Marconi's Eznec model for the AS Super Magnum 1/2 wave GP

[Marconi]

Here is my model of a Super Mag (SM) 1/2 wave vertical ground plane antenna in Free Space (FS) only, and some of my thoughts for this antenna design. If I get this model working with a good Average Gain result in Free Space...then I will set the model over Real Earth and post it later.

The (SM) model is setup in (FS) and the other model is a vertical 1/2 wave dipole in (FS) which I post for comparison. These models were done in (FS) in order to minimize the effects of Earth and material losses.

IMO the (SM) uses a similar design to the Old CLR2 and Super Penetrator 500. These were each unique because of the fact the mounting bracket design raises the GP radials up about 12" above the bottom of the radiator.

When I finished this model over real Earth, the model showed more gain than most of my other 1/2 wave and 5/8 wave models...and that seemed to me to be problematic. So I concentrated on making the model in (FS).

To start here is the (SM) model in (FS). Notice the (AG) at the bottom of the main screen indicates a nearly perfect match of the model.

Eznec Main Screen for the Super Mag.


Antenna View.


Antenna (FS) Pattern.


Wire descriptions.


Source (FP) data sheet.


Below is a vertical center fed 1/2 wave dipole in (FS). Use this to compare the gain, and angle of the (SM) model noted in the Antenna Pattern. You will notice the 1/2 wave shows a little bit more gain and the Average Gain values are almost identical.

Eznec Main Screen for the CFHW dipole.


Antenna View.


Antenna (FS) Pattern.


Wires Description.


Source (FS) data sheet.


Note that the CFHW dipole's Source Data reports the generally accepted impedance values for the dipole as read about in science, which IMO suggest the model is good enough for Government work.

My conclusions. This (SM - FS) model is fine to place over Real Earth conditions and then we'll see what happens. Before I did this project, my fist model of the (SM) over real Earth showed a, too good to be true, a with very good gain at a good angle for a 36' foot installation. I hope getting the (FS) model working right before I do the Real Earth model...works out well.

From this 1/2 wave model I will also set the model up as a Super Penetrator model, and then we can see what Eznec shows us in performance for both at 36' feet.

If I find errors these models are subject to change.

 

[Marconi]

Here is the Super Mag over real Earth at 36' feet to the feed point.

Main Screen.


Antenna View


Antenna Pattern.


Antenna wires descriptions.


Source data/Feed Point matching


Note: after I did this model in Free Space and got the Average Gain showing a very low value as noted at the bottom of the Main Screen above...I reset the model over Eznec's ideas for Real Earth, and this is the results.

When I did this model a while back, from scratch, and over Real Earth I did not test the Average Gain with a (FS) versions. Back then the model showed gain over 5.00 dbi. This is why I did not post the model, because I realized this was not the gain results this EFHW ground plane should show. It was just too much gain for a 1/2 wave GP.

I talked to DB the other say and he told me he was testing his models for Average Gain before he set the model over real Earth...and I agree that is the right way to build a real world model if you want to be accurate.

This said however, I noticed several things in the model below that I did not anticipate, and only noticed after I completed this project.

1. The model is very sensitive to the placement of the feed point (Source) as expected, but I was not aware that the number of segments assigned to your model is also very sensitive...if I follow the ideas noted above.
2. I also noticed that DB seems to design his models using very thin diameter wires throughout the model as though the antenna is a traditional wire antenna vs. using aluminum tubing for example.
3. DB's tells us he also uses a 4Nec2 design feature called Automatic Segmentation and I have not fully checked out segmentation to see how this all works with a model.
4. So, the model above may change, because I set it up with all the wire except for the radiator with the same diameter.
5. I knew that deciding the segment count for each wire was important for accuracy, but I'm thinking now proper segmentation is more important than I first considered.

 

[The DB]

I talked to DB the other say and he told me he was testing his models for Average Gain before he set the model over real Earth...and I agree that is the right way to build a real world model if you want to be accurate.

Actually, I can run AGT over earth. Nec2 sets the ground to "Perfect" for the AGT test.

1. The model is very sensitive to the placement of the feed point (Source) as expected, but I was not aware that the number of segments assigned to your model is also very sensitive...if I follow the ideas noted above.

I made the free space version of the model in 4Nec2 and noticed the same thing. If I put the feed point at the base of the radiator where I would expect it to be, AGT goes above 1.2, which is saying the model is not even close to accurate. I don't like putting the feed point at that location, but I guess that is a limit of the modeling software based on that design.

2. I also noticed that DB seems to design his models using very thin diameter wires throughout the model as though the antenna is a traditional wire antenna vs. using aluminum tubing for example.

More recently I have been using a diameter just under 1/4 inch, although I was thinking it was just under 1/2 inch. That was a mistake on my part. Remember, most of my models are based on a meter, not an inch. There is a rather large size difference I need to account for, which translates to a much smaller part of a meter than an inch for an equivalent size, so at least at first glance said diameters appear to be very small.

Also a note on wire diameter I've noticed. Thinner diameter wires often give a better AGT result. Just something to take note of.

5. I knew that deciding the segment count for each wire was important for accuracy, but I'm thinking now proper segmentation is more important than I first considered.

One change I would recommend trying is making the area of the main radiator inside the "basket" area to its own wire, and size it's length and it's segments to match wired 14 and 15. It might not make much of a difference, but when wires are close and line up like that I've noticed in the past that where the segments are can sometimes make a difference. This isn't always the case, but sometimes it does make a difference. If you are going to play with segment length, I would make this one modification.

One other note on that model, 4Nec2 gives me warnings for wires 2, 3, 4, 5, 10, 11, 12, 13, 16, 17, 18, and 19. It is basically saying the segment length divided by the radius is below 6. This is a warning, not an error, it would be an error if said calculation was below 2. What the warning is telling us is that the model might have a problem, not necessarily that there is a problem.

There are three ways to fix this normally, one would be to use fewer segments in the wires, however they all have 1 segments each. Another option would be to make said wires longer, but that would affect the model. In 4Nec2 I would go with the remaining option, and that is make the wires thinner. Not necessarily a lot thinner, but just enough to make the warnings go away.

Aside from these, my reproduction of the free space model's results come very close to your free space model's results.


The DB

 

[Marconi]

Actually, I can run AGT over earth. Nec2 sets the ground to "Perfect" for the AGT test.

My Eznec manual on AGT is attached in the PDF file below.

I won't try and interpret the meaning, but I don't see how the AGT over real Earth can produce viable information...unless the software is able to determine all the material and Earth losses in its models. Eznec won't do that. The modeler has to physically turn off the losses due to ground type as well as the material resistance losses in the structure.

DB, if you can get a AGT value = 1.00 for a free space model, are you still able to get a similar low value = 1.00 if your model is set over real Earth too?

Can the optimizer feature modify a model over real Earth to show a 1.00 AGT value also?

If the optimizer is necessary to modify the model to produce this 1.00 AGT, how far different does it make the model vs. the model to start?

I made the free space version of the model in 4Nec2 and noticed the same thing. If I put the feed point at the base of the radiator where I would expect it to be, AGT goes above 1.2, which is saying the model is not even close to accurate. I don't like putting the feed point at that location, but I guess that is a limit of the modeling software based on that design.

I found the same thing DB, but after we talked about how you model a little and how we set our segment count...I began to wonder. In the past I recall casually thinking I had to make my segment lengths for a wire shorter in order to try and get the feed point closer to the bottom of the radiator in many cases were needed. I have also used the split feed point feature to get the virtual feed point closer to the junction of two connected wire...being as Eznec, generally, places the source at the middle of a wire and not the end.

If you model a mathematical feed point match in a model and you change the location of the source...will you have to re-fix the match?

More recently I have been using a diameter just under 1/4 inch, although I was thinking it was just under 1/2 inch. That was a mistake on my part. Remember, most of my models are based on a meter, not an inch.

You explained that to me recently and I understand what you said.

However, in the past I also observed that in you models, but you never explained before. That was another apples and orange exchange between us.

That aside however, and I could be wrong, but I'm curious. What is your thinking in having a Vector model, as an example, with all the wires diameters, not the segments, set the same like you did with your recent linear loaded model?

Also a note on wire diameter I've noticed. Thinner diameter wires often give a better AGT result. Just something to take note of.

I totally agree here too DB. Making a model using very thin wires instead of using typical real world material dimensions also makes the model much more prone to violating modeling limitations and circumventing reasonable segment notices, but at what cost otherwise. I don't know.

One change I would recommend trying is making the area of the main radiator inside the "basket" area to its own wire, and size it's length and it's segments to match wired 14 and 15. It might not make much of a difference, but when wires are close and line up like that I've noticed in the past that where the segments are can sometimes make a difference. This isn't always the case, but sometimes it does make a difference. If you are going to play with segment length, I would make this one modification.

Good point and I will fix that. Maybe that is why my Super Magnum model is so dependent on having a specific segment count in order that the AGT show a good low value. If that doesn't get fixed or explained for me...I will not have much confidence in the model.

I also did not consider making the segment lengths for wires 14 and 15 equal the segment lengths in the radiator, an issue we learned about when working with the S4/NV4K models. It sounds like a good idea.

One other note on that model, 4Nec2 gives me warnings for wires 2, 3, 4, 5, 10, 11, 12, 13, 16, 17, 18, and 19. It is basically saying the segment length divided by the radius is below 6. This is a warning, not an error, it would be an error if said calculation was below 2. What the warning is telling us is that the model might have a problem, not necessarily that there is a problem.

I agree here again. I was doing the math to calculate my segment counts for my wires...trying to get all the wires with some ratio like 3, 4, 5, 6, but I did not think about it like you explain here.

If we had not been constructively talking lately...I doubt I would have ever consider any of these ideas and they are some good ones.

Thank you.

There are three ways to fix this normally, one would be to use fewer segments in the wires, however they all have 1 segments each. Another option would be to make said wires longer, but that would affect the model. In 4Nec2 I would go with the remaining option, and that is make the wires thinner. Not necessarily a lot thinner, but just enough to make the warnings go away.

Well, I disagree here. We know what happens with resonance when we make the wires thinner, and that would of course change the match too...not sure how much, but IMO that is not necessarily the right way to go.

I firmly believe in making a model as close to what we will do in real world construction. I see my models that do not require matching...showing pretty much what science predicts...and that satisfies my requirements.

That said however, I don't say you are wrong in your claim...but I need to check this idea further.

I'm reminded of the Sigma 4 model, a model that has closely space elements, which is known to violate Eznec limitation. Eznec has an error check that alerts when this happens, but I used a simple hub design that mitigates the problem within reasons, and take note: we also find this hub on the real world antenna itself, and maybe it is necessary to be there too. I seem to recall that both Homer and Booty Monster found the hub was also needed to offset the radials from the radiator a little before their homemade Vector would work as they expected.

Aside from these, my reproduction of the free space model's results come very close to your free space model's results.

I appreciate your saying that. I was amazed to find that this model, as noted below in the wires view, only produces a good AGT if the radiator is set to 33 segments over the entire length of the radiator.

I my case if I change the count to 32 or 34 the AGT shows an AG = 2.00<> with that one small change.

 

[The DB]

I won't try and interpret the meaning, but I don't see how the AGT over real Earth can produce viable information...unless the software is able to determine all the material and Earth losses in its models. Eznec won't do that. The modeler has to physically turn off the losses due to ground type as well as the material resistance losses in the structure.

DB, if you can get a AGT value = 1.00 for a free space model, are you still able to get a similar low value = 1.00 if your model is set over real Earth too?

Can the optimizer feature modify a model over real Earth to also show a 1.00 AGT value also?

If the optimizer is necessary to modify the model to produce this 1.00 AGT, how far different does it make the model vs. the model to start?

AGT isn't calculated over "real earth", if you are modeling over an earth, Nec2 uses "Perfect Ground" instead.

When modeling over an earth, the AGT result, while not always exactly the same result as running AGT in free space, is still very close.

The optimizer cannot optimize for AGT, unfortunately.

If you model a mathematical feed point match in a model and you change the location of the source...will you have to re-fix the match?

Yes.

What is your thinking in having a Vector model, as an example, with all the wires diameters set the same like you did with your recent linear loaded model?

I'm pretty sure the one I did in my I-10K modeling thread for 222DBFL was that way, actually, looking at it that is not correct. The matching system wires I added are a little thicker. If I remember correctly, that may have had something to do with correcting AGT, but I am not sure. It may have simply been a typo on my part...

Good point and I will fix that. Maybe that is why Super Magnum model is so dependent on having a specific segment dimension in order that the AGT show a good low value. If that doesn't get fixed or explained for me...I will not have much confidence in the model.

I think part of that problem has to do with how modeling tries to handle the section below where the feed point is and the area of the antenna below it, namely how modeling handles the mounting plate you are trying to simulate.

I'm reminded of the Sigma 4 model, a model that has closely space elements, which is known to violate the limitation. Eznec has an error check that alerts when this happens, but I use a simple hub design that mitigates the problem within reasons, and take note: we also find this hub on the real world antenna itself, and maybe it is necessary to be there too.

I seem to recall that both Homer and Booty Monster found the hub was also needed to offset the radials from the radiator a little before their homemade Vectors would work as expected.

Yea, you don't want a wire to have an angle of less than 90 degrees with another wire it is attached to. This is easy to get around, you just make a small one segment long wire and put it between the wires. The hub on the Vector design you are referring to is a perfect example. I had another model that had that as an issue, passed AGT, and in Free Space had something like 130% efficiency. As we both know, more than 100% efficiency is impossible... Its one of those things we just have to watch out for. I do remember that Homer's design with the hub worked much better than his design that did not include the hub. I can't say for sure if the hub is the part that made the difference, but it is a possibility.

I appreciate your saying that. I was amazed to find that this model, and noted below in the wires view, only produces a good AGT if the radiator is set to 33 segments over the entire length of the radiator.

Add that wire as I suggested above, that will be wire 20, put the feed point at the base of that 2 segment wire, first segment up from the bottom. The upper part of the radiator will still be wire 1. Move its bottom point to the top of the new wire 20, the top point stays the same. I also increased the number of segments for wire 1 from 33 to 49. Do that in your software and see what that does for your AGT numbers. 4Nec2 is getting me very close to an AGT of 1 with those changes, and the feed point is where we want it to be. This is still working with the first set of data you posted in this thread, not others that changed the height and some of the diameters, although those changes shouldn't make that much of a difference.


The DB

 

[Marconi]

AGT isn't calculated over "real earth", if you are modeling over an earth, Nec2 uses "Perfect Ground" instead.

You are right, Eznec can use Perfect ground and Free Space to generate the AGT, but my models show a little difference. I misread you words thinking, when you said 4Nec2 would generated an AGT for a real Earth model if you set the ground to Perfect. You obviously meant something else, and I missed it.

When modeling over an earth, the AGT result, while not always exactly the same result as running AGT in free space, is still very close.

The optimizer cannot optimize for AGT, unfortunately.

Again, we talked about this earlier.

I think part of that problem has to do with how modeling tries to handle the section below where the feed point is and the area of the antenna below it, namely how modeling handles the mounting plate you are trying to simulate.

I see that in my model, but the currents on that wire typically seem to progress in phase smoothly and transition to the radiator where I usually see "0" phase at the feed point. Here again I have not considered this area in this light. Thanks again.

Yea, you don't want a wire to have an angle of less than 90 degrees with another wire it is attached to. This is easy to get around, you just make a small one segment long wire and put it between the wires. The hub on the Vector design you are referring to is a perfect example.

I agree, but I find this little wire length can make a noticeable difference in performance values, and maybe even effect the match. I also find the shorter the better, until you get a geometry error.

Yes.

I had another model that had that as an issue, passed AGT, and in Free Space had something like 130% efficiency. As we both know, more than 100% efficiency is impossible... Its one of those things we just have to watch out for.[/QUOTE]

AFAIK, Eznec does not produce efficiency data.

I do remember that Homer's design with the hub worked much better than his design that did not include the hub. I can't say for sure if the hub is the part that made the difference, but it is a possibility.

Well I thought I remembered Homer adding his offset hub is what made the difference in his case, but I think Homer has disputed that...so I'm not sure. I know that BM posted images of a nice hose clamp setup he made for his Xtra long Vector that place his radials flush against the radiator, and he too had to do an offset to get his working as well. If you can find his images here you will see what I'm referring too.

That said, I'm convinced that the real antenna needs a radial offset from the radiator in order for the antenna to work right.

Add that wire as I suggested above, that will be wire 20, put the feed point at the base of that 2 segment wire, first segment up from the bottom. The upper part of the radiator will still be wire 1.

Another good idea...that I will try. I plan to revisit some of my CB models and do the Free Space model with a good AGT first and then set the model over real Earth and post any real differences noted. I'm convinced this might be good process. That said however, I have to first try and get to the bottom of this specific segment value issue, in order to get a good AGT that I talk about above.

Move its bottom point to the top of the new wire 20, the top point stays the same. I also increased the number of segments for wire 1 from 33 to 49. Do that in your software and see what that does for your AGT numbers.

Yes. 4Nec2 is getting me very close to an AGT of 1 with those changes, and the feed point is where we want it to be. This is still working with the first set of data you posted in this thread, not others that changed the height and some of the diameters, although those changes shouldn't make that much of a difference.

I will try and remember to fix my Super Magnum model that has the radiator set at .75" inches diameter for the radiator and .625" inches for all the rest of the wires.

I must have used your Linear Loading model to build my Super Magnum model. This is probably why I ended you with all the wires the same for this new model. I might post it up later if I can fix it to show similar AGT results...and for sure if I can stop the very sensitive reaction I'm getting with my segment settings.

DB, I having a terrible reaction to a couple of new drugs the Doctor put me on a few days ago, so I hope these words make some sense.

 

[Marconi]

I will try and remember to fix my Super Magnum model that has the radiator set at .75" inches diameter for the radiator and .625" inches for all the rest of the wires.

I modified the (SM) model above, and here is what I found in changing the wire diameters to what I think the real antenna might be. I also used 6" inch long wire segments on each wire that had more than 1 segment.

You are right DB, the model works and as long as most of the wires have similar segment lengths...the model should show a good AGT if the source is located correctly.

I did not have to use your short wire workaround to get the AGT = .991 = 0.04 db, which is good enough for government work.

Main Control Screen for Eznec.


Wires descriptions as changed from same (SM) model posted above. Also note that I made the 2" wire mounting bracket, wires #14 and #15, 1" inches in diameter each to represent the solid 12" x 2" back of the bracket, and it worked fine.

The AGT did loose 0.01 to 0.03 db, but the value is very low and that is what counts. So I no longer think using thin wires contributes to what I saw ill-effecting these models.

Wires descriptions.


Antenna Pattern shows about the same as (SM) model above.


Within reason, the magic here is in the wire segment lengths being as close to the same regardless of the diameter.

I have been doing this with my models for some time, and I thought I had the right idea...but I did not really know why. This discussion brought this to light DB, thanks.

It is always a good day when we learn something new and useful.

 

[Marconi]

When I changed the (FS) settings to Real Earth, the model broke. The pattern is skewed badly, lots of currents on the mast, and I don't yet know why.

 

[Marconi]

DB, I forgot to add the wires screen for my post above and WWDX would not let me edit the post.

 

[The DB]

I played with it a bit, yep, over ground the model is broke, and not even close to passing an AGT. I cam make adjustments that make small changes in the AGT, but nothing that gets it anywhere close.

I would have to do some thinking, I don't know if there is a way to save it or not...


The DB

 

[Henry HPSD]

agt can not be calculated over real earth.

So, the "line" :
over ground the model is broke, and not even close to passing an AGT. I cam make adjustments that make small changes in the AGT, but nothing that gets it anywhere close.

Makes me very worried, this is just...well...
I dont know how to put to...
Im worried if the concept of AGT is understood.
On the other hand, I am Dutch...and it could be my misinterpertation.
Or perhaps that was not what you guys were after ?
sorry...just catching up...and this is "basic" stuff you can find in the manual....
Forgive me, if I answered to `fast` ..as i know you guys have been rather bussy with all the software...
So I expect you know this, as this is basic material, and that either worries me....
(as i certainly dont want to be rude)
or perhaps you have found something that i didnt think of....that would be fantastic !
Let me know and ill catch up ! tnx !

 

[The DB]

Hey Henry, I think something was lost in translation, or something was assumed that was not picked up on. When running AGT over ground, at least in 4Nec2, it changes the ground to a perfect ground for said test. Does that not happen in EZNec as well?


The DB

 

[Henry HPSD]

Hi !

4nec2 and Eznec have more or less the same `thing` going on.

In order to calculate the AGT in 4nec2 you need to `highlight` a option.
In order to calcualte the AGT in Enzec you need to run a 3d plot in freespace

both are the same thing.

As AGT is calculated in freespace on the amount of power (loss) from a 3d plot.
4nec2 ehm...just has a additional `button` for it....

Im wondering ...why you would model it over real earth......

 

[Henry HPSD]

Ehm...
The difference in both is more or less only one step.

Eznec only calculates the AGT if you run a 3d plot.
Now, you can do this in sever ways..
Above real earth or perfect or the last one... as in free space.
So, if you want to `know` the AGT...you know...it only can do that with a 3d Plot..
So, you are allready in the `3D` `stage`.
All you need to do is `press freespace`.

In 4nec2 all you need to do is...
Run freespace etc..click on `average gain` and click on any warnings that come up in order to let the software run `its 3d freespace` plot.
(thats more `clicks` thinking of it hihi)

The only `thing` that could go wrong...
Is that a Eznec user worries about the AGT if he did a run above real earth.
But I consider that a situation where someone hears a bell but hasnt got a clue where the tower is where the bell is situated.

 

[The DB]

Hi !

4nec2 and Eznec have more or less the same `thing` going on.

In order to calculate the AGT in 4nec2 you need to `highlight` a option.
In order to calcualte the AGT in Enzec you need to run a 3d plot in freespace

both are the same thing.

As AGT is calculated in freespace on the amount of power (loss) from a 3d plot.
4nec2 ehm...just has a additional `button` for it....

Im wondering ...why you would model it over real earth......

A screenshot of the added "warning" window in 4Nec2 when running AGT...

If you actually look at the input file that 4Nec2 uses, there is a line...

GN 1

This is telling Nec2 to use a "Perfect Ground", and if you have a ground while running AGT, it will always be set to this setting. There are no exceptions, unless 4Nec2 is malfunctioning.

If you simply remove said check from the AGT checkbox the same model is now has...

GN 2 0 0 0 5 0.001

Which is "Real Earth" and then "Moderate". This ground has a conductivity of 0.001 and a dielectric constant of 5. This, as I mentioned above, gets changed if AGT is run.

To say that 4Nec2 is running, or even can run an AGT over "real earth" is not, and can not be true. Perhaps there is a translation issue here?

That is not to say that 4Nec2 cannot run AGT over an earth, that earth will always be "Perfect Ground", and this is by design. If you cannot run AGT over an earth, why does 4Nec2, which makes preparations for said case, and even mentions "ground" as part of the "losses", allow it and not just switch it to "Free Space" automatically as it already changes it to "Perfect Ground"?

Now, it is possible that their could be a problem running AGT over a "Perfect Ground" in Nec2, however, that is not what you'r text is saying, and if such a problem does exist I have yet to come across it in my studies.

I'm sure you know this, but AGT works by averaging all of the gain calculations from all directions. This average should equal 1. As long as you have no losses this is the same weather you are in free space or over a perfect ground.


The DB