Canadian TV, Computing and Home Theatre Forums banner

21 - 27 of 27 Posts

·
Premium Member
Joined
·
1,127 Posts
Where and when was there a track in Kingston? That one I don't remember.
I'm 86, so I'm about 20 years older than you.

The most popular car used was the 1936 Ford coupe. I remember the hood of one car flying open during a race, but the driver could still see through a small space just above the bottom of the windshield.

Kingston is a village and a census-designated place in the town of South Kingstown, Rhode Island,
United States, and the site of the main campus of the University of Rhode Island.

Racing started at the Lonsdale Sports Arena in Cumberland, then moved to the Kingston Fairgrounds in South County. Washington County, known locally as South County, is a county located in the U.S. state of Rhode Island.
 

·
Registered
Joined
·
13 Posts
Discussion Starter #22
Download Nikiml's python scripts, the linux tarball.
nikiml's Antenna pages - Nec optimization scripts

and follow the installation instructions..
nikiml's Antenna pages - Nec optimization scripts

After it is done, u'll probably want to edit this bash script and make it executable
'chmod +x nec_eval.sh'

here is mine...
[[email protected]]: ~/src/nikiml>$ cat nec_eval.sh
python2 -m nec.eval --validate-geometry=0 -e nec2 $*

I would then copy some examples that have been edited to be used with nikiml's python scripts
in to the same directory where that bash script lives and run it.
He has an example in his tutorial you can follow along and use.
nikiml's Antenna pages - nec.opt tutorial

say ./nec_eval.sh gh0.nec for example.
majortom,
I don't know how far you are willing to go in helping me with the operation of the software, at each stage of this process, I'm coming up with more questions. I'm wondering if I should start a separate thread for this if you're willing?

I followed your instructions, read nikiml's tutorial and then let the script run to completion. While it was running, I followed the link from nikiml's site to the user guide and began reading that (still have a ways to go).

python vs python2
in nikiml's nec-eval.sh script mikiml calls python while you call on python2. I have both installed; does it make a difference? I ran your copy, using the python2. You did not give me an nec_opt.sh script, so I used nikiml's copy which calls python. Can I combine the two like this or should I stick with one and if so which one?

Now I have (as they say here in dixie) a mess of files, what do I do with them? How do I produce the graphical files showing the antenna structure with dimension and the plots for the beam?

A question about this process. Why are we using python? Is it because nec can't read the input file directly and we don't want to maunually input each line over and over? In essence, is python providing the input to nec, capturing nec's output, then sending it back to nec for the next iteration?

(I really need to get a life, I was up to 2AM watching the computer loop and reading; before deciding the computer could run just fine without me watching, and went to bed.)
 

·
Registered
Joined
·
13 Posts
Discussion Starter #23
I've looked at a LOT of user installation reports, but I do NOT recall ANYONE using Angle Extrusion rather than Circular Rods. Obviously, Angle Extrusions would have a MUCH higher wind resistance....and may or may NOT break sooner than Rods....it would depend a LOT on the inherent strength of the chosen metal formula....I would ALSO be worried about how it holds up to wind induced vibration.

Since an outdoor antenna needs to survive high winds [esp. in your location], I would highly recommend that you stick with what we know WORKS...namely 3/8-in Aluminum Rods.
holl_ands, I've begun running the nec software following majortom's instructions. I started reading the nec user manual. I've learned that my ideas for L or square shaped elements would produce something that would be quite different from what you guys have modeled, so I am going to abandon those ideas.

About the aluminum rod; is it easily bendable? One on-line supplier's site I looked at indicated that forming and bending characteristics were poor.

How do you guys go about mechanically joining shorter pieces? The longest piece I can buy locally is 48" and the driven elements for what you recommended are longer than that. Where do you guys get your aluminum wire from?

From reading the software user manual, I can see there will be a difference between copper and aluminum, but in a practical sense will it make that much of a difference? Can I combine the two in the same antenna? Can I use copper for the driven elements and aluminum for the reflectors?

Looking at the channel reports from the two different sites, I see a problem with WPGX. One site shows I should get a good signal while the other shows a poor signal. It looks like its the only VHF channel I should be able to get. So I think I'll start with a UHF only antenna and see what happens.

I was looking through your site last night. You have UHF SBGH + 2ea 36"x18" Screen Grid Refl listed but you didn't offer that as a recommendation; I'm wondering why? Using the screen as a reflector seems as if it would be easier to construct (I like easy). Is this a viable alternative? If so, I think I might try this one first.
 

·
Registered
Joined
·
4,935 Posts
python vs python2
in nikiml's nec-eval.sh script mikiml calls python while you call on python2. I have both installed; does it make a difference? I ran your copy, using the python2. You did not give me an nec_opt.sh script, so I used nikiml's copy which calls python. Can I combine the two like this or should I stick with one and if so which one?
Whatever works in your environment.
The reason for 'python2' in mine was to be more specific.
eg - in linux OSs, there's legacy python2, and newer python3
if the user doesn't specify, the system decides based on what you had configured to be the default python environment in your environment variables.

Nikiml uses python, because it's a powerful scripting language, suitable to parse the NEC2 program's card deck inputs, program output results, automating the process of analyzing.

I agree, a separate topic covering NEC2, Nikiml's scripts, and 4nec2 in linux would be a good idea.
But we already started it here so...
Maybe you can give a moderator a heads up (report Post Tool is the 'warning icon' to the left of a post) and they can 'Spilt" the relevant posts from this thread in to a new thread for you.
 

·
Registered
Joined
·
4,935 Posts
I use copper wire for dabbling with antenna elements, since it bends easily.
It can also be straightened rather easily, by tying one end off in a vice, or some other means to anchor.
put the other end in the chuck of a drill, and let the drill run to twist the wire until it is perfectly straight. Makes for good stock to bend elements with.
 

·
Registered
Joined
·
3,396 Posts
holl_ands, I've begun running the nec software following majortom's instructions. I started reading the nec user manual. I've learned that my ideas for L or square shaped elements would produce something that would be quite different from what you guys have modeled, so I am going to abandon those ideas.

A1: Yup....NEC2 is primarily a Cylindrical Wire type model....although it does half-way support Flat "Patches"....but since you can't SEE them in 4nec2's Geometry Viewer, I avoid them like the plague. And as I said, "L" and "U" shaped extrusions will have a LOT higher wind resistance as they "cup" the wind.....although not as much an increase as square shaped extrusions. BTW: Hollow Alum. Rods would also be easier to bend...although I've tried to point you in the direction of designs that do NOT require any (or very little) bending....incl. Square Loop Folded Dipoles using Angle Connectors with either copper or aluminum type solder.

About the aluminum rod; is it easily bendable? One on-line supplier's site I looked at indicated that forming and bending characteristics were poor.

A2: Although I've built a few DIY Antennas for my own use, I'm an Antenna Designer....NOT in the Antenna Construction Business.

How do you guys go about mechanically joining shorter pieces? The longest piece I can buy locally is 48" and the driven elements for what you recommended are longer than that. Where do you guys get your aluminum wire from?

A3: Although I've built a few DIY Antennas for my own use, I'm an Antenna Designer....NOT in the Antenna Construction Business.

From reading the software user manual, I can see there will be a difference between copper and aluminum, but in a practical sense will it make that much of a difference? Can I combine the two in the same antenna? Can I use copper for the driven elements and aluminum for the reflectors?

A4: Not much difference from an electrical performance viewpoint, only minor effect, esp. SWR, seen in UHF Band.

Looking at the channel reports from the two different sites, I see a problem with WPGX. One site shows I should get a good signal while the other shows a poor signal. It looks like its the only VHF channel I should be able to get. So I think I'll start with a UHF only antenna and see what happens.

I was looking through your site last night. You have UHF SBGH + 2ea 36"x18" Screen Grid Refl listed but you didn't offer that as a recommendation; I'm wondering why? Using the screen as a reflector seems as if it would be easier to construct (I like easy). Is this a viable alternative? If so, I think I might try this one first.

A5: Yes, it's a viable alternative. I have uploaded over 800 different Antenna Designs and simply can't remember everything I've done. I try NOT to replicate what nikiml has done on his (mostly G-H) website, nor Ken Nist on his HDTVPrimer [mostly old] Antenna Analysis website, nor K6STI on his FM (and a few other) Antenna Design Website. The few G-H designs you might find on my imageevent website were either from the early days (pre-nikiml), in response to a posted candidate G-H design variation or because I wanted to generate a performance Curve for inclusion on a Comparison Chart. The UHF SBGH + 2ea 36"x18" Screen Grid Refl Design was a rare exception, per the request of a particular user:
https://imageevent.com/holl_ands/grayhoverman/uhfsbghaaxbbsplitscreengridrefl
See my embedded comments....

Note that nikiml and I have very DIFFERENT Antenna Design Objectives....he tries to find the highest GAIN with an acceptable SWR, and allows the F/B & F/R Ratio performance to be whatever it wants to be, cuz it is NOT included in his "Target Function". OTOH: I almost always TRY to design an Antenna with significant F/R (Front/Rear) Ratio performance [F/B, Front/Back is always higher since it is a sub-set of F/R for the very narrow angle directly to the "Back"]....and do several re-runs until I am happy with keeping SWR Max at JUST under an acceptable value of 2.7, which provides final maximization for Gain.

So, my UHF SBGH provided about 1 dB lower Gain, but 10 dB higher F/R & F/R Ratio performance. BTW: I've done numerous Designs with choice of either Flat Grid Reflector or Multiple Reflector Rods....the latter consistently provides higher Gain, although many are needed to match the F/R Ratio performance of the Reflector Grid (Flat or Folded).

BTW: If "someone" were to re-optimize any of nikiml's G-H designs using my type of "Target Function" [incl. with Grid Reflector], they could tailor the Gain/MismatchLoss/F_R/F_B "weighting" factors to emphasize and de-emphasize the design parameters as desired. Here is an example of my most commonly used nikiml OPT and EVAL Control Statements, incl. initial "Target Function"....before I "tweak" it as the design evolves:
[PS: "f2b_d iff" should be "f2b_diff".....something gets hozed when I copy/paste it.]

CMD--OPT -s(470,6,24) -t(13,14) --swr-target=2.7 --f2r-target=15 --f2b-target=15
CMD--OPT --target-function=(2*max_ml+8*max_gain_diff+4*max_f2r_diff+max_f2b_diff)/15
CMD--OPT --de-np=50 -r restart.log
CMD--OPT --auto-segmentation=11 --char-impedance=300 --num-cores=12
CMD--EVAL --auto-segmentation=0 --char-impedance=300 --num-cores=12
CMD--EVAL -s(174,3,15) -s(470,6,39) --total-gain --publish
 
21 - 27 of 27 Posts
Top