A Short Study of NVIS
By Warren Paulson, VE3FYN

Go to www.ve3fyn.ca/files for the EZNEC source files used for this article.

When I started building my radio shack in Atikokan, I wanted a stable way of communicating with nearby Thunder Bay, Dryden, Sioux Narrows and Kenora. These communities are all in the 200 kilometre range. That's where many of my friends are, and those are the communities I would want to contact in the case of an emergency. Even Winnipeg and North Bay (about 500 kilometres) were generally out of reach of my vertical HF antenna - I needed a way to connect to these places beyond VHF, beyond the groundwave, and within the skip zone.

Enter NVIS, or "near vertical incident sky-wave." Simply put, this system optimizes HF antennas for a near-vertical takeoff. What goes almost straight up comes almost straight down, reaching regional targets.

I read many articles on NVIS before I started. Some were very good, but many were contradictory. Some, for instance, advocated antennas seven feet above ground, while others recommended I shoot for 90 feet.I decided to do some modeling using EZNEC. I discovered what works best (at least in theory) if you have an infinite yard and a tolerate spouse, and what works best in some more constrained situations.

Here's what I found

1. NVIS only works at frequencies from 2Mhz to 10Mhz. The signal must penetrate the D layer of the ionosphere, and bounce off the F layer. Lower-frequency signals will not penetrate the D layer; higher frequencies will not bounce off the F layer at these sharp angles. (Remember the Maximum Useable Frequency, or MUF?) So, for us amateurs, we're looking at 40 metres and 80 metres primarily, and also 160 metres.
   
   
2. NVIS requires horizontal (flattop) dipoles. Verticals are out of the question for NVIS - they cannot be configured to provide the vertical takeoff angles. Variations, like vee antennas may provide some advantages, as described later.
   
   
3. The optimal height for an NVIS antenna is 1/8 of the wavelength. At 80 metres, this is about 32 feet off the ground. This height produces a steep takeoff angle, minimal ground losses, and an omnidirectional pattern. Below 30 feet elevation, the pattern remains the same, but the dipole experiences significant ground losses. Above 1/8 of a wavelength, the takeoff angle is reduced and there is a significant overhead null with strong DX-style sidelobes.Under ideal conditions, gains of 7dBi at the near vertical angles are possible. Get the antenna too close to the ground, and signal loss due to ground absorption will occur.
   
   

   	This plot shows the increasing vertical gain as the antenna moves from 5 feet (0.019 wavelengths) to 40 feet (0.15 wavelengths) above ground.
   	This plot shows the degradation of the vertical signal after 40 feet. The 60 foot plot (blue) is nearly identical, but by 90 feet, it is significantly attenuated.

   
   
4. A good NVIS antenna will not work well at DX distances. Antenna gain is a zero sum game. There is a fixed amount of energy radiating. If we push it all out in one direction (the near-vertical angles), we have to take it away from another direction (the low DX angles). If you want to NVIS and DX on 40 and 80 metres, you will need a few antennas.
   
   
5. During sunspot highs, 80 metres works well at night. However, as morning progresses, the 80m signal will become absorbed by the strengthening D layer. Forty metres can still penetrate the D layer, and works well during the day. During the current sunspot lull, however, 80 metres works well during the day, and 160 takes over at night..
   
   
6. The NVIS signal is virtually omnidirectional, so it doesn't matter how you orient your dipole. Dipoles only exhibit directionality once they reach 1/2 wavelengths above ground.
   
   

   This azimuth view of an NVIS dipole at the 70 degree point shows only 0.76 dB attenuation from dBmax off the ends of the antenna. This antenna is a dipole 32 feet (1/8 wavelength) off the ground, and tuned to 80 metres.

   
   
   
7. Up to about 50 kilometres, ground wave becomes a factor. If a nearby station receives your groundwave signal, and the skywave (which has travelled perhaps ten times further) the station will experience multi-path distortion. So your antenna design should minimize any near-horizontal signals. Still, you may find a vertical antenna with a low take-off angle better for very short hauls. (The January 2006 issue of QST has a very good article on groundwave that uses real data to show that groundwave propogation is far shorter than many of us thought.)
   
   
8. The F layer varies from 160 to 480 kilometres up. Do a little trigonometry to figure out the take-off angle for your target destinations. For example, Thunder Bay is 175 kilometres from Atikokan. The takeoff angle requied is roughly from 60 to 80 degrees. This is important when modeling, as it allows you to pick a design optimized for your target. .
   
   

   Dst (km)	50	75	100	125	150	175	200	225	250	275	300	400	500
   Low (°)	81	77	73	69	65	61	58	55	52	49	47	39	33
   High (°)	87	86	84	83	81	80	78	77	75	74	73	67	62

   
   
9. The improved gain you will get from adding a reflector is inversely proportional to the height of your antenna. In other words, the higher your antenna, the less point in adding reflectors. A 40 metre dipole seven feet off the ground can achieve an additional 2.55 dB gain with the use of a reflector at ground level. Multiple reflectors increases the gain furhter. But a reflector positioned at the optimal height of about 2 feet above ground, below a 20 foot high dipole only increases the gain by 0.81 dB -- hardly worth the effort.
   
   Even though great gains can be made at lower heights, the reflectors still won't fully offset the loss in gain from lowering the antenna in the first place. The best gain from the seven foot high antenna is 3.31 dBi, far lower than a 16 foot high dipole with no reflectors (5.31).
   
   The moral: Get your antenna as high as you can, up to the limits described earlier. Reflectors will improve gain, but may only be worth it if you are forced to place your antenna very close to the ground.
   
   

   	The primary plot is a 20' high 40 metre dipole with no reflector. Vertical gain = 6.03 dBi Ground reflector = +0.38 dB Optimal (2') reflector = +0.81dB
   	The primary plot is a 16' high 40 metre dipole with no reflector. Vertical gain = 5.31 dBi Ground reflector = +0.62 dB Optimal (2') reflector = +1.34 dB
   	The primary plot is a 7' high 40 metre dipole with no reflector. Vertical gain = 0.72 dBi Ground reflector = +2.55 dB Two ground reflectors +/- 6' from centre = +3.31 dB

   
   The models above were applied to 40 metre dipoles. As you would expect, the same principles apply to 80 metre antennas. In fact the increased gain of a low-level 80 metre dipole is even greater (because in terms of wavelength, it is closer to the ground at seven feet). The model below shows an increase in gain of 3.71 dB by applying three ground-level reflectors (at 0, +/- three feet from centre) to a seven-foot high 80 metre dipole. But remember, the overall gain of this antenna is still low: -2.24 dBi without the reflector, and 1.47 dBi with.
   
   

   This plot shows the advantage of a reflector at low elevations. The blue plot (80M_7ft) shows an 80 metre dipole seven feet from the ground. The primary plot shows the same antenna with three reflectors placed on the ground. One is under the dipole, and the other two are each three feet to a side. The gain is 3.71dB - enough to warrant running the wires.

   
   

What to do when space is tight

Wire dipoles for HF are long. And, as we've demonstrated, even for vertical skywave applications, they're high. Yard sizes, finances, family and neighbour relations often inhibit our ability to erect the dream antenna In my case, my yard is only 110 feet long. That posed a problem in trying to erect a 123 foot long antenna. Further, I'd imposed on myself a height restriction of 24 feet, assuming that an antenna much higher than that would elicit comments from the neighbours. I also wanted in my antenna to avoid the use of an antenna tuner outside that of my HF rig. So, the SWR in my antennas had to remain within 3:1 across the band. So, cutting the wire and using a random wire tuner was not an option.

That basically left me with three options for shortening the antenna:

1. Shape it into a Vee.
2. Shape it into an inverted Vee.
3. Run it as a regular flattop dipole, but drop the ends straight down, in my case at the 55 foot mark from centre.

Again, antenna modeling helped. I modeled each of the three options to assess their impact on vertical and horizontal gain.

This plot compares a full-length 123' dipole at 24 feet elevation (primary), with two other options: 80m_drop-ends_24ft is 24 feet high, but only 108 feet long. The ends drop straight down 7.5 feet on each side. 80m20deginvVee_24ft is 24 feet high in the centre, and only 3.3 feet at the ends, shortening it to 115.6' overall.

Dropping the ends is clearly the best option for several reasons:

1. The loss in vertical gain is minimal - 0.73 dB from a full-length dipole of the same height. (The inverted vee option loses 3.56 dB gain.)
2. The inverted vee option shows some gain at the lower angles. At the 18 degree mark, the inverted vee shows 2.83 dB gain over the other two options.
3. The drop-ends option shortens the antenna more while still keeping it safely off the ground. In this model, the drop-ends option is 108 feet long, and the ends are 16.5 feet high - well out of reach of kids. The inverted vee option, on the other hand, is still over 115 feet long, and the ends are a dangerous 3.3 feet off the ground. (The ends of the wire will show low current, but high voltage.)

These results make sense. Most of the current is in the centre of the antenna. Keeping the centre section high off the ground will give us the greatest gain. Dropping the ends allows us to maintain SWR without sacrificing gain, as little is transmitted from the ends in any case.

Now, if you need to shorten your antenna, but overall height is not an issue, a Vee antenna may make sense.

In this model, the primary plot is a dipole at 51 feet high. At vertical, it shows 1.77 dB gain more than the same dipole at 24 feet. However, a Vee antenna 24 feet at the centre and 51 feet at the ends shows virtually the same gain at the high angles, and less gain at the low elevations (about 2 dB less at 20 degrees).

This vee antenna retains the gain of a dipole at the high angles, and reduces it slightly at the low angles. This antenna has the advantage of reducing multi-path distortion somewhat, but having a weaker groundwave signal. It also requires less feedline. (Saving 25 feet of RG 213 cable will reduce your line loss by a whopping 0.1 dB.)

So, what did I build?

In spite of the modelling, I found that it was far easier to match SWR by running an inverted vee. For the 80 metre antenna, I ran an inverted vee from 24 feet high at the centre, to eight feet at the ends. Due to my small yard, I still had to run the ends of the wire straight down. The 40 metre antenna is a true inverted vee four feet below the larger antenna.

Both antennas cover their full band well within my required 3:1 SWR. In fact, I found that the 40 metre wire will tune nicely to 80 metres, and to 30 metres. You may get by with just the one.

This shows the antenna design: An 80 metre inverted vee is installed 24 feet at the centre, and 8 feet at the ends. In the actual installation, the last four feet drop straight down. A 40 metre inverted vee is at 20 feet.

Does it work? It works very well for communications within 250 kilometres, and will drive a decent signal up to 500 kilometres. At that point, you can swith to your vertical, or a beam. If you take the time to model it, this is a highly efficient and extremely cheap antenna project. And it fills the gap between local and DX communications.

... 73, de VE3FYN

Last update: 24-Mar-2009 3:05 AM
Web page by: Warren Paulson