EZNEC Antenna Modeling for All of Us

I missed this while it was happening. I was looking to update my old software to model my new antenna configurations and found out about the change.

As Roy W7EL decided to retire, he made public his EZNEC antenna modeling software (compiled form only).


There is also some neat stuff on NEC-5 end, where there is a collaborative update to the latest version (NEC5 X11) thanks to some great ham inputs! https://eznec.com/NEC-5.htm

The AC6LA add-on AutoEZ is something I plan to check out. https://eznec.com/AutoEZ.htm and https://ac6la.com/

AutoEZ is interesting as it an an Add-On to EZNEC which in also in essence an Add-On to NEC.  AutoEZ appears Excel-based, EZNEC is Visual Basic 6 (obsolete version) with some Fortran modules,  and the core NEC compiled Fortran (Intel Visual Fortran with the mkl math libraries).

Below the “–” line I will include some background from either Roy W7EL (website archived copy or QRZ.com) or a QSO Today Podcast interview he did with Eric 4Z1UG.

Well worth checking out all of these programs.

Also a series of dedicated books by ON5AU that should be considered if you are going to put EZNEC to work:



GL and 73




On January 1, 2022 I will be retiring. I’ll be nearly 76 and want to spend more time at other things. EZNEC is and always has been developed, sold, and supported only by me, so all development, sales, and support will end at that time. For several reasons there is no practical way to sell or transfer ownership to someone else, so here’s what will happen on that date:

EZNEC will be released to the public domain and become free of cost and can be freely copied and distributed. I do not plan to release the source code. Click here for more information.

There will be one or two EZNEC program types:

EZNEC Pro. This will have all the features of the current EZNEC Pro/2 (internal NEC-2 calculating engine) and will additionally be able to run external compiled NEC-4.2 and NEC-5 programs furnished by the user. Please keep in mind that NEC-4.2 and NEC-5 are copyrighted software that must be purchased elsewhere. At the time of this writing, the sole seller is Lawrence Livermore National Laboratory (LLNL) but they are in the process of transferring this responsibility to some other person or organization. Note that external NEC-4.2 and NEC-5 programs won’t have the full capability of EZNEC Pro/4 with its internal enhanced NEC-4.2 calculating engine, and the LLNL-supplied NEC-4.2 may be very much slower.

EZNEC+. EZNEC Pro uses some advanced memory management methods to accommodate its upper limit of 45,000 segments, and this may cause trouble on some systems. So I might also make available EZNEC+ which has simpler memory handling. However, I’ll investigate the possibility of making this unnecessary.

No programs will require a hardware key or other additional device.

EZNEC Pro/4 with its internal enhanced NEC-4.2 calculating engine will no longer be available for sale or free or as an upgrade from v. 5.0 or earlier. This is because sale of this program requires my verification that a user has purchased an NEC-4.2 license before sending EZNEC Pro/4. I will no longer be doing this. It will also not be available to replace a lost or damaged program, so users must be responsible for maintaining robust backups. The requirement for a hardware key to use EZNEC Pro programs (EZNEC Pro/2 and EZNEC Pro/4) will be removed very soon and announced on the main EZNEC web site main page and elsewhere.

Until the retirement date, EZNEC sales and support will continue as before. After that date, no support will be be available, and any email or other communications regarding EZNEC support will not be answered. NO REFUNDS WILL BE GIVEN FOR ANY REASON AFTER THE RETIREMENT DATE.

Anyone using the public domain programs will do so at their own risk and with no possibility of support or recourse.

At the time of this writing, the compiled NEC-4.2 from LLNL has been found to contain at least one bug (which has been repaired in the current EZNEC Pro/4), and the recently released NEC-5 has a substantial number. There is currently no one at LLNL able to fix these and no one outside LLNL authorized to distribute updated versions, so anyone purchasing those programs and using them with the public domain EZNEC Pro program will encounter the bugs and may not have any way to get them fixed.

Very soon after starting my own business I discovered a number of things about human nature I hadn’t previously realized. One is the following: If you were able to buy a brand new car for $500 you’d be absolutely thrilled and feel that you had gotten a terrific bargain. But if you discovered that your neighbor had gotten the same car for $300, those feelings would immediately switch to convictions that you’d gotten a lousy deal and been cheated. That’s my greatest hesitation about releasing EZNEC to the public domain and making it available for free. Rather than expressions of delight from people who will be able to use this software for free, I fear I’ll hear (but sorry, I’ll ignore) loud complaints from some people who previously thought that EZNEC was worth at least what they’d paid for it (and many have told me they feel it’s worth much more) but like the lucky car purchaser now feel they’ve been robbed because they could have gotten it free if they’d just waited or if I’d made it free a little earlier. Ironically, those people would be happier if I passed it to someone else to continue selling it or simply made it unavailable altogether when I retire. In fact, the only reason for the six month delay until retirement is to give potential new customers fair warning of the choice they can have of buying it now or waiting and getting it free later.

I sold my first copy of MININEC-based ELNEC v. 1.0 in March 1990 and two more major versions followed. When personal computers became powerful enough to run NEC-2, I introduced EZNEC v. 1.0 in April 1995. Shortly after that I quit my “day job” as principal design engineer at Tektronix, and EZNEC became my primary way of making a living which it has been ever since. The first Windows version, 3.0, came in May 2000, and versions 4.0 – 6.0 have been introduced since. My customer database which goes back to the first ELNEC sale in 1990 has nearly 13,000 different names and companies. A sizeable fraction of these are current users.

You, my customers, have been wonderful. I’ve enjoyed interacting with you and learned a great deal by doing so. You’ve been courteous, thoughtful, patient, and honest beyond all expectations. You’ve enabled me to escape the Dilbert cartoon and have the dream job of being a one person home-based company for 26 years, with no bosses and no employees. It hasn’t brought me more than a modest income, but getting rich has never been my goal — along with occasional consulting it has provided an adequate income and a much, much, more pleasant and stress-free life than a career at a large corporation (or even most small companies). I hope that what I’ve been able to provide has been a fair exchange.

This announcement may be updated from time to time. Last update July 7, 2021.

Roy Lewallen, W7EL

From QRZ.com – various threads:

Thank you for all the kind words and suggestions.

I agree that an open source-based universal version of EZNEC would be great. But there’s a problem.

The first Windows version of EZNEC was written beginning in about 1998 when Visual Basic was the current and logical choice for a Windows language, and EZNEC was subsequently upgraded using that language. By the time Microsoft discontinued support for Visual Basic in 2008, EZNEC/EZNEC+ had undergone two major upgrades and consisted of more than 71,000 lines of code and EZNEC Pro more than 76,000 lines — not counting blank and comment lines, or the Fortran-compiled NEC based calculating engines. Translation from Visual Basic 6.0 (VB6) to its successor Visual Basic Net or any other language, open source or not, would require almost a complete re-write of the program. (Although VB Net has a similar name and some commonality to VB6, it’s different enough to prevent any kind of even semi-automated translation let alone direct copying.)

v. 6.0.33 of EZNEC/EZNEC+ now consists of more than 79,000 lines of code and EZNEC Pro more than 79,000. Worse yet, the program has evolved over three decades and contains some code from the days when memory and computing time were precious, so there are portions which aren’t nearly as modular and compartmentalized as more recent tools and systems allow us to be.

Of course I’ve thought a number of times about updating EZNEC to another language but estimate that it would take about two years of full time work, at the end of which would be a program which the user couldn’t distinguish from the original — except for the massive number of bugs which would inevitably result from the translation and the poor support I’d be able to provide while spending the time programming instead.

If it was attempted, the logical way to go about a project like this would be to identify blocks of code which could be independently translated and put into external libraries where the re-written functions could be called and tested as they’re developed. As far as I’ve been able to tell, compatible libraries can’t be created with Visual Basic Net, at least not easily, although they can be made with some other languages. (EZNEC uses a number of libraries I wrote in Fortran for calculation speed.) After creating a library, EZNEC would have to be modified to call its functions and then be recompiled — but the VB6 compiler hasn’t been available for many years now. (Those of us who have it can still coax it to run under Windows up to and including 10 but with a few quirks.)

The calculating engines are easier. They’re in Fortran which has evolved to become a very modern and structured language, easily integrated into Microsoft Visual Studio. To recompile the calculating engines you’d need Intel Visual Fortran with its mkl math libraries. I think this is going for around $1400 now. But let me tell you, NEC-2 and NEC-4 (even -4.2) code is written in 1970s style and very, very tangled and hard to follow and work with.

So, any volunteers to make an open source version of EZNEC? Don’t all speak at once!

Roy, W7EL

Thank you all again.

There seems to be some misunderstanding about the structure of EZNEC. EZNEC isn’t a rewritten version of NEC. It’s primarily a human interface which enhances the usability of NEC-2 and (EZNEC Pro/4) of NEC-4.2. The program sizes quoted earlier are for just the interface which is in addition to the NEC calculating code itself which has been extensively enhanced and made much faster. I refer to the modified NEC code as EZNEC’s “calculating engines”, and each of those involves several thousand lines of code. They’re in Fortran, compiled with a modern compiler (Intel Visual Fortran). The NEC-2 engines are extensively modified from NEC-2 and don’t use the original input and output interfaces. The NEC-4.2 engines aren’t as highly modified but are still substantially different. For just one example, the NEC-2 and NEC-4 input format limits the precision of the numbers sent in and the total length of each line. Calculations using NEC networks such as EZNEC’s transformers, lossy transmission lines, and parallel connected sources benefit greatly from beginning with fully double precision values. So EZNEC sends those and some other inputs to the calculating engines as double precision numbers via separate binary files. This maintains the accuracy of the resulting calculations over a wider range of user specified object values. The free versions of EZNEC to come will be able to run external NEC-4.2 and NEC-5 programs but that higher accuracy can’t be obtained with them due to the limitations of the standard NEC input format. There are many other examples of NEC modifications made in the calculating engines to improve operation such as adding numerous features not present in NEC-2 including current sources, automatic array dimensioning, MININEC type ground, wire insulation, catenary wires, and others. The first two weren’t in NEC-4 but appeared in NEC-4.2, and the third in NEC-5 (although it doesn’t work correctly and NEC-5 doesn’t include wire insulation).

To answer UA3TW’s question — no, I have never shared the EZNEC source code with anyone so by definition it is not “open source”. Its confidentiality has been essential to my being able to make a living from it. Also, I might add, something else that has been essential is the exceptional honesty of EZNEC users. If EZNEC had been copied to any extent (“pirated”) to avoid paying for it, I would have been out of business a very long time ago. Many times I’ve received an order, with payment, for a program to be sent to a friend which illustrates the honesty and fairness with which I’ve been treated. Thank you all, very much! I’ve done my best to respond in kind.

Roy, W7EL


Roy, you’re best known for your creation of the EZNEC antenna modeling software. How did you come to create it?

Roy, W7EL:
First of all, I’ve had an interest in antennas ever since my early days of ham radio. It’s just something that’s fascinated me. Once I got an engineering education, I was able to better understand what was going on, although it’s still very difficult to get a really deep understanding. I spent many, many hours researching in the wonderful Tektronix library. Lot of papers from the IRE and the IEEE about antennas and trying to learn about them and educate myself about them. I wrote little computer programs for my HP-11 calculator. I resisted getting a computer for a long time. Not really having any justifiable use for one.

Eventually a neighbor ham, a fellow by the name of Jim Larson, I’m afraid I don’t recall his call, gave me a floppy disk and it had many NEC3 on it which was a public domain program written by Rockwell and I’m embarrassed here, Logan. Logan and Rockwell. It sat there for quite a long time, but finally I got a computer. My first computer was a XT machine. I loaded this thing in there and I started fiddling with it and holy mackerel, the power of this program as simple as it was absolutely amazing. Up until I used that program, all of these … In order to even estimate the impedance of a dipole over a wide range of frequencies with various diameters of wire, there are a good half a dozen approximation methods that have been developed to do that.

I’ve got a file an inch thick or inch and a half thick in my file cabinet of papers on people who attacked this problem. Then we start taking two antennas that are parallel with each other in the same length. Then you start getting into some pretty complex math. If you tilt one a little bit relative to the other, it’s a problem becomes almost impossible to solve. It’s because of the integral equations that are necessary that can’t be solved in close form. Anyway, this little program would solve those problems and do it very quickly. That was just amazing. I started hacking on it and it was user hostile. If you wanted to make any change, you would have to start all over again. It wouldn’t remember anything.

There was no graphics and so forth. It was written in GW-BASIC which I was acquainted with. I started hacking on it, adding graphical stuff and made it possible to save and recall antenna designs and so forth. It needed current sources so I figured out a way to incorporate current sources and on and on and just kind of kept working on it to make it more usable for myself. I had no thought of selling it. Then somewhere along the line I became aware of Brian Beasley selling his software. What he had done essentially is the same thing. I thought gee, mine will do everything his will and a bit more. I would sure like to get enough money to get a new computer because this thing is kind of sole for doing this kind of work.

I got a compiler and bought a couple of months of advertising in CW and QST. My first program. It was called ELNEC. I sold my first one in 1990, early 1990. 1995 I finally got fed up with living in the double work cartoon and I quit my day job. At that time I had EZNEC which was based on the NEC2 code which is also a public domain. I’ve been living off of that, plus a bit of consulting ever since and I’ve never looked back. That’s the story in a nutshell of EZNEC.

Eric, 4Z1UG:
What excites you the most about what’s happening in ham radio now?

Roy, W7EL:
I’m not very excited about what’s happening in ham radio now. This is something that I can spend a little bit of time talking about, but it’s just my personal opinions. The ham radio that I knew no longer exist. When I was young, ham radio was just magic. I could listen to this radio and I could hear all kinds of interesting stuff from far away. Later on I could build rigs to talk to these people far away which really I could not do in any other way. I could not meet people from other countries and on and on. Nowadays of course anybody can pick up a cellphone and dial an international number right there standing in wherever. He is talking to someone around the world and thinking nothing of it.

The magic of ham radio … I used to go down to the city dump that was back in the days when they would take all the trash and dump it in a big pile out over along the shore of the Gastineau Channel. In the daytime I would go out there and collect components out of discarded TV sets to use in my home brew rigs. Then at night I would tape a flashlight to my rifle and go out and shoot rats. Neither one of those things is really possible anymore. I guess to some extent that’s good. People who were entering ham radio are doing it for entirely different reasons. I’m not too sure what all of those reasons are. They’re different to the ones that attracted me to ham radio though because like I said the reasons that attracted me to ham radio no longer exist.

I’m a pilot. I encountered on a number of the pilot forums people who mentioned getting a ham radio license simply so they could use the APRS, the position reporting service, to put in their airplanes. That’s the sole reason that they got an amateur license. I think we probably have some people that are interested in doing public service or people who are just interested in chatting with people on two meters or whatever. I think those are the reasons that a lot of the people are entering ham radio today. It’s becoming a very, very different hobby than what attracted me. I won’t say it’s a bad thing, but it certainly is different and changing. One thing I have found very encouraging, very interesting is that every year I go to the events at Dayton sponsored by the QRP ARCI.

They have a day of seminars and other events preceding Dayton and during the Dayton Convention. They call it the Four Days In May, FDIM. One of the things they usually have is a show and tell or a home brew contest where people bring their home brew projects and show them off. I’ve really been impressed with the number and quality of the things that people have been designing and building. It’s really great. Some are just absolutely marvelous craftsmanship. Some of them are extremely clever design. As a professional design engineer, I can appreciate that. The number of people attending FDIM seems to be growing every year. Unfortunately I think it’s still very heavily dominated by older people.

I don’t see nearly as many young people as I would like to, but I think there are some young people that are coming up and getting interested in that. It seems like QRP is one of the last holdouts of the home brew portions of Amateur Radio that I really love. The other fairly encouraging thing is boy, there sure a lot of people filling with their antennas. Antenna seems to be one of the other last things that people are playing with and doing some experimenting and building of. My EZNEC sales just keep going up as more people get computers and decide to try it. That’s really encouraging from the standpoint of the future of ham radio.

Eric, 4Z1UG:
What advice would you give to new or returning hams?

Roy, W7EL:
Well, as I said, the hobby is quite a bit different than what it was when I started out, but I do have just a couple of words of advice. The first one is to join the ARRL. If you’re in another country, join the equivalent organization in your country. The reason is if no other reason, they are our lobbying group. They’re really the only thing that stand between us and a whole bunch of restrictions and other things that would really harm our hobby. I think we should support the ARRL if for no other reason for that one. It is also a great resource for a large number of things of interest to all amateurs. They sponsor contests and they sponsor a lot of other events that interest a lot of hams. I highly recommend that.

The second thing is to find a group of people that have the same interest as you do. Ham radio is a very diverse hobby. There are people like me. There’s a geezer crowd CW die hards. There are people who do QRP home brewing. There are people who like to play with microwave stuff or digital modes. People who just like to chat on two meters. There are people who like to contest. Work DX. There’s just a lot of stuff. Find a group that specializes in that where a lot of the members share that same interest and then connect with them. I think you will find that that greatly increases the fun you’ll have out of ham radio. If you go to Dayton or a similar kind of a great big convention, you’ll find that each one of the groups has its own hotel.

Find which group has the interest that are those of you and join them and stay with them and meet some people that share your interest. Those are my just two pieces of advice.

Tagged , , , , , , , ,

3 thoughts on “EZNEC Antenna Modeling for All of Us

  1. Kuby, MS-EET [Ham for 50yrs] N6JSX says:

    About four decades ago I got into antenna designing due to a very close friend, fellow Veteran and T-Hunt mentor/Elmer went SK 1985, from cancer; Ray/TEY asked me to carry on his 2m Quad design that he perfected over many years. I took his request and eventually published it in ’73’ magazine. Ray’s uniqueness was in direct-feeding coax to the Quad driver (using NO extra components, capacitors or beads that would consume RF energy).

    I was seeking a computer antenna modeling program to validate and see the cause-n-affect to VHF/UHF/SHF antenna designs. I had conversations with W4RNL/SK (considered the guru of HAM Antenna Modeling back then – at least the ARRL loved him). W4RNL used many NEC based programs but most often used NEC-Win-Plus. I tried Nec2Go, MultiNEX, MiniNEC, EZNEC, Manna-Gal and even DOS YagiMAX (I still use).

    Do you notice the common theme here, “NEC”, these Model programs are shells over the NEC-2 (public domain software) engine. NEC – Numerical Electromagnetic Code from Lawrence Livermore Lab funded by the USN & USA/USAF, 1977.

    One of the biggest issues I ran into using my 2m J antenna design across three of the most popular HAM Modeling programs including EZNEC. First, each modeling program required a different method in entering the physical antenna data – the output results of all three NEC’s were different. I sent my data file to these program Modelers asking them if I entered the data/parameters correctly? All replied positively but one of the Modelers stated their program was based off NEC that is more for HF not VHF.
    After these answers I made another query, asking these questions;
    1. How did they ‘prove’ their Model was correct?
    2. Did they ‘verify’ their HF, VHF, UHF Model predictions?
    3. Have they ever physically made a HF/VHF/UHF modeled antenna from their program AND then tested their antenna on a test-range to verify their Model’s output?

    Only in person or telephone calls did they disclose, “NO” to all my questions due to test-range time cost money (and HF antennas are too large for most ranges). I countered that a 2m beam or 2m J is not too large and would indicate via proportionality if your algorithms are close to being accurate – I got no response.

    So the moral to my search for a sound Antenna Modeling program is (as I suspected) Modeling programs output are unverified mathematical predictions yielding reasonable guesses.

    As a lifetime Electronics Test Engineer (and even a GPS Antenna designer for Magellan working with a test range in Oxnard, CA creating our Quadrafiler-helix) we soon discovered even the the NEC-2 Model will not output optimal data just near predictions. Only by tweaking the physical design and retesting did we find an optimal performing antenna. AND VSWR is NOT the singular most significant factor to an antenna – radiation patterns are.

    I submit that the tweaking we had to do was to counter the variations in the test-range; ground, coax, structures/reflections and RF signal losses (before digitizing) all had significant effects. The HAM Modeling programs often ASSume a perfect ground, perfect elevation, perfect mounting, etc – ‘an isotropic environment’. With virtually NO input parameters for environment variations IF you could measure them via a controlled method.

    Now the real trick is correlating the ‘tweaks’ back into the NEC Model algorithm (if you can get into the program to make such tweaks). And since the writers of all these antenna modeling programs charged a fee for their program, I take the attitude, it is incumbent upon these Modelers to insure they are selling a quality product. Please point me to any analysis of a HAM level antenna Modeling program, where its outputs match test-range measurements (without algorithm manipulations).

    BTW, I returned EZNEC 5.0 back for full refund since a month after I bought it he released EZNEC 6.0 and would only upgrade for an additional fee (that was half of the cost of the original V5). About this time the ARRL Modeling course was dumped due to an issue with the course based EZNEC V5 availability vs. V6 changes.

    Now that EZNEC is free – give it a try but beware it may not give you the results expected.

    Now retired I’m perfecting a RF-rainbow-viewing glasses that let us see emitted antenna radiation patterns.

    • k9zw says:

      Some updates to consider:

      Latest NEC version available to the public is the NEC-5.2 x13, released this Spring, but level 5 does require a license.

      Private personal license for the latest level is about $110

      The lower level that comes with EZNEC doesn’t have a separate NEC licensing fee.

      Few of the old modeling programs are being updated. Not aware of much that isn’t a wrap around for a NEC engine.

      None of the predictive models will fully simulate real world performance. Test Ranges may differ enough from achieved performance to a difficult measurement to stand behind.

      Most of these models are perhaps more useful to compare the differences between various configurations rather than predicting absolute values. The Delta between options looks to be the most important thing you can take away.

      Outside of certain well heeled organizations the cost to take an HF antenna through a proper HF Test Range analysis unrecoverable, so hence we model.



  2. […] offered gratis to the Amateur Radio community had some comments, and I received a few emails. (Link https://k9zw.wordpress.com/2022/05/14/eznec-antenna-modeling-for-all-of-us/ […]

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: