Receive Antenna & Impedance "Matching"

Useful information regarding antennas for SDR products.
vk7jj
Posts: 206
Joined: Thu Mar 02, 2017 9:56 am
Location: Tasmania

Re: Receive Antenna & Impedance "Matching"

Post by vk7jj » Mon Jul 15, 2019 10:54 am

.. pity about the page break, I'd better post the comparison pic again
Here is the Kenwood
jj-high.png
jj-high.png (66.37 KiB) Viewed 3757 times
Here is the RSP1A
jj3-high.png
jj3-high.png (68.3 KiB) Viewed 3757 times
c) From the spreadsheet analysis of the many hundreds of spots from the last 24 hours it appears the RSP1A's WSPR noise floor is -29dB as compared with -33dB from the Kenwood so there is not much in it otherwise as you can see from the stronger SNRs above.

Thoughts:
I don't believe the difference is lack of RF gain.
Months of testing of the Kenwood on it's own (and an FT-991 and an FT-857) shows that turning their built in preamps on or off makes no difference at all to their ability to hear WSPR spots down to the -33dB WSPR noise floor which all 3 achieved.

RF bandwidth does matter.
I tried the RSP1A with and without the toroid preselector illustrated above and the preselector improved it's performance by around 4 to 6dB.
My impression is that the AGC action was moderated by the preselector, ie. strong adjacent stations on 40m caused more AGC action on the RSP than on the other radios. Of course that's only my impression and there may be other factors might make a difference, eg. I'm not sure what part the ADC resolution might play.

--

Because my Windows box is limited to a VM on a Mac I have not been able put two independent RSPs running SDRuno up against each other so that I could test out the twin coil preselectors, the differences in the dynamics of the RSP against an analogue radio make it too hard to properly analyse the anticipated results.

It's been fun.

Reason: No reason

glovisol
Posts: 662
Joined: Thu May 10, 2018 6:42 pm
Location: Piedmont, Italy

Re: Receive Antenna & Impedance "Matching"

Post by glovisol » Mon Jul 15, 2019 8:32 pm

Hi Phil,

Even if I have not been commenting, I have been following your WSPR activity, as described in this thread, an activity which I think is more than interesting, it is fascinating. I am afraid to get involved, as I am afraid it will capture my attention completely,

However, if my input can be worth 2 cents, please understand that, among the RSP's the RSP-1A is not the best choice for the activity you are doing. The RSP2 or (better) the RSPduo with their HI Z imputs exhibit a tested noise figure that is from 6 to 9 dB better in HF than that of the RSP-1A with its coaxial 50 Ohm input. Phil, here, with WSPR activity, working with negative S/N's in the order of 20 dB's the advantage in noise figure I mentioned will go a long way to demonstrate the sensitivity advantage of the RSP devices. So my advice is to get a duo and see for yourself. From what I know you should get an advantage that will make the pre-selector advantage negligible.

Going now to pre-selector action resulting in WSPR sensitivity advantage, we are confronted, in my opinion, by an as yet unexplained effect, becaus I do not think you have such strong signals near your WSPR reception segment, so there should be no positive effect of a narrow preselector on AGC action, unless there is an as yet unknown noise modulation effect increasing the noise threshold of the receiver caused by nearby stronger signals. In any case you could test this by resorting to AGC manual settings to try and understand what is happening.

Finally, if you really determine without doubt that extreme front end selectivity is the name of the game, I would gladly offer to develop and deliver ad hoc crystal filters in a joint effort to advance the WSPR technology.

Kind regards,

Gianfranco

Reason: No reason

vk7jj
Posts: 206
Joined: Thu Mar 02, 2017 9:56 am
Location: Tasmania

Re: Receive Antenna & Impedance "Matching"

Post by vk7jj » Tue Jul 16, 2019 12:27 am

Hi Gianfranco

I do have an RSP2 but I can't run it and the RSP1A simultaneously with SDRuno because of my lack of Windows boxes and I did put an RSP2 through it's paces using it's high impedance input a couple of pages back, threads progress too quickly!
After many days of testing, comparing the FT-857 with the RSP2 is too difficult. They behave quite differently on WSPR mostly it would appear because of their different AGC action and the results are not readily comparable. If I placed the preselector inline I would not be able to tell what differences it was creating.

The FT-857 has been abandoned and am trying now to set up the RSP2 and an RSP1A for independent simultaneous operation in the hope they track closely enough to make the preselector test valid.
I can't afford a Duo! And if I could I'd have to go and buy a new Windows machine to make any use of it which doubles the impecuniosity!

--

On the subject of Windows v other OS:

I went to a lot of effort seriously trying out the RSP1A and the RSP2 using CubicSDR and Gqrx on Linux and on Mac.
They both proved totally inadequate. Great for general listening but with major limitations in design and with coding bugs.

SDRuno is so far ahead of them it's not funny.

It is really a shame Windows is the only platform that supports a viable range of RSP software, it's worse than a shame actually as I personally find the cost of keeping up to date with Windows as well as maintaining it so much of a pain I'd love to see it dead and buried. There are so many older laptops and other hardware out there that run Linux sublimely, or Macs, just begging to be used by radio hobbyists.

Phil

Reason: No reason

vk7jj
Posts: 206
Joined: Thu Mar 02, 2017 9:56 am
Location: Tasmania

Re: Receive Antenna & Impedance "Matching"

Post by vk7jj » Wed Jul 17, 2019 8:27 am

Hi Gianfranco, sorry about the delay in replying to your other points, been choring.

Thanks for you very kind offer regarding the crystals, while only a single crystal appears to give the required bandwidth, like a lot of other radio commodities we took for granted they have become very expensive.

WSPR has been fun but I only got into it as a tool to improve my antennas and shack noise and at this stage unless some magnificent new idea occurs the return on investment in further effort has reached the marginal stage so it's only out of some form of mild addiction that I keep doing it.

Regarding the effects of the various filters / AGC action on the RSPs, yes it is probably even more of a mystery than before. That's partly because I've so far only been able to compare RSPs with normal analogue receivers. That comparison shows they behave differently but those differences raise other more complex questions.

Putting two identical RSPs head to head and then introducing filtering and other changes to one of them is the only way forward, so yes, a Duo would be the ideal test platform wouldn't it and I know how enamoured you are with yours.

Hope your ferrites turn up soon, some more are on the way to me too. The batch I bought a month or so ago are hidden somewhere in my shed unless they were inadvertently thrown out and I've driven myself mad trying to find them. Interestingly when I used a coax T connector to share my second antenna between an RSP1A and an FT-857 the 857 appeared as if it was hogging the signal so another splitter is essential.

Also the results of the shoot out between the VK7JJ3 RSP1A and the VK7JJ souped up Kenwood 120S are on WSPR Challenge site at the moment. The RSP1A did a fantastic job didn't it, only a couple of dB between them. That tiny little black box called the "budget RSP" has beaten out all the rest of the world :-)
challenge.png
challenge.png (126.68 KiB) Viewed 3659 times
Regards,

Phil

Reason: No reason

ON5HB
Posts: 140
Joined: Sat Dec 29, 2018 1:07 pm

Re: Receive Antenna & Impedance "Matching"

Post by ON5HB » Sat Jul 20, 2019 5:15 pm

glovisol wrote: However, if my input can be worth 2 cents, please understand that, among the RSP's the RSP-1A is not the best choice for the activity you are doing. The RSP2 or (better) the RSPduo with their HI Z imputs exhibit a tested noise figure that is from 6 to 9 dB better in HF than that of the RSP-1A with its coaxial 50 Ohm input. Phil, here, with WSPR activity, working with negative S/N's in the order of 20 dB's the advantage in noise figure I mentioned will go a long way to demonstrate the sensitivity advantage of the RSP devices. So my advice is to get a duo and see for yourself. From what I know you should get an advantage that will make the pre-selector advantage negligible.

Going now to pre-selector action resulting in WSPR sensitivity advantage, we are confronted, in my opinion, by an as yet unexplained effect, becaus I do not think you have such strong signals near your WSPR reception segment, so there should be no positive effect of a narrow preselector on AGC action, unless there is an as yet unknown noise modulation effect increasing the noise threshold of the receiver caused by nearby stronger signals. In any case you could test this by resorting to AGC manual settings to try and understand what is happening.
Sorry but this is not correct.
WSPR will go down as low as -33db, the only thing matters is the quality of it's own noise-level, as the RSP1A does something like -140db there is absolutely no problem.
Changing to a different box or the Hi-Z will not matter at all, if you receive anything or not is totally up to the antenna as receivers coax or ladder-line are more then sensitive enough.
These signals are so far below the noise-level that it's almost impossible to measure, it's all mathematics and very slow tones.
A Duo will not matter, you can get the same results with a Dongle, just matters how clean the antenna-signal reaches your receiver.

To give you an idea, 1mW (with better conditions then now) will bring a signal 9000Km away on 40m, every night, clockwork.
The only thing that matters is a clean reception from the antenna fed into an average receiver, doesn't need to be as sensitive as the RSP1A or others.

Reason: No reason

SDRguy22
Posts: 12
Joined: Fri Jul 19, 2019 4:34 pm

Re: Receive Antenna & Impedance "Matching"

Post by SDRguy22 » Sat Jul 20, 2019 8:00 pm

ON5HB wrote: WSPR will go down as low as -33db, the only thing matters is the quality of it's own noise-level, as the RSP1A does something like -140db there is absolutely no problem.
Bas your comments on WSPR agree with my experience using this mode. I have tried it on 630M, 80M, 40M and 20M with good results using several different antennas.

At my location WSPR typically decodes down to -28 dB SNR but -30 dB is sometimes observed under excellent conditions. I find it an excellent way to observe band conditions. Anyone interested in learning more about WSPR can find some good information here.

https://en.wikipedia.org/wiki/WSPR_(ama ... _software)

WSPR is easy to install and get running with an RSP.

Bill
Last edited by SDRguy22 on Fri Jul 26, 2019 2:39 pm, edited 2 times in total.

ON5HB
Posts: 140
Joined: Sat Dec 29, 2018 1:07 pm

Re: Receive Antenna & Impedance "Matching"

Post by ON5HB » Sat Jul 20, 2019 8:39 pm

Well you say the same as I but in more detail.

But I do not agree with 1 point, and that is the noise-floor.
WSPR will go under the noise (whatever DBM it is) by -33.

Mostly the noise on the band is pretty high, so most receivers have no problem capturing it.

In short....WSPR goes below noise by 33db at best....and the RSP's have no problem handling such.

Take CW and then twice as much noise and you are still not at the same level :lol:

Reason: No reason

glovisol
Posts: 662
Joined: Thu May 10, 2018 6:42 pm
Location: Piedmont, Italy

Re: Receive Antenna & Impedance "Matching"

Post by glovisol » Wed Jul 24, 2019 11:15 am

CLARIFICATION ON SENSITIVITY PERFORMANCE COMPARISON RSP-1A VS. RSPduo & FOCUS ON NEGATIVE S/N

I believe the previous two posts have introduced confusion on several important questions and with this post I wish to clarify some ESSENTIAL concepts.

1. RSP-1A and RSPduo Sensitivity & Noise Figure specifications

The beauty of handling equipment by SDRplay is that you do not have to speculate on performance parameters, because they are readily supplied by the Manufacturer. Here we are concerned with the sensitivity & NF specifications of the above mentioned RSp's and here is the data readily available from the RDSplay site.
RSP 1A-Duo compare.png
RSP 1A-Duo compare.png (7.96 KiB) Viewed 3420 times
From this table it is evident that at LF the RSPduo HI Z input has a 10 dB Noise Figure advantage over the RSP-1A.

2. Significance of Noise Figure

If a device, any device, receives a signal with a S/N ratio of 10 dB and has a Noise Figure of 8 dB, then the device will provide at its output the same signal with a S/N ratio of: 10-8= 2dB. If the same device receives a signal with a noise ratio(S/N) of -10 dB, it will provide at its output the same signal with a noise ratio (S/N) of: -10-8= -18 db.

In other words the device DEGRADES the received S/N by an amount in dB EQUAL TO ITS RATED NOISE FIGURE.

3. Significance of the advantage of the RSPduo over the RSP-1A in WSPR reception

If I am receiving a WSPR signal with the RSP-1A and the S/N coming from the antenna is -20 dB, the total S/N ratio will become: S/N= -20-21.2 = -41.2 dB.

If I am receiving a WSPR signal with the RSPduo and the S/N coming from the antenna is -20 dB, the total S/N ratio will become: S/N= -20-11.75 = -31.75 dB.

Thus the RSPduo has an extremely significant advantage (10 dB!) over the RSP-1A.

4. Significance of receiving bandwidth

The detection bandwith is neutral in the comparison. If we go from the 1800 Hz referred to in the SDRplay specifications, to, for example 800 Hz, the noise figure in both receivers will improved by:
NF1800 = 10*LOG(10) (1800) = 32.55 dB
NF800 = 10*LOG(10) (800) = 29 dB
Improvement = 32.55 - 29 = 3.55 dB

So, at the end of the day, the RSP-1A NF will become: 21.2 - 3.55 = 17,55 dB and the RSPduo NF will become: 11.75 - 3.55 = 8.2 dB and the corrected S/N's will become:

RSP-1A: -20 - 17.55 = -37.55 dB
RSPduo: -20 - 8.2 = -28.2 dB

AND THE SDRduo will still give an ADVANTAGE ON WSPR RECEPTION OF 10 dB as before!!

5. Conclusion

It is not true that with negative S/N the receiver sensitivity has no relevance: in fact the contrary is true and surely the RSP2 and the RSPduo have the potential of providing an extremely significant advantage of a 10 dB Noise Margin on WSPR reception.

6. Unexplained anomaly

If, as reported by vk7jj, it is possible to detect and confirm an improvement in WSPR reception results, not because of a restriction of the DETECTION BANDWIDTH, but because of a selectivity improvement at the receiver's input, then we are in presence of an unexplained anomaly that needs further investigation.

For further reading on these ESSENTIAL concepts, please read the thread shown below, which includes an extremely clear, masterful and praiseworthy tractation on Noise Figure measurement and how SDRplay interacts with S/N, NF and Noise Floor by Roger.

https://www.sdrplay.com/community/viewt ... =assessing


glovisol
Last edited by glovisol on Thu Jan 01, 1970 12:00 am, edited 0 times in total.
Reason: No reason

glovisol
Posts: 662
Joined: Thu May 10, 2018 6:42 pm
Location: Piedmont, Italy

Re: Receive Antenna & Impedance "Matching"

Post by glovisol » Wed Jul 24, 2019 4:39 pm

CLARIFICATION ON SENSITIVITY PERFORMANCE COMPARISON RSP-1A VS. RSPduo & FOCUS ON NEGATIVE S/N

7. SDR Noise Figure performance @ 80 KHz

A legitimate objection that could be opposed to the data published and the quantities computed in the previous post is
that the SDR Noise Figures taken from the published SDRplay specifications are referred to 1000 KHz, while some users are concerned with reception at 77.77 KHz and therefore previously examined results could change at a lower frequency. For this reason Noise figures for the RSP-1A and for the RSPduo have been investigated as per the reference shown in the previous post.

Frequency: 80 Khz
Detection bandwith BW: 1800 Hz
Factor 10*LOG10(BW) = 10*LOG10(1800) = 32.44
Measured terminated Noise Floor
RSP-1A (50 Ohm input): -107 dBm
RSPduo (HI Z input): -139 dBm
Calculated Noise Figure
RSP-1A - NF = 174 - 107 - 32.44 = 36.44 dB
RSPduo - NF = 174 - 139 -32.44 = 2.44 dB

The above results show that in reality, at LF, the Noise Figure advantage of the RSPduo over the RSP-1A is an enormous 34 dB. This happens because while the Noise Floor of the RSPduo changes little, the Noise Floor of the RSP-1A deteriorates drammatically. Thus for very weak signal reception of incoming LF signals with negative S/N's, by means of the Autocorrelation Function, the superiority of the RSPduo makes the difference between success and failure. Furthermore this value of NF for the RSPduo at LF is also exceptional when compared to other commercial LF/HF receivers.

These findings agree with what written on the sensitivity of the RSP-1A at LF by Roger in the thread referenced below.

https://www.sdrplay.com/community/viewt ... A+longwave
The RSP1A does not perform as well on LF as the RSP2 and the RSPduo when using a poor antenna. This is because the RSP1A has a higher noise floor
Last edited by glovisol on Thu Jan 01, 1970 12:00 am, edited 0 times in total.
Reason: No reason

glovisol
Posts: 662
Joined: Thu May 10, 2018 6:42 pm
Location: Piedmont, Italy

Re: Receive Antenna & Impedance "Matching"

Post by glovisol » Wed Jul 24, 2019 5:44 pm

CLARIFICATION ON SENSITIVITY PERFORMANCE COMPARISON RSP-1A VS. RSPduo & FOCUS ON NEGATIVE S/N

8. SDR Noise Figure performance @ 7-8 MHz

Finally we can look at the 40 m Band, because on this frequency vk7jj published most of his results for WSPR reception. The SDRplay published N.F. data here is as follows:

RSP-1A (50 Ohm input/1800 Hz BW) - NF = 18 dB @ 8 MHz
RSPduo (HI Z input)/1800 Hz BW - NF = 16.65 dB @ 7 MHz

Here the NF advantage amounts to 1.35 dB, small, but still significant.

According to my experience, Noise Figure does not tell the whole story at HF frequencies, because the balanced HI Z input allows very simple balanced input antenna configurations that naturally reject local noise, while the more complex circuitry and provisions for antenna connections in coaxial mode tend either to increase insertion loss, or to increase local noise input level, or both.
Last edited by glovisol on Thu Jan 01, 1970 12:00 am, edited 0 times in total.
Reason: No reason

Post Reply