could show that reversing any wire could make any of your models work with the field tested 90 degree collinear phase shift. Understand where my mind is at in this area now?

 

[bob85]

interesting article on how nec handles phasing stubs

An exploration of a cophased collinear array with coax phasing stubs

 

[Marconi]

i have thought about what cebik meant when he told me it would be very difficult to get accurate results for the vector using eznec,
he did not say impossible or give me an idea of what magnitude of error we can expect or why

Bob, I only have an opinion as to possibly why Eznec might cause some concern in the process of modeling the S4/NV4, and it has much to do with how wire ends are handled at the input of the data, and the results we see at the output...in particular the currents and phase issues noted in the tabular currents log.

Nobody I've talked to about modeling has ever seriously addressed the issue of currents and phase with me...even when I try to talk about it. So, I can't be totally sure my thinking is right on the subject, but I have my ideas and I see issues when I do things differently in my modeling.

I was hopeful the other day that ghz24 was going to possibly provide me with some insight in his dealing with 4nec2. He came back with a no answer answer, so I pressed him to be specific. I did not want to say anything that described what I was seeing before he suggested to me what he saw. That all was to no avail and I've not heard from him since. Excepting he said he switched ends in a simple L shaped dipole, posted same, and said that doing so it mattered, without being specific. He did post is models and I checked one of them, but I'm not efficient at working that software and I did not take the time to start without getting more form him. He told me he works a lot, so he doesn't have much time. I forget too much at my age to be very productive that way, so I carry on. BTW, he did tell me that his switching end in his model did matter and that it changed the gain and the angle. I knew immediately that is absolutely not what I see here, and I just blew the idea off.

If I say more here I'm afraid I'll influence him in my thinking and then the cat is out of the bag. Plus we are working on two platforms and I don't know his product well, and he doesn't know Eznec well. That to could be a problem, even though his model results thus far have been very close to mine.

If I see what looks to muddle the end product in the currents area while the antenna gain, angle, swr, and match all remain exactly the same for the model...then maybe Cebik realized that would likely be an issue and may be difficult to understand a topic that very few ever get involved with...currents and phase, and we both suspect that this antenna depends some how on how the current work on raising the radials. I just posted something new today in my words to Donald, but that too is lost to the Ether.

i don't use software so im just sharing my thoughts from what i have read about nec,

if i remember correctly nec samples in the middle of each segment, which means the sample is not at the very end of a conductor,

could that cause the model to not be showing the real difference in magnitude between the top of the sleeve and the corresponding point on the monopole?,

if it does work like that how much difference is there?
is it enough to stop eddie seeing the claimed free space gain in his models or not?

Bob, could you be more specific on what you see when you say the "real difference.....?" Are you seeing something that I posted? Are you talking about the point on the model where the red line current indicator crosses the radiator relative to the top of the area at the hoop?

I have said already that I cannot get these models to work right trying to add tapper to the radiator and the radials.

I have been working with a new model that looks promising with a little improvement I think. With this one a tapered the first element that is 71.5" inside of the cone. I made it the stock diameter because it is so long...I thought it might need to be there and took a chance that one tapper connection on the radiator would not be so destructive to the results as tapering all the tubing proved to be.

The big problem I will have is being able to ever produce a vertical S4 that has a free space gain much over the 2.29dbi gain at 18* degrees much less a 4.15dbi gain in FS like Donald suggests is what the New Vector shows. I'll probably also find out someday from Donald that the NV4 model in CST also shows this gain of 4.15 at 0* to the horizon too. All of my CB vertical models over 1/2 wavelength show FS models at higher angles above 0* degrees. The FS gain may be higher, but the angle is always about 3* degrees and higher...all except for the center fed Gain Master. Hooray for symmetrical center fed antennas.

I agree with Donald on one thing though...this fact note above sure looks like Eznec does not even consider any far field gain except in the 1/2 wave part of the antenna. I see the same thing with my .625/.64 wave models, and I know for sure they don't work based on any collinear benefit. I'm talking only Free Space models now.

also in the real antenna there is significant difference in the diameter of the monopole to radial conductors,
how does nec handle that if you try to model the real antenna?

For the most part I always try to keep the antenna to specs, but I have to make an educated guess and average the longer radiators so Eznec does not get all unset.

what i don't understand is when you lengthen the monopole to more than 1/2wave the corresponding point on the monopole becomes a lower impedance point in my minds eye,

the magnitude of differential should increase which should increase the magnitude of radiation from the sleeve, does nec show that to be true?
if not what am i missing?

I have not done iterations for the difference in radiator lengths, but I did do one on the .75 wave model with 107" radials a while back. I was reading some old discussion back in the day when you, FC, 228, and Master Chief were fussin' and discussin' the j-pole and this style antenna. I think ever body was talking about how and what changed the values of R, X, gain, angle, bandwidth in mhz, at the bottom and higher up the radiator when using a gamma match. I don't know if you can see or read this, but here it is.

View attachment Iterations of radial angle.pdf

when you add extra conductors to a j-pole to make a skeleton sleeve monopole does the bit of extra gain in the direction of the short leg of a wider spaced j-pole go away in an averaging type function or do we end up with the same peak gain but omnidirectional?

it seems to me that it makes sense to use as many small equal length segments in the radials and monopole as possible to minimise the
disparity between sampling points and the real end of the sleeve and its corresponding point on the monopole,

maybe two models of the same size vector, one with a low number of equal length segments and one with as many smaller segments as possible would confirm or refute my idea.

I'll get to the last section tomorrow maybe. Thanks for the questions and ideas.

 

[bob85]

i did not realise you were that far away from 4.15Dbi eddie,

i meant that where eznec samples the impedance is not at the radial tip where impedance is highest,
impedance disparity at the top of the sleeve is less than it would be if the sample was taken at the tip of the radial where impedance is VERY high but not infinite due to the diameter of the conductor,

am i missing something about how eznec works?.

 

[Lil'Yeshua]

The thread title implies what?

 

[Marconi]

i did not realise you were that far away from 4.15Dbi eddie,

i meant that where eznec samples the impedance is not at the radial tip where impedance is highest,
impedance disparity at the top of the sleeve is less than it would be if the sample was taken at the tip of the radial where impedance is VERY high but not infinite due to the diameter of the conductor,

am i missing something about how eznec works?.

I can't tell you how Eznec works at the program level. I only have a skimpy idea how this stuff works. I do understand how to get around in the environment however and I have an idea how to use some tools.

I mentioned in my post to Donald that the .75 wave Demo model I used was very abbreviated with the segment count used. I did that so I could present a smaller picture of the currents...and not blow everyone's mind with a lot of confusing details that most likely won't consider or try and understand.

The actual model I made for the .75 wave has 384 segments on 10 wires and my real S4/Vector models with a loop have close to the 500 segment limit.

The Demo model has only 36 segments on the same 10 wires, so of course the data detail for the current segment distribution across each wire has its effect on the results we see at the 1st and last segments. When I reduced the segments in the .75 wave model to make the Demo model...the gain did go down some from 4.00 dbi at 8* degrees to 3.81 dbi at 8* degrees over real average Earth.

I have not analyzed the idea that I think you have in mind here, but I can tell you that the idea of your question has merit, because the number of segments does effect the results to a point...at both ends of the range.

I found some models produced by Cebik with some modeling detail sometime back, and I was able to duplicate the model very close to get his reported results. I wanted to try and understand just what he did and then maybe try and figure out why. I used trial and error logic.

I then looked at his model to see if I could determine any idea of a scheme he might have used. At that time I saw him making each segment as close to the same length for all wires.

However, since then I've seen him also use very small segments lengths on the primary wire near the feed point in a model...and then follow the prior idea with the rest of the wires.

If you've read anything about CST, do you know if their product uses MOM's procedures in their engine or does it use something entirely different?

If you think the fully segmented model will help you see what I see, then I'll email it to you. Just let me know and I'll also give you the end to end switch fix that I used for this model.

 

[Henry HPSD]

Marconi.

QUOTE:
Nobody I've talked to about modeling has ever seriously addressed the issue of currents and phase with me...even when I try to talk about it.
END OF QUOTE.

Thougth i gave it a good try, must have been mistaking ?

In other words, im "*" that my effort didn't help.
This of course is my fault as im probarbly unable to explain it correct.

Not to be an "x" but In aspect to many of your questions: use google or the help file.

Google: cst & mom .....lets see if you get some hits.
Yes, cst does have a mom option but also fem etc.

As i believe i have explained earlier:
Before we get a "see its probarbly the mom versus another methode thingy"
You can guess, what the guys at CST say to use. (i hope the hint is clear hihi)

In aspect to 4nec2 and questions like:
average gain function, radials etc.... all described in the manual......
Press the second word in the left top corner... "help" it says.

In aspect to the "difficulties" to model a sigma 4 story.
That has nothing to do with input output etc.

The difficulties lies in the fact we have CLOSED SPACED WIRES.
Besides those CLOSED SPACED WIRES are moving AWAY from each other under a ANGLE.
Again....somthing that can be found in the manual.
That is something the engine doesnt appricate.
With that said, this is not a "not done" thing..but you need to be aware of it..and know what you are doing.

questions regarding the number of segmentations (and how to verify them) etc...all in the manual.

Now, with your permission..i just wont continu over the forum.
As last time didnt help.
(And in all honesty, one of your posts takes for me very long to "sink in".)
Migth be me not beeing a native at ur langue.

So, an idea:
Call me.... !
Im at UTC +1. Just take a "normal" hour...not beeing 0200 hihi.
We can both sit down behind a computer and talk through 4nec2 and eznec.
Maybe that can help in your phase current search...

Kind regards,

H.

 

[bob85]

yes eddie im aware cst uses mom, i had a look at antenna magus in the hope i would find a similar antenna for some clues,
i still get emails inviting me to lectures and webinars from webinars@cst.com, next one nov14 "EMC Simulation in the Design Flow of Modern Electronics",

the guys at cst seem like a very professional bunch, i won't pester them with cb antenna questions.


henry,
thanks for your input, no problems here understanding you

some years ago i posted that eznec can have some issues with
close spaced wires at acute angles and dissimilar diameter conductors & tapers,

right or wrong on my part, what cebik told me and what i read made me think an inexperienced guy like myself without a full understanding of the issues would likely not get accurate results trying to model the antenna exactly as its constructed in real life so i never had the confidence to try.

 

[Marconi]

Marconi.

QUOTE:
Nobody I've talked to about modeling has ever seriously addressed the issue of currents and phase with me...even when I try to talk about it.
END OF QUOTE.

Thougth i gave it a good try, must have been mistaking ?

In other words, im "*" that my effort didn't help.
This of course is my fault as im probarbly unable to explain it correct.

Not to be an "x" but In aspect to many of your questions: use google or the help file.

Google: cst & mom .....lets see if you get some hits.
Yes, cst does have a mom option but also fem etc.

As i believe i have explained earlier:
Before we get a "see its probarbly the mom versus another methode thingy"
You can guess, what the guys at CST say to use. (i hope the hint is clear hihi)

In aspect to 4nec2 and questions like:
average gain function, radials etc.... all described in the manual......
Press the second word in the left top corner... "help" it says.

In aspect to the "difficulties" to model a sigma 4 story.
That has nothing to do with input output etc.

The difficulties lies in the fact we have CLOSED SPACED WIRES.
Besides those CLOSED SPACED WIRES are moving AWAY from each other under a ANGLE.
Again....somthing that can be found in the manual.
That is something the engine doesnt appricate.
With that said, this is not a "not done" thing..but you need to be aware of it..and know what you are doing.

questions regarding the number of segmentations (and how to verify them) etc...all in the manual.

Now, with your permission..i just wont continu over the forum.
As last time didnt help.
(And in all honesty, one of your posts takes for me very long to "sink in".)
Migth be me not beeing a native at ur langue.

So, an idea:
Call me.... !
Im at UTC +1. Just take a "normal" hour...not beeing 0200 hihi.
We can both sit down behind a computer and talk through 4nec2 and eznec.
Maybe that can help in your phase current search...

Kind regards,

H.

Sorry Henry.

I did not want to go into all the back and forth details trying to explain the results of your trying to help me better understand currents. Donald could say the same thing, because we discussed how Eznec handles currents differently when I connect wire 1 end 1 to wire 2 end 1 vs. connecting wire 1 end 1 to wire 2 end 2. I also tried to talked to Roy about currents and that too went nowhere.

So when I hear it said that Eznec cannot model the S4 design correctly, I say please explain why I see the same thing happen with the FS models for my .625 and .64 waves. My Eznec models will not produce the gains reported by Sirio for any of their antennas longer than 1/2 wave either, so I tend to question such statements and I relize that some folks think I'm just hard headed and maybe a little nuts.

Henry you talk about close wires as being a potential problem. I agree that Eznec does not like wires at sharp angles and that get too close together. I have seen the warnings many times.

I'm probably wrong, but at times I've found, in some cases, where a work around for this problem is possible, just by adjusting the wire diameter and the segment count a little. That said however, the inital problem of wires that are at sharp angles and too close may still exist even with a work around solution...that I don't know for sure.

When you added the radials to the monopole in your S4/NV4 model did you first build a small hub on the monopole to accommodate and offset for the radial ends on the radiator like the real models have, or did you just connect the radials directly to the monopole?

 

[Marconi]

yes eddie im aware cst uses mom, i had a look at antenna magus in the hope i would find a similar antenna for some clues,
i still get emails inviting me to lectures and webinars from webinars@cst.com, next one nov14 "EMC Simulation in the Design Flow of Modern Electronics",

the guys at cst seem like a very professional bunch, i won't pester them with cb antenna questions.


henry,
thanks for your input, no problems here understanding you

some years ago i posted that eznec can have some issues with
close spaced wires at acute angles and dissimilar diameter conductors & tapers,

right or wrong on my part, what cebik told me and what i read made me think an inexperienced guy like myself without a full understanding of the issues would likely not get accurate results trying to model the antenna exactly as its constructed in real life so i never had the confidence to try.

I got my notice today. I've yet to join in on one of these CST meetings however.

If I had been Cebik talking to you about Eznec, I probably would have made comments about limitations for the ideas you presented...but I would have also tried and encouraged you to pursue your ideas.

I read a review on a ham site or a webpage a while back and the guy was giving Cebik praise for his work and contributions in radio and many other fields of endeavor. He said specifically that Cebic basically got his start in the world of ideas by promoting the new tools available to the masses... antenna modeling using PC versions of MoM's algorithms, and that he was prolific in writing on the subject.

I had the thought to send you one more shot at trying to explain what I see in the currents with Eznec, but I don't think it will be worth the effort.

Sorry, but I'm tired of trying.

 

[Marconi]

when you add extra conductors to a j-pole to make a skeleton sleeve monopole does the bit of extra gain in the direction of the short leg of a wider spaced j-pole go away in an averaging type function or do we end up with the same peak gain but omnidirectional?

IMO, a lot depends on the space between the bottom elements. I've not model the skeleton sleeve, but from the looks of it I'd guess the radials need to be within an inch or less to the radiator, like the S4/NV4 and my J-Pole models that look to work right. Wide spacing for a J-pole is no-way-no IMO.

I didn't check the actual dimension, but the radials look to be offset from the radiator by several inches in the Open Sleeve model in the article, and if they were closer together we might see more gain...the same with the J-Pole. Again, the closer the better is what I think.

it seems to me that it makes sense to use as many small equal length segments in the radials and monopole as possible to minimize the disparity between sampling points and the real end of the sleeve and its corresponding point on the monopole,

And to a point, I think you're right. With only 500 segments allowed I typically try to make my segment length 2.5" to 3.0", and I get them all as close to the same length as possible. However I recently saw a Cebik model that showed the wires view and he used the sameO'/sameO' idea on all wires except the radiator. I noticed when I raised the feed point the Average Gain changed rather dramatically, so he was trying to get the source as close to the bottom as possible and that made the difference.

maybe two models of the same size vector, one with a low number of equal length segments and one with as many smaller segments as possible would confirm or refute my idea.

I did that with the .75 wave model I just posted. I talked about doing that already in my post to you, and I gave you the number differences so you could get a sense of modest difference it made. I did not check the differences in currents, but they probably changed more dramatically and it's not hard to figure why.

Below are two reports in my notes. One was done back when I talked to you about the 3.2:1 ratio limit in the Open Sleeve article where you said I misunderstood the math back a year ago. The other that I did today is pretty similar as it turns out and is the one I promised you I would do. You can compare some data but the one I did a year ago was over real Earth and the one I did today was in Free Space where I removed the mast and the losses.

I don't recall if I ever posted the older report, because I seem to remember you and Donald were discounting any possibility of such a report telling anybody anything concrete and that I did not understand what I was reading in the article.

I didn't go looking in my notes for this old work...it just so happened to fall out of my notebook when I picked it up to put into my printer. Go figure that just happening out of thin air.

View attachment Bob's idea for S4 change in length..pdf

Sorry, the last report must have flipped over in my printer for some reason. You'll have to print it out if you care to look it over.

 

[SmackDown]

"Best" is such a subjective word. I can say that with my A/P 30' off my 36' high home I have made contacts on 10 & 11mtrs across the country with good audio & signal reports and lately been enjoying a good deal of dxing across the Atlantic to 387 and 347 in Manchester UK on the bowl. I also had a weak contact with 457 in Belgium from my place in the Phila. suburbs. Great antenna without a doubt.

3's

525 philly aka the unknown soldier

 

[Marconi]

"Best" is such a subjective word. I can say that with my A/P 30' off my 36' high home I have made contacts on 10 & 11mtrs across the country with good audio & signal reports and lately been enjoying a good deal of dxing across the Atlantic to 387 and 347 in Manchester UK on the bowl. I also had a weak contact with 457 in Belgium from my place in the Phila. suburbs. Great antenna without a doubt.

3's

525 philly aka the unknown soldier

SD, the very first A/P I ever tried was an all copper homemade antenna I bought from a friend who built them many years ago. I bought it for another buddy and he used it until he died a while back. Both of those old men talked all around Texas for years on 10 and 11 meters.

I have a Sirio New Top One and several older models. I can't see a shred of difference when I get those two up and side by side. The New one is much better built, and has a tad less bandwidth.

 

[nosepc]

I would argue that the antenna A / P needs a diameter similar to the mast to function properly.
Smaller diameters makes more work flows on the mast and not the antenna.
It is my humble opinion.

op:

no see.