Friday, 6 February 2009

The Rybakov 806 vertical

I have no idea why this is called a Rybakov 806, but essentially it is a 7.6m vertical fed with a 4:1 Un-Un (unbalanced to unbalanced transformer) as described by IV3SBE.

It works very well with a fibreglass fishing pole and can be put up in a few minutes. The pattern (right) is what you get on 20m.

You need to put down an earth stake, and it really, really needs radials – the more the merrier.

The idea is that the antenna represents a non 1:1 SWR match at all frequencies – 7.6m is chosen as it isn’t actually a half wave (high impedance) or quarter wave (low impedance) on any band. The Un-Un (see below right) transforms the impedance to something closer to 1:1 and therefore reduces coax losses.

Does it work? I have tried a 7.6m Rybakov at a few locations and yes it does. Is it as good as a dedicated resonant antenna – no. You get losses in the Un-Un and there will always be a residual SWR on the feedline, which adds to the losses. You will also need to use an ATU to get the SWR down to 1:1, although most internal ATUs can usually cope with the mismatch.

As a matter of interest, here is the SWR I found with the 7.6m vertical (with the 4:1 Un-Un) with a single earth stake and two 20ft radials:

7.50MHz - SWR 3.2:1
10.1MHz – SWR 3.6:1
14.2MHz – SWR 5.6:1
18.1MHz – SWR 5.4:1
21.2MHz – SWR 2.6:1
24.9MHz – SWR 2:1
28.5MHz – SWR 2.1:1

The radiation pattern on 10MHz and 14MHz is typical of a vertical and good for DX. It is not so good on the higher bands as the antenna is longer than a quarter wave and gets complex.

It is also rotten on 80m (SWR 330:10) as it is waaaaay too short – if you extend the wire to 8.6m it will be better on 40m but you may as well make it a proper quarter wave (about 10.4m).

In tests it could hear local CB stations that were inaudible on my Windom, W3EDP or horizontal dipole. I know most CBers use vertical polarisation, but it shows that it might be good for low-angle DX radiation on 10m too.

It is a bit short for 40m, but does work. It was ok on 20m - some stations louder than on a Windom at 30ft, some worse. Same on 17m.

For 20m–10m use I think you are better sticking to 7.6m. The modelling shows that the radials are critical. You may get different SWR readings depending on how many you have. A single earth stake might give you a low SWR, but will be lossy.

If you have little space and like experimenting it might be worth a go, but I never found it matched a proper half-wave dipole at 30ft. More radials might help –if laying on/under the ground they don’t have to be resonant, just make sure they are roughly the same length as the radiating element and aim for as many as possible (32+ is good). If you try this antenna with just a ground stake I think you will be disappointed.

Let me know how you get on.


  1. Rybakov allegedly means 'fisherman' or 'son of the fisherman', I was thinking of using one for my garden as a take down and put up antenna using a 9m roach is a true postage stamp size.....


  2. OK! I get the "fisherman" reference, but this now begs the question "Why 806?"

  3. Perhaps it is because it could work from 80m to 6m (albeit not very well!). Can't think of any other reason.


  4. I have always liked the idea of non-resonant aerials to give the matching unit an easy life. In this case I would mount it well off ground, say at fence height with one radial towards the dx, ie an L. For more consistent swr readings I always use a good choke followed by a stake in the ground on the tx side then the feeder on the ground is not picking up noise and spoiling the modest lobe to the dx. David G3UNA

  5. Hi,
    Whats the best (maximum) length of the wire for Rybakov?
    The longer the better?

    1. You want a length that is neither a half wave or a quarter wave at the frequency of operation, so neither a high impedance or low impedance. It depends on what frequency you wish to operate on. If you make it too long you will get a poor radiation pattern at HF, hence the suggestion of 7.6 or 9.6m. If you want it for 40 or 80m you may need to experiment, but more than 10m long for 40m and 20m for 80m.