Posts by PA3FYM

    No it is not difficult to combine a horn LNB with the circular wave guide:

    1. saw off the horn

    2. drill a 22mm hole in the horn dielectric and press the LNB on the waveguide.


    Ready (lots have done that already, me included ; -)

    Like I said .. forget about the (working of the) patch. That is a different story but in the explanation above I tried to simplify it as much as possible (perhaps too much) in order to explain the CP effect of the antenna and how the feedpoint was iterated.


    Google on articles where the precise working of patch antennas is explained, like in this Czech doc where a picture below summarizes what I wrote above:


    OK, now we get into the tricky point (the real secret of the POTY) ... the feed point .., which is the holy grail ; -)

    First .. I go 'back to basics' ... (forget about the 'patch', just let's simplify it down to earth) ... consider one of the 'antennas' (let's take the 'too long' one) ... as a 'dipole' .. or a 'radiator' grounded in the middle, like many 2m Yagi's and imagine a 'gamma match' (or 'shunt feed'). We know that Z = 0 in the grounded middle and that Z = (theoretically) infinite at the end of the radiator.

    Thus .. there is a 'tap' where R = 50 Ω (leave the jX for a while.. come back to it later).


    So, imagine an infinite small diameter 'waveguide' (call it a short) in the middle of the patch (because Z = 0, there can be an electrical short). The distance from this middle towards the vertical 'end' of the patch determines R (the real part of the impedance).

    This is (with the infinite small waveguide, say a M3-bolt ; -) along the vertical axis. Say that the 50Ω point will be at 28mm from the (grounded) center of the patch.


    However, we don't have an infinite small diameter waveguide (or M3 bolt), we have a 22mm (OD) copper tube in the center, which 'disturbes'.


    We go back to the infinite small wave guide in the center (or M3 bolt ; -) .. given the 28mm vertical 'tap' ... if we go (in the picture, see above) to the left we introduce a -jX (we run to the shorter side) .. if we go to the right .. we introduce a +jX. For a M3-bolt the jX = 0 is along the vertical axis.

    However, with the 22mm copper tube we do not have a M3 bolt, so there is 'some place' where Z = (almost) 50 + |j0| <-- absolute) Ω

    At the desired frequency (2400.175 MHz) jX is not 0 but a compromise .. such that the return loss (RL) is acceptable.


    Therefore the RL @2400 MHz is less than the two resonances (too short and too long antennas) but the phase differences are 90 degrees apart .. creating CP (with a good axial ratio)


    The 'horizontal' 'pick' (8mm in the POTY) is in fact a 'match' to eliminate the 'disturbance' of the 22mm OD copper tube (wave guide) and the 28mm 'pick' is the R = 50Ω point (in conjunction with the 22mm waveguide)

    Consider the cropped corners as a 'too short' antenna (so Fres > 2400 MHz) and the square corners as 'too long' (so Fres < 2400 MHz).

    Also imagine that the phase shift of every antenna is either -45 or +45 degrees , so the combination is 90 degrees --> delivers CP at the center frequency (Fcenter = 2400 MHz)


    (This is a very simple explanation, but hopefully enough to get a feeling how -and why- this antenna works/produces CP)


    In my experiment I used a 'ferrite pearl' (Ferritperle <-- German, or 'Ferrietkraal' <-- Dutch) of which I measured the AL and then wound it with windings so that XL of the windings was high enough. I only tested it with the Inverto 'Red Extend', not with other (non Inverto) LNB's.

    At our club we measured phase noise of (modified) LNB's last year summer. One of our members brought a R&S analyser (cost: around 200.000 Euro !) and we got nice pictures.


    Having a 'benchmark LNB' I determine p/n of other LNB's qualitatively by recording a tone with the same receiver/settings.

    I analyse the tone in Audacity and from that it can be seen/heard of the 'new' tone has more , or less, distortion on it.


    Bad phase noise reveals as 'raspiness' of the tone (in fact it's super imposed FM).

    BTW: In Spain this antenna would be illegal for QO-100 Uplink as it has < 22.5dB gain and less directivity... ;-)


    Rules which can't be enforced are useless rules. Also the social/economical impact of our 'Scooby Doo' experiments is below the noise / radar .. so it's like Orange Free State (Oranje Vrijstaat) (having a Dutch origin ; -)


    As long as you do not exceed the agreed uplink budget it's all fine imho.

    SV1BDS .. we're on the same page ... but .. it doesn't work in the same (one) Arduino. I managed to squeeze the LCD routines and NMEA -with QTH locator etc etc- AND Lars' GPSDO c0de into 30 KB .. but the GPSDO software screwed up with the softserial interrupts ..

    Hi George, no .. no news ; -) Busy here with other 'non RF/electronic' stuff but after that this OCXO and minimalistic transverter are next in the project queue.