Question on the Vector

[The DB]

DB, I asked you earlier what kind of match you got with the model on page 10. I had to change your model to Meters where it belonged. I did that by leaving the data numbers the same and changing the units. Then I think the model presented without errors. I guess you also set the element diameters to .0001 mm in order to help eliminate segment errors, but this is just a guess, because I have no idea what you were thinking or doing.

The model on page 10 has a match of 82.6 - j96.3, as well as an SWR of 4.22. I haven't modeled a gamma match yet. It is on my list of things to figure out.

I have your model in Eznec now, and it looks like what you might have intended at the time. I posted it compared to what I thought you guys were talking about...adding wires. I was wrong again...that is what happens when we try and guess what others mean.

All I recall you posting earlier was some results that showed the model's gain compared to a dipole and the gain (6.00+dbi) was about what Sirio reported for their NV4K. This was the bases of my concession. I still can't be sure I have your model correct as noted on your computer however. Can you show me a model with an antenna view that shows the gain results you described earlier?

The gain and antenna data can be found in post 172 on page 12 of this thread.

I also think that is where I first saw the model had units set to mm's too.

I figured out that your model's units were really set at Meters and not mm's, but I wasn't sure that I didn't use a model of mine to enter in the data by hand...that was already set using mm's, even though I never use mm or Meters as units in my models. I also try and set my antenna in the CB range so maybe some can easily compare any dimensions noted...to what they might already understand like 102"-108" inches as a 1/4 wave length at CB frequencies.

I'm sure you had some idea about the use of adding a feed line and seeing if it too would show CMC's by adding the wire idea from the Eznec Manual that I posted for Bob a while back. I just imagined such an idea as the wires being parasitic and not a physical part of the antenna.

We can see the results in the pattern you got, it does not look right, and your radiator is almost a wavelength long at 9.1 meters...just like Henry noted earlier. But I can wait for you to post the other results that shows us a good match, gain, and angle from your Vector model with wires added that you, Donald, and Bob came up with.

You can't be seriously thinking that Eznec's...add a feed line feature to a model, mucks the model up like we see here...just trying to show currents flowing on the outside of the cone area.

When it comes to the length of the overall radiator, did you miss the post where I posted that already? I'm not sure what your pattern results show so I can't say one way or the other if they match the results I posted back on page 12 or not.

I will now try your other model posted on page 13.

No DB I won't be posting any more of my results, but I do appreciate your work in this discussion. This is your model to prove by showing all of your results that coincide with the comparative results you showed us earlier...with a pattern of a 1/2 wave dipole compared to your Vector model with wires added showing CMC's flowing on the radials.

I've just been trying to guess what you guys were thinking and doing here, like I did earlier, and I don't wish to get caught up trying to guess or change what may be proving 4Nec2 is able to predict how the S4 design works according to your model.

You got upset with me before in a PM I sent you...when I tried to predict what you may be doing wrong, my asking questions, and me making assertions without really knowing. I don't want to get me and you in that position again. I'm trying to be helpful.

When it comes to that, don't worry about asking questions. What got me to that point was you asked a question and I answered it. You then assumed the answer I gave was wrong and pushed that on me multiple times. You also posted that I was wrong in a public forum, accidently or otherwise. All of this before I posted any real data (plots/currents/phase) relating to the model.

You also seem to ignore me mentioning that certain aspects of the model tell me that this version wasn't a perfect match to the Vector design. I cited the long vertical elements length, the short radial lengths compared to the Vector, among other things as evidence. If anything I have been critical of this version of the model. I have been very critical of this version of the model in several posts, including once stating that it may even be a step in the wrong direction. Call it my model if you wish, but I posted the data for it so that others could see it and see what I did, and what is going on, and play with it themselves. Maybe someone will see something I missed...


The DB

 

[The DB]

DB, I looked at the model descriptions you posted on page 13. The dimensions look to have gotten shorter a bit and there is some new code added to the .txt file that has a line descriptor noted by "LD".

Can you tell me what this LD descriptor means?

Is it part of the antenna or does LD describe something else like a setting?

I can look it up in the manual.

Let me know if I did not make myself clear.

The LD discriptor. I'll parse that out if you like. It is essentially making the elements act like they are t6 aluminum...

The first column after LD appears to be the "wire conductance" option, as opposed to other forms of loading. The second is the tag number. The third column is the first segment to apply this loading to, and the fourth appears to be the conductance, in Siemens, of the material.

One thing that I have noticed is that the main vertical element is not listed with the aluminum "loading" in this final model. I should be, it was initially. I may have to recheck this model again...


The DB

 

[The DB]

DB, I added the new wire descriptions from your .txt file on page 13 in a new Eznec model. It looks similar to the other model on page 10, but it is a little shorter in the radiator at nearly 35' feet instead of 36' feet tall.

The Eznec pattern over real Earth shows the gain, the match, and the currents are all terrible.

However, the Eznec model in Free Space reports an Average Gain = .03 dbi. This is similar to what you indicate, and that is because the feed point is down close to the base of the antenna where the complex impedance is low. This FS models also shows a 4.74 dbi gain at zero "0" angle, which is good, but this is misleading. If we add losses the gain goes to heck in a hand basket. Your model is desperately outside of any reasonable limitations. Sorry, but when i saw the good reports of gain compared to a dipole that you posted earlier, I decided to concede to your better model, but after you gave us some detail...I can't overlook what I see now.

Like Henry suggests...as soon as you added losses to the model it falls apart and now I understand his point.

I set your model to meters instead of mm's, leaving the data values as they were, and that fixed all the segment errors your model showed at first scan.

For the life of me I cannot explain how this 4Nec2 model can show me -65.08 dbi gain using Eznec over real Earth. Yet you claimed good results similar to what Sirio publishes in the specs. Then you posted a good gain and pattern when you compared a 1/2 wave in an overlay of the two patterns.

If my Eznec model of your model is correct, then no wonder the 4Nec2 currents image you showed earlier in the antenna view indicated hardly any currents flowing on the top 1/2 wave radiator. The model is in error.

My Eznec model of your wire descriptions shows a match for R=529300 - J 28720 OHMs, and SWR >100, which is similar to your first model I posted above.

I will also note that in every case before where you or I compared our models...we noted that the results were always very similar.

DB, if I'm wrong on any of this...please work with me to figure this out.

1) Have you figured out what those LD lines represent and implemented that into your model? If not, then your model is off as the entire model has been modified because of the inclusion of that data. Therefore, if you have not implemented the loading data on every wire to your model then this whole post of yours becomes irrelevant.

2) The model that has no currents flowing in the vertical element IS NOT THIS MODEL. That, as I said at least twice above, WAS FROM A VERY EARLY MODEL, ACTUALLY THE INITIAL MODEL BEFORE I MADE MANY CHANGES. The current data you should associate with the data on page 13 is IN THE SAME POST AS THE DATA.

3) THE ONLY REASON THE MODEL ABOVE WITH ALMOST NO CURRENTS WAS MENTIONED IS BECAUSE IT WAS IN RELATION TO SOMETHING I SAW HAPPENING AND AS A DESCRIPTION FOR SOMETHING ELSE ENTIRELY TO WHAT YOU ARE TRYING TO ASSOCIATE IT WITH!!!

4) Eddie, are you actually reading what I am posting?

You seem to be constantly making links between things that don't exist. Further, if you read the associated text it is clear that the links you are making do not exist. You are then being told that the link does not exist, and you continue to keep making statements where you clearly think that the link still exists. It is very frustrating having to tell you time and time and time and time again that something you are stating is wrong, and every time it being about the same thing. You get an idea and don't want to give up on it even when you are told and shown it is wrong, and there nothing to support it.

You need to keep up with what is being said or I'm done. I'm not taking the time to explain all of this just to have what I'm saying apparently be ignored. It isn't worth my time to tell you that a link you made in error is wrong multiple times, especially when reading the original messages the link you made came from clearly shows that it can not be even before you make the link.

I am as clear as I can be in how I say what I am saying, and you seem to be ignoring half of it.

Your references to the currents image above and your link to a model it has nothing to do with, and its text clearly explains not only as much but also where it is from. The explanation shows that it cannot be from the model you are trying to associate it with is a perfect example. I can go on with this example if you wish, this isn't even the half of it.

I have to ask again, are you actually reading what I am posting?


The DB

 

[Marconi]

I am as clear as I can be in how I say what I am saying, and you seem to be ignoring half of it.

Your references to the currents image above and your link to a model it has nothing to do with, and its text clearly explains not only as much but also where it is from. The explanation shows that it cannot be from the model you are trying to associate it with is a perfect example. I can go on with this example if you wish, this isn't even the half of it.

I have to ask again, are you actually reading what I am posting?

DB, to answer your last comment first. I sure hope I read every word you have posted on the recent subject of adding wires to your Vector model hoping to prove there are CMC flowing on the outside of the radials. That was my intentions at least.

I also realize you have explained several times that things in the model you posted on page 10, were not right, but I was having trouble importing the wire descriptions into Eznec, so I could not see or tell what the model looked like.

I even asked Henry to help me understand how to import...after he suggested that he had done so and saw problems with the model. At that point I didn't even understand everything that Henry was describing.

After you reported your success in the beginning with very good numbers, supported by a pattern image comparing a 1/2 wave dipole with your Vector showing very good gain and low angle, I made my concession post. I really figured that you had a model that would show the gain that Sirio publishes in their specs. I was seeing the pattern compared to a dipole...and it was more than I had ever seen in my own work on the S4.

I do however agree with you that I was trying to relate the good numbers you reported with a model that you had not posted as yet. I think if you looked back, I asked you if you could post your Vector model...before you made any modifications for adding wires. I wanted to see what the model looked like before you added the wires.

Then you posted your model on page #10. After I finally got the model into Eznec, where I could manipulate the results and see how the antenna looked, I was very surprised at what I saw, and I said that in my posts above several times.

DB, this is when I started back asking questions. I did not understand why you made a claim that your new 4Nec2 Vector model showed some good numbers similar to what Donald was suggesting, and also showed similar numbers to what Sirio publishes in their specs, yet the model you posted did not show any such numbers after I got it into Eznec.

In my head, I was wondering why would you make a claim and support the claim with such a poor performing model. Maybe I did misunderstand that the model you posted had little to nothing to do with the new Vector model that showed higher gain...like Donald had been telling us.

I still have to ask...why would you post a model that had nothing to do with your previous claim of very good performance? I must have missed this explanation,

 

[The DB]

I still have to ask...why would you post a model that had nothing to do with your previous claim of very good performance? I must have missed this explanation,

Are you talking about the one with almost no currents on the vertical element? If that is the case, the full explanation of the model, why I posted it, and where it came from is in the post itself.

If you are talking about something else then I am not aware of what you are referring to. Both the original, and the later version that was modified to use T6 aluminum, have descent gain over the dipole reference according to 4NEC2.


The DB

 

[bob85]

im lost with this modeling,

i don't understand why anybody would think a good match can be had without a gamma when the impedance at the bottom of the monopole when the radials are added in parallel to the monopole is well below 50 ohms,

nor do i understand how moving the feedpoint up the monopole in the model simulates the braid been connected to the radial/monopole junction with the center tapped up the monopole,

moving the feedpoint up the monopole looks to me like it screws up the antenna mode current distribution in the same way that moving the feedpoint in an ocf dipole would change the current distribution,

i could be wrong but the models with added wires to simulate the outside of the sleeve look to be treating the radials as 1/2wave wires with current maxima in the middle,

what am i missing.

 

[The DB]

im lost with this modeling,

i don't understand why anybody would think a good match can be had without a gamma when the impedance at the bottom of the monopole when the radials are added in parallel to the monopole is well below 50 ohms,

nor do i understand how moving the feedpoint up the monopole in the model simulates the braid been connected to the radial/monopole junction with the center tapped up the monopole,

moving the feedpoint up the monopole looks to me like it screws up the antenna mode current distribution in the same way that moving the feedpoint in an ocf dipole would change the current distribution,

i could be wrong but the models with added wires to simulate the outside of the sleeve look to be treating the radials as 1/2wave wires with current maxima in the middle,

what am i missing.

The model as it exists after a few more edits of the one on page 13, I'm at 119 +j195. This is less than 10:1 SWR before any form of matching network has been applied. This version of the antenna also has 4.72 dBi gain, and if you add in an L-Network for matching, which is to simulate a gamma, still has 4.69 dBi in gain. This is not a gamma, mind you, just the addition of a matching network to simulate some losses for matching purposes. The natural tune isn't so bad that you can't efficiently match the antenna.

Where the feedpoint is does affect the antennas tune, and even a small change in its location has pretty serious effects on the antennas gain. It is between 0.1 and 0.2 wavelengths up from the bottom of the the vertical element. That is where everything comes together for the best performance. Mind you that is the model simulating a standard feedpoint and not a gamma match. Applying a gamma match, which is still something I plan on doing with this model, will be an interesting experience...

The models I have made treat the radials as longer than 1/2 wavelength elements as they go up the inside, down the outside, and then travel on the mast some. On my model I would put them at closer to 3/4 wavelengths total, if not a little more.

Also a note on something I posted yesterday, my segments. The segments in my models are not 10 mm long, they are 100 mm, or 0.1 meters long.


The DB

 

[Marconi]

Are you talking about the one with almost no currents on the vertical element? If that is the case, the full explanation of the model, why I posted it, and where it came from is in the post itself.

I'll go back and re-read the post to be sure, but I do seem to recall you talking about your just trying to consider what the radials were doing and not being particularly concerned about the rest of the antenna. If this is right, then we are more on the same path in understanding.

DB, you might see CMC's flowing in these 1/4 wave radials, but I see an end fed folded 1/2 wave of sorts and Antenna Mode currents flowing. I talk about this below.

Plus this 1/2 wave would show a very bad match without some sort of matching device added since it is an EF 1/2 wave. When I import your model into Eznec, it showed me a very bad match and a lot of negative gain. I wondered why.

Note: I also have my model set for material resistance to aluminum tubing, and it still shows a bad match.

As an example of what I see: back in the days of base station CB antennas there was made a 1/4 wave folded ground plain that resembles what IMO you have constructed as a 1/2 radial wave of sorts. This is just a guess, but this may be why you made the radials shorter than a 1/4 wave in the models on page #10 and #13. It could also be describe as a skinny quad too...as I also suggested in earlier remarks.

You and I were just on different wavelengths in thinking, and I did not pick up on your point of view until we talked. To me, sometimes talking tends to make things clearer.

In the image 4-22a below is a model of a 1/4 folded Ground Plane and I see this in the radial area for the cone. We just see things differently, but we should be able to talk about our differences.

If you are talking about something else then I am not aware of what you are referring to. Both the original, and the later version that was modified to use T6 aluminum, have descent gain over the dipole reference according to 4NEC2.

If I'm even close in my comments above...then I don't think I was talking about something else. This also might help explain some of my corn-fusion. I was expecting that maybe the model you posted for me to import was going to show us the nice gain you reported earlier, but I was wrong obviously and I pray you understand I was not trying to be mean.

After you posted the patterns for the Vector and the dipole showing us the nice gain, I asked you if you could post the original model, before you modified it with the added wires. Do you recall that?

When I was finally able to get your wire descriptions into Eznec...I did not see the nice gains, and the model also had the added wires you were telling us about. I was thinking, if your model showed a nice gain that I was not able to get with my Eznec model...I wanted to see your Vector model that you started with...and compare it to my Eznec model. That was my frame of mind, not something else. I appreciated the fact you finally might have the answer...how the S4 designed worked, and that you were seeing better gain.

Can you show me the antenna view, pattern, gain, SWR/match reports from 4Nec2 that shows the nice gain for your Vector? I can't explain why your starting Vector model would be so different than mine...others I have compared to your models in the past were always very close.

Don't get upset with me DB. If you are right and you have a model that shows a superior gain like Donald and Siro report for the NV4K, then I would like to try and understand more about the model, and see if I could do the same with my Eznec model.

Thanks.

 

[Marconi]

im lost with this modeling,

i don't understand why anybody would think a good match can be had without a gamma when the impedance at the bottom of the monopole when the radials are added in parallel to the monopole is well below 50 ohms,

nor do i understand how moving the feedpoint up the monopole in the model simulates the braid been connected to the radial/monopole junction with the center tapped up the monopole,

moving the feedpoint up the monopole looks to me like it screws up the antenna mode current distribution in the same way that moving the feedpoint in an ocf dipole would change the current distribution,

i could be wrong but the models with added wires to simulate the outside of the sleeve look to be treating the radials as 1/2wave wires with current maxima in the middle,

what am i missing.

Well Bob, maybe this will help.

I see the same and alluded to the radials looking like a folded 1/2 wave dipole in my post above this one.

Here is the match and other data that I see with DB's model that I imported into Eznec by hand. DB, told me it was alright to post my data. Sorry I did not have the currents on.

I also posted the wire definitions so folks can compare my numbers with the model posted on page 13. I show the antenna, pattern, wires, and the feed point source data.

I can't disagree with DB's explanation above either. I was typing my response while he posted. IMO, he tells us the model was designed to only show what was going on in the cone area and I think he told us that, I understand him better now, and I believe him.

Maybe that is what he was suggesting in the very beginning when he explained why he was uing so many segments in his model. I just didn't connect all of his dots.

 

[Marconi]

The model as it exists after a few more edits of the one on page 13, I'm at 119 +j195. This is less than 10:1 SWR before any form of matching network has been applied. This version of the antenna also has 4.72 dBi gain, and if you add in an L-Network for matching, which is to simulate a gamma, still has 4.69 dBi in gain. This is not a gamma, mind you, just the addition of a matching network to simulate some losses for matching purposes. The natural tune isn't so bad that you can't efficiently match the antenna.

DB, my Eznec model shows a much more terrible match than your model. I think I've tried moving the feed point tap from the base up to about 35" above and saw very little change in the match. I can't explain why we see this difference between our models, but it could all be due to your adding wires that are connected to the radials instead of wires that are passively attached like the directors on a bean antenna.

My S4 model shows the area near the bottom of the monopole is very low impedance and as I raise the tap point up the radiator the resistive part goes higher until I can find a 50 ohm match at about 32" inches. My model does not show the sensitivity you note in the area around the bottom of the monopole. I don't think my real antenna shows a big difference in this area either.

So much so, that I have been tempted to see if I could tap the bottom of my S4 and direct feed the antenna.

I hope this makes sense.

 

[Marconi]

The LD discriptor. I'll parse that out if you like. It is essentially making the elements act like they are t6 aluminum...

The first column after LD appears to be the "wire conductance" option, as opposed to other forms of loading. The second is the tag number. The third column is the first segment to apply this loading to, and the fourth appears to be the conductance, in Siemens, of the material.

One thing that I have noticed is that the main vertical element is not listed with the aluminum "loading" in this final model. I should be, it was initially. I may have to recheck this model again...

The DB

I suggested in my question the LD might be the code for a switch.

This function using Eznec is a switch with several selections like copper, tin, aluminum with several qualities to select, and an option to custom select another material resistance if you know the values including "None."

Eznec selection for this feature, "Material Resistance" is one of those listed or none. None is used to remove the losses necessary to accurately run the Average Gain Report.

As a reminder, this is the area I was addressing earlier when I thought you got upset with my questions.

Thanks.

 

[bob85]

DB
if im understanding your model with the near 1 wavelength monopole the antenna mode impedance of a 1 wavelength monopole is very high so little current flows in the monopole above the sleeve as described in the arrl article which is what your model shows,

what i don't get is it also shows very little current in the lower portion of monopole that forms the center conductor of the transmission-line,

its not important that i understand the model but it would be nice if i did,

i wish you good luck with adding the gamma to the model,

you will be introducing a 3rd pair of currents into the mix with antenna mode and transmission-line mode currents flowing within the gamma section as i mentioned years ago.

 

[Shockwave]

Marconi, your folded 1/4 wave groundplane example may represent a similar physical and electrical circuit, however it in no way accounts for the PAIR of currents present on the cone that form the 1/4 wave pattern with a current peak towards the center. You also cannot expect a parasitic element to represent the phase we see on the cone. The phase of parasitic elements mimics the phase of the driven element in your example.

You recognize other end fed antennas would have CMC issues if it were not for the addition of decoupling radials. Why would you think the half wave end fed portion of this antenna did not develop CMC of if you think it does, why wouldn't they be present on the resonant cone?

One you understand that they are present and their path is from the base of the inside of the cone and fold over the top of the cone onto the outside surface, you understand how that current is 1/2 wavelength long and the phase shift between the bottom of the inside and the bottom of the outside is 180 degrees.

You would also understand why it is important for DB's model to show the extra wires connected to the top of the cone since that mimics the current flow and phase much better than a simple parasitic element.

 

[bob85]

eddie,
you have suggested that the currents in the radials are antenna currents not cmc,
they are two names for the same thing eddie,

the other currents i imagine flowing in the vector are transmission-line or differential mode currents.

 

[Marconi]

Bob, you will be draging me out of my commitment to concede my position on the S4 design until DB has finished his project that shows us more...regarding the improved gain he has reported in his thread, "Question on the Vector".

I agree with what you suggest, but I think of CMC only exist on the outer shield of coax feed lines, and not the out-side of these antenna elements as noted on the S4 cone. Do CMC's act like the currents on the antenna elements...yes the can.

Adding a wire directly to the antenna can change everything for the model, and adding a feed line by adding a wire so we can see the currents as I have modeled the idea...does not alter the model much if at all. If you better understood what this feed line feature does and how...I think you might agree with me on its use.

I think this idea of adding wires is mistakenly applied, but I'm waiting for DB to show us more, and I want to see the performance reports for his model that shows the improved gain he has reported.

I can agree with you except....I'll explain more later if I can remember.

I also think there are transmission line currents flowing between the radials and the out of phase monopole. This happens, because these currents are virtually equal in magnitude and opposite in phase...according to what I see with my Eznec model of the S4.

These currents are a bit different on my Vector model, because I can not fix the NEC requirement to always connect, wire 1 end 1 to wire 1 end 2, and visa-versa with every wire in the model. The 3 radial S4 allows me to fix this issue completely, but the 4 radial Vector does not, even so the differences are insignificant at best.

On this very point I might consider that Cebik knew this when you two talked, and this was what he considered to be a dubious effort...that might make such a model difficult to impossible. Considering that you don't know for sure Bob, my guess might be as good as yours.