CENTRE FED MULTIBAND HF DIPOLES - G5RV DERIVATIVES
EASY AND SIMPLE HF WIRE ANTENNAS THAT WORK
If you don't feel like reading the rubbish on this page - try the short version.
INTRODUCTION - THIS PAGE
The best description for the HF antennas this page details is probably just a "centre fed dipole." Someone else who built one and found it works gave it the name "non-resonant twin wire". It is not a G5RV because thats a special case with a given length of radiator and feeder. These antennas can be set up as a straight suspended wire or erected in inverted V format. The principle is the same.
I often wonder why anyone would buy such an antenna. They are so cheap and easy to build there seems to be no sense at all in not making your own homebrew version. Amateur radio operators should be able to build their own.
THE CONCEPT AND ORIGINAL G5RV
The basic principle of a G5RV is that, an antenna nicely resonant on one frequency/band will not be on another. A centre fed (or offset) dipole either needs traps to make it reasonant on each required band or, in the case of the G5RV, a compromise can be struck where it isn't so good on any band but tunable on all. The concept requires an antenna tuner and, because the SWRs can be high, balanced feeder is essential. (It doesn't matter what kind.) With high SWRs, the losses in coax can be high. Because of the opposing signals, balanced feeder can't radiate anything significant. You can be sure that anything that makes it out of the tuner is radiated with only small amounts going into heating wires and so forth. IT IS A MYTH THAT BALANCED FEEDER RADIATES A LOT BECAUSE IT HASN'T GOT A SHIELD.
THE SWR ON THE RADIO SIDE OF THE TUNER MIGHT BE 1:1 BUT, NOT ON THE ANTENNA SIDE. THAT'S WHAT A TUNER DOES. The tuner's main job is to balance out reactive components (and matching impedances in so doing) making the system look purely resistive. The reactive components still exist on the antenna side of the tuner.
Everything which carries an Alternating Current of any kind radiates something. The only way nothing is radiated from balanced feeder is if both wires occupy the same physical space which is, of course, impossible but the opposing signals mean the amount of signal radiated is so small as to be negligible. The amount radiated over the entire length increases as the SWR increases.
An ideal antenna needs to have a current antinode somewhere along its length. It is the point of maximum current and the point of maximum radiation. Given there is always a voltage antinode at the ends of the wire and that the current antinode will be ¼ wavelength away, the ideal centre fed dipole is ¼ wavelength on either side with the current antinode at the centre.
If you decrease the length of the radiator, an equivalent increase in the electrical length of the feeder is also required.
The trouble is, a quarter wave on 40 metres is a half wave at 20m. An exact multiple of a half wave length from the reflection point (end of the wire) to the radio, means the SWR will be so high as to be almost impossible to tune.
If only 80m and 40m were considered, the idea is to make an antenna half way between 20m (¼ of 80m) and 10m (¼ of 40) ie. 15 metres, or thereabouts, long. The SWR will be well over acceptable levels without a tuner but will be tunable on both bands without too much effort. The same antenna will work well on 20m (¾ wave either side) but we get into trouble on 15m because it is 1 wavelength long on both sides. Multiples of a half wave length can have SWRs so high as to be untunable.
The actual wavelength of 21MHz is 14.27 metres so adding the extra over 15 metres means the original G5RV will have 1.27 metres making it 1⅛ wavelengths long at the bottom of the band leaving it still tunable with a good tuner. The same problem exists on 10m where the antenna is near 1½ wavelengths long but again different enough to tune. 10m seems to be the frequency most have trouble tuning and some say this antenna won't work on 10m. IT WILL WORK EXCEPTIONALLY WELL ON 10M IF YOU GET IT RIGHT
Throughout this description I have not mentioned the feeder. THIS MUST ALSO BE CONSIDERED in the calculations. It forms a part of the antenna electrical length even though it doesn't radiate. If you add the length of the original G5RV to the feeder described and take a velocity factor of roughly 90% into account, the full electrical length from antenna tuner to the end of the antenna is about 27 metres.
The compromise found by G5RV is only one of many. To find others, a plot of antenna length desirability can be made. The closer an antenna length (including feeder) is to an odd multiple of ¼ wavelength the better. The trick now is to find a length which is close to an odd multiple of ¼ wavelength on all bands of interest.
Suppose a plot is made of the antenna length desirability for one side of a half wave dipole for 80 metres. The perfect length will be either ¼ or ¾ wavelengths whilst the worst are ½ and 1 wavelength.
If 40 metres is now added to this, the most desirable length is now somewhere between.
If we now add 20 and 10 metres, the plot becomes:-
The original G5RV APPROXIMATE electrical length compromise is shown with the asterisk. NOTE, THIS IS THE TOTAL ELECTRICAL LENGTH OF ANTENNA AND FEEDER. The length of the antenna itself is really only relevant in so far as there needs to be a reasonable amount of current antinode along the length of the antenna. If there isn't, the feedpoint impedance may be too high to match no matter what electrical length the feeder is. The above figure and calculations yields the nearest length for the feeder to best match the amount of radiator present. To match the rest, an antenna tuner is used.
On the radio side of the tuner, the SWR can be quite low but the high SWRs still exist on the antenna side of the tuner. For this reason COAX SHOULD NOT BE USED FROM THE TUNER TO THE ANTENNA. For this system to work, you must use balanced feeder simply because there are still very low losses even with very high SWRs.
The above examples assume exact wavelengths for the bands. The actual wavelength of the Australian 80 metre band is about 81 to 85.6 metres and the 40 metre band is from 41.1 to 42.8 metres. Similarly the 20, 15 and 10 metre bands vary significantly from the exact wavelengths. In order to find other tunable compromises, the exact wavelengths need to be plotted. Once both ends of all bands have been included, the compromise figures become much more accute so a good match on one end of a band does not mean the entire band can be tuned easily. ON SOME BANDS, LENGTHS BECOME EXTREMELY CRITICAL.
The fewer the bands and the narrower the band widths, the easier it is to find a compromise. Even better compromises can be found if you eliminate those parts of each band you don't use. If you aren't going to use CW, then don't bother with that part of the band. On 40metres, for example, if you only need to attempt a good tune for SSB use, then only bother with 7.04 to 7.3 and similarly for the other bands.
On the lengths I use, I get a better than 1:5:1 SWR right across the 10m band but need to re-tune at several locations across the 20m band.
GETTING THE LENGTHS RIGHT
It would be useless to describe the exact lengths of the other compromises. Each batch or type of balanced feeder has a different velocity factor so lengths will depend on the type used. Lengths will also depend on which parts of the band you intend to use. YOU WILL HAVE TO FIND YOUR OWN.
All that can be done here is to give several electrical length compromises for 80, 40, 20, 15 and 10 metres under several circumstances. If you measure the lengths of the radiators then calculate the lengths of the feeder for a HIGH velocity factor feeder, the feeder length can then be trimmed 50mm or so at a time until a good match can be obtained on all HF bands. At 20 or 30 cents a metre, a couple of metres thrown in the rubbish shouldn't break your budget.
In any case, the radiator should be as long as possible. If you can't get more than about 10 metres per side, it is a very good idea to hang a metre to a metre and a half at each end. There will be some losses if this is done but you gain more on the lower bands by, figuratively speaking, pulling some of the current anti-node out of the feeder. More than 1.5 metres on each end will increase the losses too much on 10 or 6m.
I CAN ALSO OBTAIN A GOOD MATCH ON 6M BUT THIS BAND HAS BEEN OMITTED FROM THESE CALCULATIONS. Unfortunately, I am limited in space and couldn't erect an antenna large enough to be tunable on 160m without considerable effort.
ELECTRICAL LENGTHS TO ATTEMPT
For 80m, 40m 20m, 15m and 10m using all modes
8.50, 12.70, 17.00, 24.50, 26.00, 29.55 or 33.80
For 80m, 40m 20m, 15m and 10m SSB only
8.45, 12.65, 16.85, 24.85, 27.10, 29.55 or 33.70
For 80m, 40m 20m, 15m, 10m and the warc bands 30m, 17m and 12m
9.25, 12.85, 27.10 (ssb only) 26.00 (all modes) or 34.10
CALCULATING THE LENGTH
Le (electrical length) is calculated by Lo x Vw (overhang length IF USED) + Lr x Vw (length of ONE SIDE of radiator) + Lf x Vf(length of feeder)
Le = (Lo x Vw) + (Lr x Vw) + (Lf x Vf)= Vw x (Lo + Lr) + (Vf x Lf)
El is the overall electrical length
Vw is the velocity factor of the wire (usually about 97% so negligible)
Vf is the velocity factor of the feeder (could be anything from 70% to 90% or more - SEE TEXT)
Lo is the length of overhang on ONE end IF YOU USE ANY AT ALL
Lr is the length of ONE SIDE of the radiator
Lf is the length of the feeder
1) Other than, as long as possible, the radiator length is irrelevant. First decide where you can put it and therefore how long you can make it. It can be bent in the middle and errected as either a straight wire or inverted V format with the ends closer to the ground. Unless it is 20 metres in the air, it will be omni-directional enough no matter where you put it.
2) Work out the minimum physical length of feeder from the antenna's centre to the tuning unit. Calculate it's electrical length using a largish velocity factor then ADD SOME until the length of the radiator (ONE SIDE) and feeder comes to a number a larger than one of the figures given above. RADIATOR LENGTH (ONE SIDE) PLUS FEEDER ELECTRICAL LENGTH SHOULD BE A GOOD COUPLE OF METRES LONGER THAN one of the above numbers and A GOOD COUPLE OF METRES LONGER THAN YOU REQUIRE. If you know and are very confident that you know the velocity factor of the feeder you are using, you can save some time and make the length much closer to that required.
3) Plug it in to the tuner and attempt a tune on the upper and lower bands of your choice. If you can't get a good tune on all bands, CHOP 50MM off the end and try again.
All text and images on this site are Copyright to John Langsford (vk5ajl).
You may provide links on other sites or use the information and pictures for your own personal use.
You may use the text or images for redisplay or quotation provided you acknowledge the source ie. vk5ajl.com.
I think that's pretty fair, don't you?