Page 3 of 12

Re: Receive Antenna & Impedance "Matching"

Posted: Sun Jun 09, 2019 10:20 am
by vk7jj
That's looking encouraging isn't it, you are such an enthusiastic worker you put me to shame.

Regarding the tap at 2 turns, that was chosen as a guess as my antenna has an impedance of around 300 ohms on 40m and I'm feeding the high impedance input of an RSP2, perhaps a better approach would be to tap at 1 turn so as to prevent loading the tank circuit as heavily. Hopefully improve Q and insertion loss. Most people would be looking at 50 ohms in and out wouldn't they.

Alternatively instead of a tap what do you think of trying a simple coupling loop? eg years a go I came up with a tank circuit to end feed a resonant halfwave 14MHz wire on a squid pole and found it was the best compromise. That would make it easier to build and experiment with lighter loading.

Excuse the old unchanged museum of a website from 2008

Re. your excellent articles on your splitter design and construction, yes I read your work with interest but I'd already built two quite a while ago as per that article. I mentioned the article because of your interest as it was so similar in principle. I like your implementation with the binocular core. I've got some new binoculars and toroids on their way from Mouser so I can have more of a play.

Re: Receive Antenna & Impedance "Matching"

Posted: Sun Jun 09, 2019 3:18 pm
by glovisol

you MUST STOP giving me ideas!! Now this squid pole antenna is a fantastic idea: unfortunately it is Sunday now and I shall have to wait until tomorrow to go and look for a fishing pole.

Coming back to the pre-selector, in my opinion it is really much better to use links to couple in and out and this is what I have done with my prototype. Advantages of using links are as follows:

1) You have two ways to change coupling: (a) number of turns & (b) placing it nearer or farther away. This is very convenient when building the coils on a PVC pipe, like in the example under consideration.
2) You are really handling a three winding transformer, with extremely low capacitance between input and output. You save a ferrite isolation transformer (and its losses) and you get rid of all local man made noise coupled through parasitic capacitive effects.
3) To change the coupling soldering and desoldering on a tap is a sort of nightmare, so you do not do it and never optimise the thing. Changing a link is a very esy life.



Re: Receive Antenna & Impedance "Matching"

Posted: Sun Jun 09, 2019 4:10 pm
by glovisol

In the test prototype I ended up with a coil Qu of 300 and I have measured a 3 dB bandwith of 40 KHz, so my Ql is:
Ql = 7,100/40 = 177.5.

My ratio is: Qu/Ql = 300/177.5 = 1.69 and my loss:
IL = 20*LOG(1-(1/1.69)) = -8 dB

I found -10/-11 dB, which can be accounted for the ceramic capacitor Qu, coil measurement error and measurement setup wideband transformers loss: so my results agree with the machine mathematics. What I did, was that I coupled less, trading more loss for higher selectivity, so the essence of this game is to work on coils, in order to get the best possible Qu.

Uploaded below screens of performance at 200 / 500 & 1000 KHz measurement bandwith. An added bounus of the pre-selector is the ULTIMATE ATTENUATION, which is extremely good far away. This means that strong signals even one octave below and above center frequency are wiped out and the poor wideband RSP loafs along with no worry of unwanted strong signals anywhere.

Re: Receive Antenna & Impedance "Matching"

Posted: Mon Jun 10, 2019 3:08 am
by vk7jj
Nice to see your refinements brought benefits; I whizzed up a single turn coupling version just for fun and the loop is pretty close to 50 ohms, haven't had a chance to do much more though.

The single turn is the 1mm wire at the bottom. As you can see I'm using ceramic trimmers. I've got a few beehives in the junk box I could be using but I went for parts that people could easily buy and the locals have nothing air spaced.

Re: Receive Antenna & Impedance "Matching"

Posted: Mon Jun 10, 2019 3:22 am
by vk7jj
OK on the squid pole, they are great for fixed mobile or portable because they collapse to a small ultra light size and if you have a socket ready to mount one you can be on the air in less than two minutes from a standing start.

Here is one in all it's glory. Forgive the lattice tower joke it's the only pic I could quickly put my hands on.

One drawback for use at home is that homes tend to be lots noisier locations, so the low angle vertical polarisation hears it all.

I reckon your beautiful vineyard would look even more splendid with a nice horizontal delta :-)

Re: Receive Antenna & Impedance "Matching"

Posted: Mon Jun 10, 2019 11:33 am
by vk7jj
Cheap wire delta loop antennas work, along with roughed up preselectors and an RSP2, how is this for an Rx effort...

I modded a test version of my WSPR website to track ham balloons. The power out of the balloon Tx is 10mW on the 30m ham band. As you can see every point on the map was a transmission received by me in Tasmania on an RSP2.

The best SNRs have been -21dBm at 1,584,700 Kilometres per Watt. The worst SNR was -32dBm. The distances varied from 8,000Km to 18,000Km.


Re: Receive Antenna & Impedance "Matching"

Posted: Mon Jun 10, 2019 6:22 pm
by glovisol
Very good performance of the delta loop and of the RSP2. I am getting organized to set one up, but it will be in september next, as I need to put the poles up and help will be available then.

As a young designer in the early sixties (modern filter theory was being perfected just then) my chief would always tell me: "if you do not have enough Qu, just use more stuff". This lesson is valid to this day. Even with a double section tuneable preselector you end up with low unwanted signal rejection and high insertion loss: this is shown by my prototype. Even supposing you succeed in building two reasonable size coils with Qu=400, you end up with an insertion loss of 13 dB approx., but with a modest attenuation in the broadcast band above 7,200 KHz. The other problem with a tuneable pre-selector is....tuning: it is difficult to see what you are doing on the SDRuno screen.

Uploaded below is the calculated response of a Chebyshev N=4, four tuned circuit, fixed pre-selector. This filter covers the entire 40 m Ham band, but wipes the broacdcasters out completely, not to speak about signals anywhere else in the LF/HF frequency range. But the greatest advantage is the ability to get these results with a significantly lower coil Qu. In fact with a Qu in the order of 250 (easily achievable) we should have an insertion loss of 6 dB or less. Then, if this design works, you stick the preselector at the antenna and you forget about it, unless you change band. Input and output coupling also done with links here.

Of course between saying and doing, the sailors say there is the ocean in between, but I think this is worth trying.

Re: Receive Antenna & Impedance "Matching"

Posted: Tue Jun 11, 2019 5:37 am
by vk7jj
Hello again, been winter pruning fruit trees all day.

Just wound a second tank the same as the first, the plan being to use the 50 ohm loops as "in one end and of the twin section out the other" but have only started experimenting with inter-tank coupling.

Part of my problem is the only Windows box I have is an older iMac running VMWare Fusion and that's busy with the RSP2 on WSPR so it's hard to experiment. My other two WSPR bands are running on Linux on old laptops. I'd love to ditch Windows but of course it's not effectively possible.

Your band pass Chebyshev shows promise and would make a lot of sense for ham users as it avoids the messy tuning issue. Not good enough for weak signal WSPR on a busy band where even two series antenna tuners make a really significant difference, especially with the RSPs.

Re: Receive Antenna & Impedance "Matching"

Posted: Tue Jun 11, 2019 6:52 am
by glovisol
What is the size of the coil you have shown? The diameter look greater than 25 mm.... If you give coil physical data I can build one and tell you what the measured Qu is.

Re: Receive Antenna & Impedance "Matching"

Posted: Tue Jun 11, 2019 7:22 am
by vk7jj
That would be splendid, thank you! It's annoying not being able to readily use the Windows box, hence the photo of the antenna analyser instead of antscope.

Exactly as per my original post, the 40m version as below. The PVC is a standard size that I had lying round here and the coil length is accurate to a mm.

PVC outside diameter = 2.7cm, wire = 0.4mm enamel copper, the single turn is 1mm copper.

20m => 17 turns, coil length approx 0.71cm
30m => 21.3 turns, coil length approx 0.87cm
40m => 27 turns, coil length approx 1.11cm