Page 10 of 12

Re: Receive Antenna & Impedance "Matching"

Posted: Sun Jun 30, 2019 6:57 am
by vk7jj
That's a pretty good question Gianfranco!

A core i5 refurb with a decent screen where we live is closer to $400 and then you start thinking it's 5400rpm snaildrive needs replacing with an SSD, and a new Pi4 is $60 but that's not the real story.

I've got two remote Pi running irrigation pump, watering etc which means driving relays via the Pi I/O pins. In the shack another one drives the FT-857D and an old FRG-9600 doing TTL CAT directly from the pins, the FRG-9600 works really well as Pi driven scanner. The FT-991 is connected via USB and my homebrew software controls every function and all memories and all prefs/param settings.

No Pi has a screen, mouse or keyboard, they are all headless running SSH for OS level work and have onboard node.js web servers for their user interfaces, all web-socket driven, updating multiple users from anywhere. One Pi in the shack streams full time audio from scanner and any other rig onto the LAN using multicast so I can listen anywhere. In the shack they are powered by linear 5V regulators connected to the 12V 1KW solar battery bank, so tiny footprint and useful. Coding/programming is also a big part of my hobby and a Pi makes a great development platform if you happen to be using the same programming language out in the wild, as I do.

I take my Pi3B to our rock-n-roll club when it's my turn to do the music, it runs (free) Volumio and has the entire club music collection on board, use it at home otherwise. Volumio is great for anyone wanting to carry a complete portable music system, it creates it's own local WiFi network (or joins one) allowing phones or tablets or computers to control it via web browser, has an optional app, supports web-sockets for multi-user. Handily it alternatively accepts audio streams via WiFi straight from any music on an iPhone or Android device or computer. Accepts USB music too. /end plug


Watch out for your data on those Iomegas! Very popular when they first came out but my two died prematurely from the dreaded click-of-death with no warning. I've still got their little blue power packs with the obligatory bent pins.


The continuing WSPR shootout between the FT-857 and FT-991 is interesting, the 857 is around 1dB better SNR on almost every spot after 12 hours with largest SNR difference for any single spot being only 2dB.

That tiny 1dB average advantage meant the 857 had 47 more spots which demonstrates just how important an extra 1dB SNR is, for WSPR anyway.

Over the last 12 months I've been using the same technique to slowly improve antennas and feedlines and that has been really meaningful, don't overlook the usefulness of WSPR if you are inclined to do some testing. Anyway I'll leave those two rigs running for a full 24 hours UTC then start switching radios.

My thought was to put the RSP2 up against the FT-857 for a day or two to get a known SNR relationship between them and then put the preselector in series with the RSP2 and fiddle.

After that I might try a shootout between the RSP2 and RSP1A...

If anyone wants to follow the current radio shootout, open in two web browser windows simultaneously and configure them as illustrated below.

The FT-991 is VK7JJ and the FT-857D is VK7JJ2

The auto-search needs to be on (on both) to ensure they both get the same bunch of spots as new spots are uploaded every even two minutes. Make sure the VK7JJ band is set to 30m because 40m antenna testing is ongoing.


Re: Receive Antenna & Impedance "Matching"

Posted: Mon Jul 01, 2019 10:17 am
by vk7jj
The winner of the WSPR 30m first round shootout is the FT-857D

Both radios achieved 88 total unique callsigns heard over a 24 hour UTC period.

The win was thus decided on the basis of extra spots from the same stations being received by the 857 that the 991 missed rather than new stations being heard. Those extra spots were all close to my SNR floor which is around -32dB.

The numbers:

The FT-991 achieved a total of 699 spots in the last 12 hours.
The cumulative distance for those spots was 8,196,755 kilometres.

The FT-857 achieved a total of 762 spots in the last 12 hours.
The cumulative distance for those spots was 9,190,539 kilometres.


Sometime over the next day when I get some time I'lll swap out the FT-991 for an SDRplay RSP2 using it's high impedance input. It's callsign will be VK7JJ3.

I'll let that run for at least a couple of days before putting a preselector in line with the RSP2.

Assessing the performance of a preselector in front of an RSP2 and an RSP1A while running SDRuno software is the goal of these tests.

For those who might wonder how it is possible to listen to transmissions more than 30dB below your normal noise floor check out WSJT-X software's digital modes. WSJT-X is freely available available for Windows, Mac and Linux including the Raspberry Pi and is really great for playing around with SDRplay and SDRuno in particular .

It is hopefully common knowledge that it is easy to listen to two different bands simultaneously in SDRuno and configure the audio output of one band to use the Right and the other the Left audio channel of the stereo pair.

It's then easy to run two instances of WSJT-X and configure their preferences so one listens to the L and the other the R.


Re: Receive Antenna & Impedance "Matching"

Posted: Mon Jul 01, 2019 10:36 am
by glovisol
Looks very interesting...I shall look at it after the holidays! Thanks Phil...


Re: Receive Antenna & Impedance "Matching"

Posted: Tue Jul 02, 2019 11:54 pm
by vk7jj
Hi Gianfranco, enjoy that magnificent holiday!


For anyone following the WSPR tests a software problem took down VK7JJ3 last night and there are other issues as well.

More later, Phil

Re: Receive Antenna & Impedance "Matching"

Posted: Fri Jul 05, 2019 3:05 am
by vk7jj
Finally got around to resolving the issues I've been having...

The inexpensive chassis mount UHF sockets on my home brew antenna splitter are faulty, the inner female part has been rotating slightly each time I inserted and screwed in the coax plug, finally causing the wires from the toroid to break off.

The fault was invisible being inside a diecast box and the proximity of the broken ends still allowed enough RF coupling to make it appear it was functional, yet the performance of the radios was not up to expectations.

All fixed, so VK7JJ and VK7JJ3 are now live.


Re: Receive Antenna & Impedance "Matching"

Posted: Wed Jul 10, 2019 6:06 am
by vk7jj

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.

That can't be done using SDRuno as there is only one computer running Windows and that's in a VM on a Mac so I've been trying to set up two linux based CubicSDR devices even though it falls over with memory leaks from time to time.

My Pi4B which has just arrived was to be one of them but I can't get Cubic running on it with an RSP; Rasbian Buster is a bloody hopeless OS and that's all that Pi will run at this stage. Therefore it's going to have to be done with two Macs but can't spare one of them at the moment.

Frustrated, Phil

Re: Receive Antenna & Impedance "Matching"

Posted: Thu Jul 11, 2019 1:47 am
by vk7jj
Update: Been running an RSP1A and an RSP2 side by side using CubicSDR on two separate iMacs, with audio piped directly from Cubic into WSJT-X in software using Soundflower.

Have tried CubicSDR versions 0.2.4 and 0.2.5 and neither seem to permit me to choose smaller bandwidths like 250k along with anything other than zero IF mode, on my computers Cubic fails to respond and does not receive so I was forced to stick to 2.048MHz in 1620 IF mode.

In that configuration either I've been doing something wrong or both setups were uncompetitive as they have received nothing on WSPR in the first two hours alongside an FT-857 on the same splitter which received continuous spots so I've discontinued the experiment until I think of something else.

Both RSPs work just fine in normal use with SDRuno.


Re: Receive Antenna & Impedance "Matching"

Posted: Fri Jul 12, 2019 5:05 am
by vk7jj

In order to confirm that RSPs do indeed work well on WSPR I've set up an RSP1A with my original simple fairly broad toroid preselector on 40m with the callsign VK7JJ3 using SDRuno.

It's up against the winning Kenwood TS-120S with a high Q antenna tuner with the callsign VK7JJ. As mentioned the Kenwood is set up with a programmable GPS-DO as external VFO.

They are both fed from one 160m loop via a balanced splitter with around 3dB insertion loss. They have been running for 30 minutes as I type this and it would appear the RSP has the edge.

VK7JJ callsign is also being used on 30m by another radio on a second 160m loop; anyone searching for the results of the test above needs to select 40m for VK7JJ or they will get both 30m and 40m results.


Re: Receive Antenna & Impedance "Matching"

Posted: Sat Jul 13, 2019 10:58 am
by vk7jj

12 hours of RSP1A versus TS-120S revealed the TS120S as clear winner with typically 10 to 15 percent more spots.

Rather than let that experiment run for a full 24 hours I put a new and better toroid preselector inline with the RSP1A.

Over the last hour that has more than reversed the spot count in favour of the RSP1A
eg. 83 spots on the RSP1A with the preselector versus 54 for the TS-120.

Be interesting to see how it goes over 12 hours.

On WSPR, the RSP1A is VK7JJ3 and the TS-120 is VK7JJ

Any search must be for 40m only.


PS: Here's the preselector, the tubular device is a ceramic screw adjustable capacitor trimmer.
And it's vinyl underneath, not checker plate aluminium. :-)
toroidpreselector2.jpg (253.36 KiB) Viewed 12834 times

Re: Receive Antenna & Impedance "Matching"

Posted: Mon Jul 15, 2019 10:31 am
by vk7jj
OK, I have finished comparing an RSP1A with a good conventional receiver on WSPR 40 meters, a Kenwood TS-120s with GPS-DO VFO.

Despite all my spreadsheets of results it boils down to just a few things:

a) a few dB of SNR makes a very large difference to hearing very low level signals.

It's worth remembering I'm trying to decode digital transmissions that are as low powered as 10 milliwatts as far as -33dB below the noise at a distance of something between 10,000 and of 17,000 kilometres and every transmission must be heard for a full two minutes in order to be successfully decoded. Thus AGC pumping from strong adjacent stations in particular, plus fading and interference can very easily cause a problem in that two minute period.

b) the most recent hour of reception is typical of the differences between the two.

Exhibit 1 shows the SNR comparison of the SNRs of the lowest 20 spots. VK7JJ is the Kenwood and VK7JJ3 is the RSP1A
jj-low.png (195.7 KiB) Viewed 12738 times
jj3-low.png (198.4 KiB) Viewed 12738 times
Exhibit 2 shows the SNR comparison of the highest 6 posts. See next post for the rest of Exhibit 2 (each post can only contain 3 images)