EFHW Antenna: A Detailed Analysis (NEC4.2)

Discussion in 'Antennas, Feedlines, Towers & Rotors' started by W9XMT, Mar 29, 2018.

ad: L-HROutlet
ad: l-rl
ad: Subscribe
ad: Left-2
ad: K5AB-Elect-1
ad: MessiPaoloni-1
ad: L-MFJ
  1. K4DJM

    K4DJM XML Subscriber QRZ Page

    Did you read the portion of my posts discussing the "worst case" of the feedline being 1/2 wavelength long? I addressed that. Do you have any idea why your simulation case with a 1/4 wavelength feedline doesn't even apply to my thesis? What is the problem with the SOURCE in your model? Do you know what is ABSENT from your model? In the context of our discussion, do you know the limitations of EZNEC?

    EZNEC uses feed impedance only in computing VSWR. You have not modeled the impedance mis-match scenario I described. I don't think you can test my thesis directly with EZNEC.

    IMHO, using a voltage source is worst case for simulation accuracy for this antenna. Going to a current source might not be enough, though, unless a way is found to model some capacitive parasitics early in the "counterpoise" (feedline) structure.

    I looked at "loads" in EZNEC and it appears like they always need to be in line with the wires. If the tool would let you put in capacitors to ground, then it might be instructive to change the drive source to a CURRENT source (making it Hi-Z) and then add a few tens of pF along the first few meters of feedline. I think this would come close to testing my assertions, but the tool won't do it!

    Keeping it 73,
    Don
     
    AI3V likes this.
  2. W5DXP

    W5DXP Ham Member QRZ Page

    Don, I wasn't arguing with you - just trying to understand. Here's a question for everyone based on the attached .EZ file. Why is the current in one 66 ft leg so much higher than the current in the other 66 ft leg on this antenna in free space used on 40m?

    BT1.png

    Or better yet, how about this 165 ft OCF fed at 66 ft and used on 40m. It has virtually the same radiation pattern as a 40m dipole.

    40mEFfs2.png
     

    Attached Files:

    Last edited: Apr 1, 2018
    KD6RF likes this.
  3. K4DJM

    K4DJM XML Subscriber QRZ Page

    The two cases are almost the same. Looking at just the right-hand leg, you've created a voltage anti-node (voltage minima) 1/2 wavelength out by virtue of one or two 1/4 open-circuit stubs. That voltage "anti-node" is replicated at the feed (it being 1/2 wavelength away) -- so you've basically created a good counterpoise that radiates very little. Those two voltage minima locations are where the current "wants" to be minimum for the line to radiate -- those points "want" to be voltage maxima.

    Consider your first case with the two 33ft stubs. What if you made them slightly shorter, like 28-30ft? Now you will see some radiation in the right-side of the antenna, because now you do not have an ideal "counterpoise". OK, so then what if you increase the number of those 28-30ft stubs to 16, 32 or more? As you increase the number the current in the right half of the antenna gets closer to zero. You've demonstrated using multiple radials to improve your "ground" system -- "counterpoise".

    I guess that you've schooled us that there can be some unfortunate choices of feed position in an OCF antenna. I had no idea.

    73,
    Don
     
    Last edited: Apr 1, 2018
    W3JJW likes this.
  4. W5DXP

    W5DXP Ham Member QRZ Page

    Can that same concept be used for an end-fed antenna? Assuming the outside of the coax is the counterpoise, why can't we choose a length that results in "a good counterpoise that radiates very little"?
     
  5. W4KJG

    W4KJG Subscriber QRZ Page

    This whole thread is a lot like a late night discussion by a bunch of people who have had too much beer, arguing what the best beer in the land actually is.

    The discussion started as a discussion of end feeding a half wave length of of wire which can do a credible job with half wave length wires using a 49:1 transformer.

    The came the discussion regarding 9:1 transformers. The 9:1 transformers are designed to work best with wires that are between odd multiples of 1/4 and 1/2 wavelengths, and their odd multiples on the intended operating bands that are usually harmonically related. The impedance at the end of that wire, against some kind of counterpoise, is selected to provide a feed point impedance of approximately 300 to 900 ohms, maybe wider.

    That is a big difference compared to using a half wave length of wire through a 49:1 transformer on some band where the end impedance against a counterpoise will be much higher (2000-5000 Ohms). It only becomes a problem in multi-band operation, like when being on another band when that half wave length becomes an odd multiple of a quarter wave length. Dividing that quarter wave feedpoint impedance (which can typically be 20 to 100 Ohms) by 49 ain't so good. Dividing those numbers by 49 kind of ruins your VSWR.

    There is no single antenna and transformer that will be perfect. Study and understand how these things work. As we've gotten well into my AARP days, my wife have decided to move into homes with minimal land, and in one case, where we have an HOA.

    I spent several years on our farm experimenting with reasonably efficient and "invisible" antennas before we retired. I spent my early engineering years doing 1/48 scale antenna modeling over assumed 1/48th scale environments on large turntables. Just like today's many antenna modeling programs, not everything can be assumed and modeled like in the real world. It takes real world measurements. The real proof comes when you can switch between multiple antennas while using programs like PSK Reporter.

    Good luck,
    Ken
     
    W3JJW likes this.
  6. K4DJM

    K4DJM XML Subscriber QRZ Page

    Ahhh. You've just pointed out an error on one of my past posts. When I asserted that the 1/2 wavelength feedline was the worst case, I should have said 1/4 wavelength. Rookie mistake. One half the way around the Smith chart is 1/4 wavelength.

    There is a best length, but there is one very important caveat! I ask --- is the ground at the shack cable entry really ground, or is there, in general some electrical length between the shack entry "ground" and real "ground"?

    Presumably, if the coax braid is 1/2 wavelength long and is connected to honest-to-God ground precisely at that 1/2 wavelength (shack entry?) point, then that voltage antinode (minimum) at that point will also occur 1/2 wavelength away at the feedpoint.

    The electrical length between shack entry ground and true ground is a worry. We have a theoretical answer here, but I do not know that we have a practical one.

    Your sims did not have this problem. An open is an open after all.

    But wait! As you have seen in your sims, a 1/4 wavelength to an open (or any odd multiple) provides the required voltage antinode at the feed. What this means is that it would make sense to place an agressive 1:1 unun isolator 1/4 wavelength away from the feed towards the shack. The remaining length from that point to the shack would not matter. I'm telling Regis "This is my final answer!". If you use the antenna on other bands -- all bets are off though!

    I need to ask: "Are you trying to implement an EFHW without the 49:1?". I sure hope not. I'm telling you that device is necessary for an efficient system.

    I think with the high-Z feed, capacitance to ground from the coax shield really does fix the counterpoise problem. I am running 100 feet (not the magic number) of coax from the transformer to the shack, and I have no RF problems in the shack running 1KW PEP.

    73,
    Don
     
  7. K4DJM

    K4DJM XML Subscriber QRZ Page

    Agree with the points in your post. I just wanted to add to the comment above that the EFHW 40m antenna has band coverage gaps precisely for the reason you describe. Folks that deploy that shorter antenna are often puzzled by the fact that the LONGER (80m) antenna is better at covering some of the higher bands.

    73,
    Don
     
  8. W5DXP

    W5DXP Ham Member QRZ Page

    :) Heck, I thought that you believed 1/2WL was worst case and I was trying to change your mind.

    http://www.w5dxp.com/OCFDLLL/OCFDLLL.htm

    Nope, personally I prefer balanced antennas. For what I was illustrating using EZNEC, no transformer is needed. EZNEC just forces 100 watts into the system at the feedpoint. But I would like to know the typical efficiency of a 49:1 transmission line transformer.
     
  9. 2E0VSS

    2E0VSS Ham Member QRZ Page

    This keeps cropping up re efficiency of these 49:1 transformers, i know that one of the sharpest ham minds on the planet VKOMD, Owen Duffy seems less than enthusiastic for sure..I also note that... "Generally" speaking most of the elmers who contribute on these pages use balanced fed systems....Why?
    More input gents on this particular thread would be very interesting re the 49:1 xfmr

    2e0vss
     
    N0TZU likes this.
  10. N0TZU

    N0TZU Platinum Subscriber Platinum Subscriber QRZ Page

    I think you meant VK2OMD (formerly VK1OD, and his old website is gone too. BTW, he seems to follow QRZ discussions like this and occasionally comments on his blog).

    http://owenduffy.net/blog/

    Another good source of antenna and related information is Charles Rauch, W8JI.

    https://www.w8ji.com
     
    Last edited: Apr 1, 2018

Share This Page

ad: GNOHF-1