500 metre radio communications

Here's a nifty antenna that will add performance to your data link. It's not much harder to build than a dipole, but provides a theoretical 4.8 dB gain over a 1/2 wave dipole. That's a bunch of gain for very little work, particularly if you use one of these at each end.

You can orient these either horizontally or vertically, as long as you use the same orientation for both the transmit and receive antenna. (If you have no overriding criteria, orient them horizontally and drop the coax feedline straight down for at least a couple of wavelengths.)

Good luck!

Tom
 

Attachments

  • Collinear Dipole.gif
    Collinear Dipole.gif
    4.8 KB · Views: 270
Thanks Tom2000. Something like this http://martybugs.net/wireless/collinear.cgi perhaps? Or http://www.rason.org/Projects/collant/collant.htm

Is it mounted vertically, ie with the antenna 90 degrees to your drawing? I gather a collinear gets its gain (when mounted vertically) by sending less energy up into the atmosphere or down into the ground. Yagis are great, but for a wireless mesh, it is better to have an omni, and a collinear might be just the right answer - some gain but in an omni direction. I'll see if I can put one together.

The links didn't use a dipole. Is the other half of the dipole needed?
 
Last edited:
Dr_Acula said:
Thanks Tom2000. Something like this http://martybugs.net/wireless/collinear.cgi perhaps? Or http://www.rason.org/Projects/collant/collant.htm

Is it mounted vertically, ie with the antenna 90 degrees to your drawing? I gather a collinear gets its gain (when mounted vertically) by sending less energy up into the atmosphere or down into the ground. Yagis are great, but for a wireless mesh, it is better to have an omni, and a collinear might be just the right answer - some gain but in an omni direction. I'll see if I can put one together.

The links didn't use a dipole. Is the other half of the dipole needed?
Click to expand...

Thanks for the links! I really like the coax collinear design. I've never seen one of those. It appears to be relatively simple to construct, and provides a whopping amount of gain.

Any "gain" antenna gets its gain by concentrating the radiation in one or more preferred directions. For a vertically-mounted collinear mounted in ground proximity, the radiation pattern will be sort of a truncated toroid. Think the top half of a sliced bagel. The more antenna gain, the flatter the bagel. For a simple dipole or 1/4 wave vertical, a very fat bagel.

Yes, you can certainly cut "my" antenna in half and mount it vertically. At the antenna feedpoint, connect the coax braid to four equally-spaced 1/4 wave radials, or perhaps a square yard or two of hardware cloth, chicken wire, or whatever you might scrounge. However, once you're fooling around with radials or a ground plane, mounting arrangements, perhaps a decoupling sleeve, etc, you're probably better off building that 9 dB coax collinear. Mechanically, it's a simpler task.

At least, with some sort of ground plane arrangement, you won't need a balun.

By the way... I see that the 2.4 GHz collinear uses coiled matching stubs. On one of the versions of "my" antenna I built for 144 MHz, I tried coiled coax matching stubs. Didn't work at all. I think they acted like chokes, not like stubs.

I'd recommend that you build the stubs out of stiff copper wire, configured as shown on my sketch -- out, squared off, and back. At 315 or 433 MHz, maybe 15 or 20 mm spacing.

By the way... I built these as dipoles to neatly solve the problem of preventing the coax shield from radiating. When you eliminate "the other half" of the antenna, shield radiation becomes problematic, and can destroy the efficiency of the antenna. Note the measures that the designer of the coax collinear has taken: a 1/2 wave decoupling sleeve and the toroid cores which act as an RF choke for unbalanced RF energy on the shield.

The "half dipole" configuration over either radials or a ground plane is a compromise, but one that should be reasonable for this application. The better solution would be radials or a ground plane and a decoupling sleeve. But that's approaching the ridiculous. Mechanically, it's much simpler to construct the antenna as a dipole, avoiding the shield radiation problem completely.

Have fun!

Tom
 
Last edited:
I once read a delightful book (whose name eludes me for the moment) where the main character was in charge of designing radios somewhere in Africa, and after he had presented the factory with the design, they would proceed to systematically remove one component at a time, and if the radio still worked they would leave that component out in the manufactured device.

Now I have a long distance link working, I'm doing the same thing! First thing is to shorten the preamble, and 6 U's seems to still work fine. Second is to delete the Manchester coding, and to present the Tx unit with a wakeup preamble, then 13 0s and then a single checksum which has to match. This seems to work fine as well with no difference to the range. Minimalist code below:

'Transmitter

main:w0=0
w1=0
w2=0
w3=0
w4=0
w5=0
b13=123
serout 1,N2400,("UUUUUUXYZ",b0,b1,b2,b3,b4,b5,b6,b7,b8,b9,b10,b11,b12,b13)
pause 60000
goto main

' Receiver

start:high 2
pause 1000
low 2
low 0' tx
main: serin 4,N2400,("XYZ"),b0,b1,b2,b3,b4,b5,b6,b7,b8,b9,b10,b11,b12,b13
if b13=123 then
for b1=1 to 55
high 2
pause 50
low 2
pause 950
next
endif
goto main
 
You must log in or register to reply here.
Back
Top