After a bit of playing, I found that none of the matching techniques tried on the ADS-B ground plane antenna gave a usable device, and with the DVB-T stock mini mag-mount antenna being actually usable at 1GHz, as evidenced by having the system running on my laptop autonomously over yesterdays night shift, and it being perfectly capable of tracking at least those aircraft within a close enough range (about 10NM) despite the very much indoors location, the very high RF environment, the huge amount of local RFI, AND the antenna being stuck on a bit of a bolt head beside me on the desk, I have decided to abandon this design direction in favour of gain antennas.
The wire collinear proved to be a usable design, but is remarkably flimsy in its open state. Putting it in a PVC or fibreglass tube would solve this, but the ground radials would need replacing with thicker rods to survive any knocks.
So the idea is to make up a Coaxial Collinear (CoCo) antenna. I have investigated these before, but only at VHF, where unfortunately the dimensions become a bit too long when aiming for decent gain. At 27cm though, the dimensions for even a high gain version become quite manageable.
A CoCo antenna uses the collinear principle of feeding halfwave sections out of phase by 180 degrees, but whereas in a normal collinear this is achieved by quarterwave stubs, here it is done by reversing the connections between the half wave sections, which are formed from coaxial cable.
Designs online suggest a CoCo formed from eight 1/2 wavelength coaxial sections will have a gain of about 3.5dBd. But how long physically would that be?
Well, our design frequency is 1090MHz -
300/1090 = 0.275m
so a halfwave is 0.138m
Now, we cant calculate the length from this, as theres a couple of other things to take into account. One is the addition of a whip section on the top, which adds another half wavelength. The other is the cables velocity factor.
The cable im using is CTF100, which has a VF of 0.82 (Rf travels through it at 82% of free space velocity), so we must multiply our halfwave by this, except for the whip, which is 'free space'
0.138 x 0.82 = 0.113m = 113mm
we have 8 half wave sections, plus the whip,
(113 x 8) + 138 = 1041mm = 1.04m
Thats a pretty manageable length for a decent bit of gain. If we improve the match by including a quarterwave tube at the feedpoint, we'll only increase this by about another 7cm. But for receive only, this isnt vital.
I have the various coaxial sections cut ready, including an extra 20mm on one end and 15mm on the other. The reason for this is to allow enough of the inner conductor to connect between the sections (the difference in end lengths makes the next job easier!). To join these, the bared inner conductor is simply inserted between the outer sheath and the braid. This creates the alternating connection which provides the phase change.
I would normally prefer soldered connections, especially at high UHF! But the outer sheath will be tight against the connections and it seems provides a decent usable joint. I will cover each joint with heat shrink tubing anyway which will provide even more of a squeeze onto the connections.
I have a few feet of extra CTF100 to act as the feedline. This has an F connector on it at present, but I might replace that with a BNC to make interconnection easier. Once built and tested, the whole antenna will be slipped in and sealed into a length of PVC conduit, it will then be mountable somewhere up clear.
A worry with these DVB-T dongles is that the receive front end is absolutely wide open. To help prevent desense of the receiver from high powered broadcast stations, I will be adding to the antenna one or more quarterwave stubs. A quarterwave coaxial stub, open circuit at the end, will provide a notch filter at the design frequency. Not what I need here. But if it is short circuit at the end, it forms a band pass filter at the design frequency. Taking velocity factor into account, a coaxial stub filter for this frequency is a piddling 5.6cm long!
Apart from building one of them into a shielded box, this will be my last build as goes the ADS-B system. I need to get on with other projects, especially the 10m WSPR transceiver,
I have spent too long trying to get the Wispy design working for me, so have decided to abandon that also. Not in its entirety though. I will replace the mixer with an SBL-1, so I have then confidence in that part of the circuit. I am happy with the performance of the oscillator/doubler, so will keep this part of the design, except for adapting it to feed the SBL-1, which I expect will require a 5dB pad. Knowing that the parameters of the modulator are fixed and out of my control, means I can concentrate of setting the drive levels such that I have a working DSB modulator. Once this is done, I can build on it with the amplifier chain to get the desired output power. Whether I modify the Wispy board, or start a fresh layout I havent decided yet, but a fresh start will allow me to better shield the oscillator, and control the harmonic output.