[TowerTalk] GM card in NEC

[jimlux]

On 4/24/19 10:37 PM, Greg Best via TowerTalk wrote:
> Thanks to you all for your questions, thoughts, and input.  My original question was about interaction with the tower and really asking the question as to whether someone besides me has addressed the issue before. (Better to take advantage of work done before rather than recreate the wheel.). For your further consideration, the tower face at 60 ft is about 30” and it is 3 sided. The mount would likely be fixed on EU or a small robust  sidearm with “rotator on a post” realizing that I would not be able to turn it fully around. All of the bracing is  “X”
> braces using 1.5” angle steel above and below the 60 ft level.  The spacing between the DE & D1 is 106” so i do have some room. The balance point is abt 63” from the DE. The 20 m beam is a Wilson 204CD. I am not familiar with the NEC command to “mirror “ the wires. Is it “copy” or something else?
> Tks. Greg N9GB
> 
> 
> Greg Best


it's the GM card in a NEC deck - various front ends like EZNEC, etc. 
provide a slightly different interface, but that's how it winds up.

It is a bit confusing - using a tool like 4nec2 that shows you a picture 
of the geometry helps a lot.

The GM card says "move" a bunch of wires that have already been defined. 
You have the choice of just "moving" them, or making multiple copies 
while moving

If you make multiple copies you have to say what the increment in wire 
number is: for instance, if you've defined wire 101,102,103,104, and you 
make 2 copies, you could specify an increment of 10, and it would 
generate 111,112,113,114 and 121,122,123,124

You can specify the move as rotation around an axis (specified as 
rotx,roty,rotz)  and translation (shiftx, shifty, shiftz).

Rotations are tricky, because they're around the axis.  Say you built a 
corner of your tower centered on X=0,Y=0.  If you rotate it around Z it 
stays in the same place. So if you make copies, all the copies wind up 
on top of each other.  So what you do is either build it off center, or 
use one GM card to move it horizontally (say to X=1,Y=0), then the next 
GM card to rotate it and make the copies. Getting everything to line up 
is non-trivial.

Translations are much easier - it's how you build up things like grids 
fairly quickly. You put one wire in, say from (0,0,0) to (0,1,0), then 
copy it with a GM card that shifts in X, so you'd wind up with a wire at 
(0,0,0)-(0,1,0), then (0.1, 0,0)-(0.1,1,0), then (0.2,0,0)-(0.2,1,0) and 
so on.

There's also options to say "move everything before the GM card" or 
"move only certain wires".  The latter has some restrictions, so i tend 
to build the stuff that's going to need moving first, then the stuff 
that won't need moving later in the deck.


And, while it is theoretically possible to build almost any geometry 
with judicious use of GW and GM cards, at some point, it's easier to 
just explicitly generate GW cards for each wire in some other program. 
This is particularly true if you're making things that are curved or at 
non-trivial angles.

 From a run time standpoint, structures that are symmetric are faster, 
but for the most part, my time is more valuable than the computer's, so 
I don't worry about that.

I have been running a model that has roughly 1000 segments over the past 
few days, and running 200 frequency steps using the Sommerfeld-Norton 
ground takes <15 seconds on a 3 year old Dell desktop.