Call Search
     

New to Ham Radio?
My Profile

Community
Articles
Forums
News
Reviews
Friends Remembered
Strays
Survey Question

Operating
Contesting
DX Cluster Spots
Propagation

Resources
Calendar
Classifieds
Ham Exams
Ham Links
List Archives
News Articles
Product Reviews
QSL Managers

Site Info
eHam Help (FAQ)
Support the site
The eHam Team
Advertising Info
Vision Statement
About eHam.net

   Home   Help Search  
Pages: [1]   Go Down
  Print  
Author Topic: modeling stranded steel core copper wire  (Read 4546 times)
W7NUW
Member

Posts: 35




Ignore
« on: March 16, 2013, 12:34:30 AM »

I am using The Wireman's #532 "Silky" 18 AWG to make a 40 m dipole. EZNEC modeling vs careful measurements gives curious results.

For operation at 7.000 mHz, EZNEC shows resonance with a length of  66.4 feet, at a height of 12 feet. (This is not the ultimate intended height, but one at which I can easily make direct measurement and trim the length). But with this length, measured resonance is at about 6.80 mHz. I am taking the jacket dielectric constant (2.2) and thickness (0.02") into account in the model, and specifying 'copper' for wire loss.

I must reduce the physical length to 64.5 feet to obtain measured resonance at 7.000 mHz.  EZNEC displays 7.199 mHz with this length. Changes in dielectric properties, ground conditions, etc have minimal effect.

The only change I can make to the EZNEC model inputs that leads it to predict 7.000 mHz from this shortened antenna involves altering the wire permeability from 1 (the assumed value for copper) to something between 100 and 200 -- typical values for steel. I wonder if the measured results indicate that the steel core is playing a significant role. Using 100 as a permeability also results in a gain reduction of about 1 db.

Single core Copperweld might not show the same anomaly. The "Silky" has 19 strands and I'm thinking individual strands may not have sufficient thickness of copper to cope with the 7 mHz skin depth. Or that the individual conductors are interacting with the steel cores to increase series inductance and energy loss.

Any comments or insights on this would be welcome. I like this wire for its durability and ease of handling, but it is tedious getting wire lengths correct when the model is so far off (~3%) -- and I don't like giving away even 1 db.

Ken Lawson W7NUW
Logged
K2DC
Member

Posts: 1362


WWW

Ignore
« Reply #1 on: March 16, 2013, 03:38:18 AM »

Ken,

  A model is, after all, only a model.  It would be extremely difficult for any model (especially a PC-based model with limits to processing power) to accurately include all of the factors that will impact accuracy and resolution.  Before retiring, I was an RF/Microwave Antenna Engineer and I worked with a number of models based on NEC, Method of Moments, Uniform Theory of Diffraction, and others.  I would have considered a model that matched measured performance to within +/-3% to be excellent results and a reliable basis for design.

  BTW, there is a skin depth calculator at:

http://www.rfcafe.com/references/calculators/skin-depth-calculator.htm

73,

Don, K2DC
Logged
GW3OQK
Member

Posts: 145




Ignore
« Reply #2 on: March 16, 2013, 04:13:01 AM »

Where and how do you make your measurement of resonance Ken?
At the bottom end of a balanced feeder for instance, using a balun to maintain the balance? Or even at the centre of the dipole? Just wondering if it could be unbalanced feeder pulling the frequency down.

I have sometimes wondered how accurate are the beautiful plots eznec produces. Don's 3% could be it. 
Andrew
Logged
KB4QAA
Member

Posts: 2373




Ignore
« Reply #3 on: March 16, 2013, 07:03:59 AM »

Compared to ground conductivity and perhaps effect of nearby objects, the nuances of wire skin effect are minimal.  It is well known that the modeling software is not great at modeling ground conductivity.

In the end, a model is just a model. Don't try to read more accuracy than is obtainable with the software tools and estimation.
Logged
W8GP
Member

Posts: 207




Ignore
« Reply #4 on: March 16, 2013, 07:25:24 AM »

What did we do before EZNEC? Let's see, 468/f, make it a little long, put it up, trim for SWR,get on the air. Who cares about the wire material, the best wire to use is SOME. This is a classic case of too much information. And you're an extra class? Wow, I used to be disgusted, but now I'm just amused!
Logged
K4SAV
Member

Posts: 1840




Ignore
« Reply #5 on: March 16, 2013, 07:42:11 AM »

Looking at the Wireman's Wirebook, it's unclear to me whether this wire is all polyethylene or not.  In one place they say polyethylene and in another place they say PVC and polyethylene.   If this wire is a combination of the two, then that could account for 50% of your error. 

If you are looking for an exact answer from EZNEC, then you have to model the antenna exactly.   The little wire loops thru the end insulators contribute a small amount of end loading.  That loading will be a little different depending on whether the wire is insulated or not.  The width of the center insulator also contributes.  These will lower the resonant frequency a little, maybe as much as 50 to 100 kHz.  It's also possible that the thin copper coating on the steel wire is not thick enough to contain all the fields.  That coating is probably much thinner than the coating on the single strand #12 or #14 copperclad, and some of that is not sufficient thickness at 160 meters.  Proximity of other conductors can make a very large difference in the answer, especially when they happen to be resonant close to the frequency of the antenna.  Then there are the other minor things like proximity of your body to the antenna, the length and impedance of the little cable between the antenna and the analyzer, the stray characteristics of the balun used in the measurement, height above ground, antenna sag, and ground characteristics.

Most people ignore these minor contributors and just prune the antenna, but if you accurately model everything, then you should get an answer that is very close. 

Jerry, K4SAV
Logged
K7KBN
Member

Posts: 2803




Ignore
« Reply #6 on: March 16, 2013, 09:59:24 AM »

What did we do before EZNEC? Let's see, 468/f, make it a little long, put it up, trim for SWR,get on the air. Who cares about the wire material, the best wire to use is SOME. This is a classic case of too much information. And you're an extra class? Wow, I used to be disgusted, but now I'm just amused!

I'd been licensed for ten years before I first used one of them "SWR" meters in the late 1960s. Discovered that most of my antennas showed SWR well over 3:1, yet my DX-100B loaded full power and I worked over 250 countries.
Logged

73
Pat K7KBN
CWO4 USNR Ret.
W8GP
Member

Posts: 207




Ignore
« Reply #7 on: March 16, 2013, 10:11:40 AM »

I guess I should apologize because it appears that you are interested in the PROCESS and not the PRODUCT. It might be be more of a challange to put up your dipole FIRST and then modify your EZNEC variables intil they match reality!
Logged
W7NUW
Member

Posts: 35




Ignore
« Reply #8 on: March 16, 2013, 12:04:28 PM »

I guess I should apologize because it appears that you are interested in the PROCESS and not the PRODUCT. It might be be more of a challange to put up your dipole FIRST and then modify your EZNEC variables intil they match reality!

I have put up many dipoles since starting ham radio in 1956. With the help of modeling tools like EZNEC, I have found ways to improve the results way beyond what I achieved in the 50s.  My best and most effective performer at this point is a Coupled Resonator that covers 40, 30, 20, 17, and, at 3/2 wavelengths on the long element, 15 meters. Now I am making a lighter-weight version that I can string up on a higher, but of necessity flimsier, pole (far end is a tree).

The CR elements interact in a way that makes trimming especially tedious, much more so than just getting one dipole right. The tedium is considerably reduced by careful modeling.

And yes, I am interested in 'process'. I find it fun. Amusing. Sort of the same reason I enjoy Ham Radio.

73 Ken W7NUW
Logged
KB1GMX
Member

Posts: 773




Ignore
« Reply #9 on: March 16, 2013, 01:33:16 PM »

I just went through that exercise and at 7mhz skin depth on all but the cheapest
copper plated material is enough.  Went as far as to elevate both bare stranded
copper weld and the same gauge copper.  No measurable difference between
them.  Neither agreed 100% with the model though I tried to be accurate and
include real ground and real heights.  A free space model will be further off.

In practical situations often the model is wrong as the real factors like actual soil
conductivity and dielectric properties not in the model.  Add o that the feed system
and measurement interactions you get a answer that may agree in general but not
in specific to n-many decimal places.

Antennas interact with the environment, computers interact with people.  The latter
are far precise but less accurate.


Allison
Logged
W8GP
Member

Posts: 207




Ignore
« Reply #10 on: March 16, 2013, 05:13:40 PM »

It may have sounded like I made my "reverse engineering" comment tongue-in-cheek but in practice that might be the best way to make your antenna match the model.I have used this technique when doing modeling in other disciplines and it does work. Good luck.
Logged
W7NUW
Member

Posts: 35




Ignore
« Reply #11 on: March 16, 2013, 07:53:29 PM »

It may have sounded like I made my "reverse engineering" comment tongue-in-cheek but in practice that might be the best way to make your antenna match the model.I have used this technique when doing modeling in other disciplines and it does work. Good luck.


I agree with your view of how reality improves models: reverse engineering is exactly what I did. I put up the antenna and trimmed it for resonance, and then tried to work out why the model didn't agree. I think I know why. Skin depth at 7 mHz is about 25 micro-meters, and you need 3 skin-depths to make the wire "essentially copper". Each of the 19 strands in 532 silky is #31 wire and has a x-section radius of about 114 micro-meters. If the copper cladding were 3 skin-depths thick, there'd be very little room left for the steel core. I think 532 silky is essentially steel wire at HF. Model it as such and the model works.

Standard Copperweld is different: there's plenty of room for enough cladding thickness when you've got a single #12 or #14 conductor.

A tip of the hat to Owen, VK1OD, for his insightful email comments.

73 Ken W7NUW
Logged
Pages: [1]   Go Up
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.11 | SMF © 2006-2009, Simple Machines LLC Valid XHTML 1.0! Valid CSS!