-
You can now help support WorldwideDX when you shop on Amazon at no additional cost to you! Simply follow this Shop on Amazon link first and a portion of any purchase is sent to WorldwideDX to help with site costs.
Here is a intetesting Antenna to think about
- Thread starter AudioShockwav
- Start date
Now that you put this thread back in front of me, I started wondering why we are getting a vertical component from the super sinner antenna models. Even the test model I made early on, which was made up of 100% horizontal wires, shows a definite vertical component. As of yet, no one has posted a plausible reason as to why this is happening, at least as far as I can tell.
Can anyone explain why this is happening? Or even have any ideas? Any tests I should model?
One thought on my end was that the modeling software is in error. That being said, I have yet to see any evidence on this either way. To try and confirm this I will use another modeling program based on another engine, in this case Mmana-Gal, which is based on the mininec engine. The thing is, the mininec engine has its faults. It does a few things better than the nec2 engine, such as calculating a wire with taper, but anything outside of that very small list we have to assume that the nec2 engine is more accurate. This is clearly one of those cases that, if there is a difference, we will have to favor the nec2 results. Therefore, while this should still be done, I feel it may end up being inconclusive.
The next thought is, is the earth is playing a part? At least in the models I have made, in all cases so far this antenna is being modeled at less than 1/4 wavelength above the earth. I don't think this will actually be the cause, but it is something to look in to. This will be simple enough to test by modeling the antenna is free-space.
The next thought is it has something to do with the antenna's layout. I wonder what the model will look like if I point the part of the antenna after the bend straight up instead of it being horizontal? That will give the antenna vertical elements, so how will that model look compared with the same elements being horizontal?
The next thought of the antenna's layout is what if one of the bent elements is reversed to point the other direction? Will this have any effect on the vertical polarized radiation?
This is where I am at so far. Does anyone else have any thoughts on tests I should run with the modeling software?
The DB
Can anyone explain why this is happening? Or even have any ideas? Any tests I should model?
One thought on my end was that the modeling software is in error. That being said, I have yet to see any evidence on this either way. To try and confirm this I will use another modeling program based on another engine, in this case Mmana-Gal, which is based on the mininec engine. The thing is, the mininec engine has its faults. It does a few things better than the nec2 engine, such as calculating a wire with taper, but anything outside of that very small list we have to assume that the nec2 engine is more accurate. This is clearly one of those cases that, if there is a difference, we will have to favor the nec2 results. Therefore, while this should still be done, I feel it may end up being inconclusive.
The next thought is, is the earth is playing a part? At least in the models I have made, in all cases so far this antenna is being modeled at less than 1/4 wavelength above the earth. I don't think this will actually be the cause, but it is something to look in to. This will be simple enough to test by modeling the antenna is free-space.
The next thought is it has something to do with the antenna's layout. I wonder what the model will look like if I point the part of the antenna after the bend straight up instead of it being horizontal? That will give the antenna vertical elements, so how will that model look compared with the same elements being horizontal?
The next thought of the antenna's layout is what if one of the bent elements is reversed to point the other direction? Will this have any effect on the vertical polarized radiation?
This is where I am at so far. Does anyone else have any thoughts on tests I should run with the modeling software?
The DB
I have to say you guys are great taking the time to model these antennas.
DB, you and Eddie, and guys like Bob85, Henry and a lot more have devoted plenty of effort to help the rest of us " see" what these designs are doing.
Thanks Guys.
73
Jeff
So, as no one else had anything to add, I will show some of my initial results here.
The first image is created with 4nec2. It is looking at the radiation pattern on the horizontal plane, or like you are looking down on it from above. The antenna was mounted 10 feet above the earth, and this version of the antenna is made out of 100% horizontal wires. Also, the radiation pattern that is shown is at an angle of 59 degrees above the horizon. There are two reason this angle is what has been shown. One, it is the default angle that the modeling software used, which happens to be where the antenna's peak total gain happens to be. And two, it is able to show where both the vertical and horizontal components in one screenshot, so there is no confusion that the modeling software is showing that this modeling software is showing both polarizations being radiated.
Also, the conventions for these images. The orientation, the Y axis, or from our pints of view up and down, will always be in line with the vehicle facing its front and back. The X axis is the same vehicle, but the left and right sides. This is in conjunction with how the pictures above show how this antenna is mounted to a vehicle.
[photo=large]6732[/photo]
The green line is the total combined gain, the red line is horizontal gain, and the blue line is vertical gain.
This is a test model that shows, again, that this antenna is in fact producing vertically polarized radiation.
So the question is why? I stated above that this model, which is similar to the Super Sinner antenna is made up of 100% horizontal wires. Where is this vertical component coming from?
The next step was to check these results with another program, so I made a model with the same dimensions using Mmana-Gal, a mininec based modeling software, and got these results.
[photo=large]6731[/photo]
In this case, the model that matches what is shown above is on the left. As far as vertical and horizontal polarization, the colors are reversed with this program. I don't know if there is a way to change the color layout, and this is likely the only model I will use from this program so I will just let it be this time. So while in 4nec2 horizontal polarized gain is shown with a red line, here it is shown with a blue line. The opposite is true for vertical radiation. I should also point out that this software is showing gain at 45 degrees off of the horizon, again the program's default when running this model. However, while this is a different angle, the result is clear, there is vertical polarized gain present.
So what does this mean. These programs use two different implementations of what is called the "Method of Moments" or MOM. Because of this, as both show a vertical component, it is very unlikely that this is just a software issue.
Because of what I stated above, its likely not a software error. So, where does this vertically polarized gain come from?
One thought I had was that the earth might play a part, any while I said I didn't think it was, it was easy enough to look into, so here I am going to model the antenna in frees-space to rule out the effects of the earth.
[photo=large]6730[/photo]
Here we see a free-space model of the antenna, again at 59 degrees above the horizon, just like the initial test model above. And here we also see that the vertical component is still present. This rules out the earth below the antenna being the cause of said vertically polarized radiation..
That is all I have for now, from here I will try changing the model layout in various ways to see what happens. That, however, will have to wait until I have more time.
The DB
The first image is created with 4nec2. It is looking at the radiation pattern on the horizontal plane, or like you are looking down on it from above. The antenna was mounted 10 feet above the earth, and this version of the antenna is made out of 100% horizontal wires. Also, the radiation pattern that is shown is at an angle of 59 degrees above the horizon. There are two reason this angle is what has been shown. One, it is the default angle that the modeling software used, which happens to be where the antenna's peak total gain happens to be. And two, it is able to show where both the vertical and horizontal components in one screenshot, so there is no confusion that the modeling software is showing that this modeling software is showing both polarizations being radiated.
Also, the conventions for these images. The orientation, the Y axis, or from our pints of view up and down, will always be in line with the vehicle facing its front and back. The X axis is the same vehicle, but the left and right sides. This is in conjunction with how the pictures above show how this antenna is mounted to a vehicle.
[photo=large]6732[/photo]
The green line is the total combined gain, the red line is horizontal gain, and the blue line is vertical gain.
This is a test model that shows, again, that this antenna is in fact producing vertically polarized radiation.
So the question is why? I stated above that this model, which is similar to the Super Sinner antenna is made up of 100% horizontal wires. Where is this vertical component coming from?
The next step was to check these results with another program, so I made a model with the same dimensions using Mmana-Gal, a mininec based modeling software, and got these results.
[photo=large]6731[/photo]
In this case, the model that matches what is shown above is on the left. As far as vertical and horizontal polarization, the colors are reversed with this program. I don't know if there is a way to change the color layout, and this is likely the only model I will use from this program so I will just let it be this time. So while in 4nec2 horizontal polarized gain is shown with a red line, here it is shown with a blue line. The opposite is true for vertical radiation. I should also point out that this software is showing gain at 45 degrees off of the horizon, again the program's default when running this model. However, while this is a different angle, the result is clear, there is vertical polarized gain present.
So what does this mean. These programs use two different implementations of what is called the "Method of Moments" or MOM. Because of this, as both show a vertical component, it is very unlikely that this is just a software issue.
Because of what I stated above, its likely not a software error. So, where does this vertically polarized gain come from?
One thought I had was that the earth might play a part, any while I said I didn't think it was, it was easy enough to look into, so here I am going to model the antenna in frees-space to rule out the effects of the earth.
[photo=large]6730[/photo]
Here we see a free-space model of the antenna, again at 59 degrees above the horizon, just like the initial test model above. And here we also see that the vertical component is still present. This rules out the earth below the antenna being the cause of said vertically polarized radiation..
That is all I have for now, from here I will try changing the model layout in various ways to see what happens. That, however, will have to wait until I have more time.
The DB
I'm not so sure this antenna is radiating vertically polarized radiation anymore.
In both nec2 and mininec, if I have a straight vertical center fed dipole (or in this case a straight wire on the Z or vertical axis), in free-space, all of its radiation is showing up as vertically polarized, as we would expect.
The thing is, if I take the same wire and put it on either the X or Y axis, both of which are horizontal, again in free space, I am getting just as much vertically polarized radiation as I am horizontally polarized radiation.
Being in free space, the Z axis should act exactly like the X and Y axis, but for some reason, in this case, it isn't. The gain is the same, the total patterns are the same, but the polarizations are coming up different.
When dealing with gain, as the total gain figures are the same in all cases, I see no reason to question those numbers. The only thing I see the need to question is weather the radiation is vertically or horizontally polarized.
This is unexpected, and something that in all of my research on modeling I don't recall ever reading about this variance. This is definitely not something that has been documented anywhere that I am aware of. I have spent way to long trying to make sense of it, and all I can come up with is it is something to do with the limits of the day said programs were developed.
All this being said, with the new evidence in play, I am going to take the side that the super Sinner Antenna is only horizontally polarized, at least until I see evidence that suggests otherwise.
Unfortunately, everything available is based on either Nec2 or Mininec, unless I want to spend quite a bit of money. I suppose I could get Nec4 or Nec5, but there is no guarantee that this has been fixed in said versions of the software...
The DB
In both nec2 and mininec, if I have a straight vertical center fed dipole (or in this case a straight wire on the Z or vertical axis), in free-space, all of its radiation is showing up as vertically polarized, as we would expect.
The thing is, if I take the same wire and put it on either the X or Y axis, both of which are horizontal, again in free space, I am getting just as much vertically polarized radiation as I am horizontally polarized radiation.
Being in free space, the Z axis should act exactly like the X and Y axis, but for some reason, in this case, it isn't. The gain is the same, the total patterns are the same, but the polarizations are coming up different.
When dealing with gain, as the total gain figures are the same in all cases, I see no reason to question those numbers. The only thing I see the need to question is weather the radiation is vertically or horizontally polarized.
This is unexpected, and something that in all of my research on modeling I don't recall ever reading about this variance. This is definitely not something that has been documented anywhere that I am aware of. I have spent way to long trying to make sense of it, and all I can come up with is it is something to do with the limits of the day said programs were developed.
All this being said, with the new evidence in play, I am going to take the side that the super Sinner Antenna is only horizontally polarized, at least until I see evidence that suggests otherwise.
Unfortunately, everything available is based on either Nec2 or Mininec, unless I want to spend quite a bit of money. I suppose I could get Nec4 or Nec5, but there is no guarantee that this has been fixed in said versions of the software...
The DB
Last edited:
@ DB,
If you want ...i could check with NEC4 ..for you...just send me the file.
I do have a few "other" options that i could try.
I havnt read the thread yet...but will do later...work calls...
Kind regards, H.
If you want ...i could check with NEC4 ..for you...just send me the file.
I do have a few "other" options that i could try.
I havnt read the thread yet...but will do later...work calls...
Kind regards, H.
Here is an attempt to show what I am seeing. I had to get out an older windows based laptop to use this tool as it doesn't work through wine on my current linux machine...
Anyway, here is total gain, we are looking down the Z axis, the dipole antenna is on the X axis, which goes from left to right, and the Y axis goes up and down.
[photo=large]6742[/photo]
Its the standard dipole in free-space layout, only instead of looking at it in a 2d slice, we are looking at it in 3d from end on broadside of the antenna. This is what we would expect to see from a dipole at this angle.
The horizontaly polarized gain should look exactly the same as this antenna is a horizontally polarized antenna, but...
[photo=large]6743[/photo]
It clearly doesn't. From this angle, it actually looks like we are looking at a 2d slice, not a 3d representation of this antenna.
And finally the vertically polarized component.
[photo=large]6741[/photo]
What seems to be happening is, anything that lines up with the YZ axis plane is being treated as horizontally polarized radiation, while any radiation along the XZ axis plane is being treated as vertical radiation. As I stated above, this does not happen if the antenna is on the Z axis, just the X and Y axis, and likely any layout on the XY axis plane.
Henry, I will post the 4nec2 and the raw nec2 file text of the test models below. The reason there are three wires on each is because I started with the model that was similar to the super sinner antenna, and then straightened the other two wires out to create the dipole.
This is the model that is lined up on the X axis, fist the 4nec2 file, then the raw nec2 file.
Next the model that lines up on the Y axis.
And finally the Z axis.
Also, while you are here, have you seen or done anything with nec5 yet?
The DB
Anyway, here is total gain, we are looking down the Z axis, the dipole antenna is on the X axis, which goes from left to right, and the Y axis goes up and down.
[photo=large]6742[/photo]
Its the standard dipole in free-space layout, only instead of looking at it in a 2d slice, we are looking at it in 3d from end on broadside of the antenna. This is what we would expect to see from a dipole at this angle.
The horizontaly polarized gain should look exactly the same as this antenna is a horizontally polarized antenna, but...
[photo=large]6743[/photo]
It clearly doesn't. From this angle, it actually looks like we are looking at a 2d slice, not a 3d representation of this antenna.
And finally the vertically polarized component.
[photo=large]6741[/photo]
What seems to be happening is, anything that lines up with the YZ axis plane is being treated as horizontally polarized radiation, while any radiation along the XZ axis plane is being treated as vertical radiation. As I stated above, this does not happen if the antenna is on the Z axis, just the X and Y axis, and likely any layout on the XY axis plane.
Henry, I will post the 4nec2 and the raw nec2 file text of the test models below. The reason there are three wires on each is because I started with the model that was similar to the super sinner antenna, and then straightened the other two wires out to create the dipole.
This is the model that is lined up on the X axis, fist the 4nec2 file, then the raw nec2 file.
Code:
CM
CE
SY el=73.287
SY ah=120
GW 1 1 30 0 0 -30 0 0 .125
GW 2 1 30+el 0 0 30 0 0 .125
GW 3 1 -30 0 0 -30-el 0 0 .125
GS 0 0 0.0254
GE 0
GN -1
EK
EX 0 1 50% 0 1 0 0
FR 0 0 0 0 27.2 0
EN
Code:
CM
CE
GW 1 15 30 0 0 -30 0 0 0.125
GW 2 17 103.287 0 0 30 0 0 0.125
GW 3 17 -30 0 0 -103.287 0 0 0.125
GS 0 0 0.0254
GE 0
GN -1
EK
EX 0 1 8 0 1 0
FR 0 0 0 0 27.2 0
XQ
ENNext the model that lines up on the Y axis.
Code:
CM
CE
SY el=73.287
SY ah=120
GW 1 1 0 30 0 0 -30 0 .125
GW 2 1 0 30+el 0 0 30 0 .125
GW 3 1 0 -30 0 0 -30-el 0 .125
GS 0 0 0.0254
GE 0
GN -1
EK
EX 0 1 50% 0 1 0 0
FR 0 0 0 0 27.2 0
EN
Code:
CM
CE
GW 1 15 0 30 0 0 -30 0 0.125
GW 2 17 0 103.287 0 0 30 0 0.125
GW 3 17 0 -30 0 0 -103.287 0 0.125
GS 0 0 0.0254
GE 0
GN -1
EK
EX 0 1 8 0 1 0
FR 0 0 0 0 27.2 0
RP 0 181 361 1003 -180 0 1 1And finally the Z axis.
Code:
CM
CE
SY el=73.287
SY ah=120
GW 1 1 0 0 30 0 0 -30 .125
GW 2 1 0 0 30+el 0 0 30 .125
GW 3 1 0 0 -30 0 0 -30-el .125
GS 0 0 0.0254
GE 0
GN -1
EK
EX 0 1 50% 0 1 0 0
FR 0 0 0 0 27.2 0
EN
Code:
CM
CE
GW 1 15 0 0 30 0 0 -30 0.125
GW 2 17 0 0 103.287 0 0 30 0.125
GW 3 17 0 0 -30 0 0 -103.287 0.125
GS 0 0 0.0254
GE 0
GN -1
EK
EX 0 1 8 0 1 0
FR 0 0 0 0 27.2 0
RP 0 181 361 1003 -180 0 1 1Also, while you are here, have you seen or done anything with nec5 yet?
The DB
hello DB,
I havent looked at the model yet..
But i did read back...
Could it be: what you are running into is the aspect of time...and a 2d screen.
Besides thinking "3D" there is another aspect...which is time.
We know the combined EM field produces our "far field pattern".
The "total field" you are referring to is correct (as you suggested).
That field is made up out of that vertical and horizontal component.
Those two have a relation.
So if the total field is the sum of both...You can not expect that the horizontal plane would be similair to it... as it is : total minus the vertical.
To provide an indication of either means you need to illuminate the other.
.....but that is .... within 1 cycle...
Now since it a "cycle" there will be a point where it is zero ...and a point where it is max.
If you would use a time domain solver....you could see the "horizontal" or "vertical" component "moving" in amplitude
Take a look at ...roughly 30 seconds ...
Method of Moments (MoM) vs. Finite-Difference Time-Domain (FDTD) antenna
That gives an impression of 3d in time.
4nec2 and eznec etc, dont allow you to use "time domain".
You will need FDTD programms, there are a few "open source" but it is a steep learning curve to use ( i dont).
Does that sound plausible to you ?
PS, i did "catch" up on NEC5 but nothing more...as im not really bussy with modelling at present times.
I havent looked at the model yet..
But i did read back...
Could it be: what you are running into is the aspect of time...and a 2d screen.
Besides thinking "3D" there is another aspect...which is time.
We know the combined EM field produces our "far field pattern".
The "total field" you are referring to is correct (as you suggested).
That field is made up out of that vertical and horizontal component.
Those two have a relation.
So if the total field is the sum of both...You can not expect that the horizontal plane would be similair to it... as it is : total minus the vertical.
To provide an indication of either means you need to illuminate the other.
.....but that is .... within 1 cycle...
Now since it a "cycle" there will be a point where it is zero ...and a point where it is max.
If you would use a time domain solver....you could see the "horizontal" or "vertical" component "moving" in amplitude
Take a look at ...roughly 30 seconds ...
Method of Moments (MoM) vs. Finite-Difference Time-Domain (FDTD) antenna
That gives an impression of 3d in time.
4nec2 and eznec etc, dont allow you to use "time domain".
You will need FDTD programms, there are a few "open source" but it is a steep learning curve to use ( i dont).
Does that sound plausible to you ?
PS, i did "catch" up on NEC5 but nothing more...as im not really bussy with modelling at present times.
I don't think the time domain is the cause. If it were, it doesn't explain why the X and Y axis does this, while the Z axis, alone, does not. I would still expect a dipole on any of the axis in free space to create the same pattern in relation to the dipole, and expect that pattern to have the same properties throughout. I don't see how the time domain would explain or resolve this.
That being said, I don't have access to FDTD modeling either. As far as I have found until recently, most FDTD modeling was done with Matlab, which is even further out of my price range than some other options. Although I have noticed a python extension for FDTD, and I do have some very limited python programming experience, but I haven't played with it yet.
The DB
That being said, I don't have access to FDTD modeling either. As far as I have found until recently, most FDTD modeling was done with Matlab, which is even further out of my price range than some other options. Although I have noticed a python extension for FDTD, and I do have some very limited python programming experience, but I haven't played with it yet.
The DB
Understood, i was hoping that was the (easy) solution.
Ill look into it.
Hoping this weekend.....
Not that i have the solution up forehand.
I really thought that direction would be it
.....But we will get there
Now past 12'o clock local time here...im off duty, and we have the Dutch equivalent of Santa Claus coming up... so...need to get presents asap.
Regards, H.
Ill look into it.
Hoping this weekend.....
Not that i have the solution up forehand.
I really thought that direction would be it
.....But we will get there
Now past 12'o clock local time here...im off duty, and we have the Dutch equivalent of Santa Claus coming up... so...need to get presents asap.
Regards, H.
I once upon a time thought I would homebrew a Homer version of this antenna type, but never did. . .
