Pluto as a Digital Frequency Converter

[G4NZV]

Pluto maybe used as a Digital Frequency Converter to move a chunk of spectrum up to ~2.5MHz bandwidth from one band to another. This example shows reception of a 8Degree sidelobe of a 6cm DATV DVBS2 transmission from G4FRE over a 17.5Km path between Malvern and Tewkesbury using a Portsdown4 and Minituner tuned to ~2322.5MHz.

The technique of shifting chunks of spectrum is very useful as it opens up possibilities of DATV reception on frequencies outside the tuning range of the Serit tuners used in various projects.

Degradation of signal was crudely assessed by converting the QO-100 beacon IF from a non-locked LNB at ~742MHz to 2322.5MHz improving the MER by ~0.4dB.

The GUI shows the spectrum display of G4FRE's 5762.5MHz signal, blue line is the live spectrum, green line is with the maximum hold function enabled. (This is very useful in peaking the dish as the 'best' can be easily observed). If '0Hz offset' is selected a receive frequency can be directly entered otherwise it serves as an offset for fine tuning.

Caveats on this technique are:
- Certain frequencies will not work e.g. When the RX and TX are close together, when the RX is a multiple of the TX
- Band Pass Filters maybe needed for reception in some circumstances
- Approx. 30dB of attenuation is needed between the Pluto TX output and the Serit tuner input
- If the TX output is connected into an antenna then it should be remembered that the whole of the bandwidth converted will be transmitted plus harmonics

I used GRC v3.7 with gr-iio blocks on a i5 windows10 PC

Steve G4NZV

[Pa3fbx]

Nice,

I still struggle with gnu radio to understand the simple thing of getting it to work for me..
I see ideas of DATV relays 70cm in and 13 cm out using a pluto and a raspberry.

Any chance you create some image for a RPi to download and start easy?

Best regards Benno

[G4NZV]

I like the idea of using a Raspberry Pi4 & Pluto as a Digital Frequency Converter and I am working on a solution.
However, there is still an unresolved issue when using the Pi4 in that the TX output spectrum has discontinuities. IF the Pi4 work is successful I am happy to share the .grc file and will investigate the possibility of providing an image.

Steve G4NZV

[M0DTS]

Yes this works well.
You can make some very sharp filters in software to remove unwanted signals close in to the desired one, like we have on 2m and 70cm for example.

Rob

[radiogareth]

This thread needs greater exposure as a very handy way of shifting Non-Serit frequencies (9 & 6cms at least) into range.
I wonder if it would run OK on a RPI4 or 5, so you could do away with all that pesky downconverting, LOs, filters etc.....

Although to TX DATV on 6cms you either need to upconvert or usa a laptop/PC and DATV Easy 2.17, so it might as well run on a PC....don't suppose even a RPI5 would manage the work of DATVEasy 2.17...

Gareth (currently doing it the LOs, filters etc way...)

[g4eml]

I have just bought a raspberry pi 5 8Gb and can confirm that GNU radio companion seems to work well on it. It used to struggle when running on a raspberry pi 4.
I did a quick test running a simple frequency converter and it appeared to work OK.

Benno, to get this working I did the following...

Create a SD card using the raspberry pi imager utility. Selecting the 64 bit recommended Raspberry Pi OS.

Boot the RP5 to the desktop then use 'add programs' to search for everything that includes the word GNURadio. Select all of the results and install them. It may not be necessary to install everything but it probably doesn't do any harm.

GNURadio Companion will then appear in the programming menu and will support both Lime and Pluto.

As it works fine from the desktop environment it should also be possible to build a stand alone version that would work without a monitor.

If I get the time I may investigate if Langstone can be made to work on the RP5. At the moment is won't work because the framebuffer on /dev/fb0 is not available on the standard build.

Colin G4EML

[g8lce]

I like this idea a lot!

Tried with my RPi5 but not a GNU expert and could not get it to run (yet) but showing promise.

Here is the zipped two files I generated.

Data.zip

(5.38 KiB) Downloaded 13 times

This project would make the Portsdown and Ryde with an extra RPi5 and Pluto into a really 'can do anything' Machine!

Martin G8LCE

[g0mjw]

I have tried this, quite a long time ago now, (around 2011) using the Ettus USRP SDR as a transponder. The main issue is you need good filtering of the front end or you just have the same issues as with the tuner from out of band signals. The gain needs to be set correctly too and that can be difficult to get right in the field. It has to be such that strong in-band signals do not cause intermodulation that overwhelms the weaker signal you are trying to decode, while being enough that the added noise doesn't become a limitation. You can select narrower bandwidths, but be careful with the filtering you use. The phase response and group delay is very important for decoding phase modulated signals.

A better solution is to demodulate the signal with the SDR. Then a NIM is not required. All it needs is a decent efficient modem implementation that can do the same job as the NIM is doing. That is the difficult part, decoding the LDPC forward error correction with minimum implementation loss* is hard and uses a lot of processing power, so efficient coding is required, the sort of task FPGAs or dedicated hardware like in the NIM are good at, but I know Simon, G4ELI has been working on it and if anyone can manage it, he will.

Mike

*The implementation loss is a measure of how well the decoder does compared to an ideal decoder. An ideal decoder would search every possibility within the 64800 bit frame to determine the most likely match to what was sent. This is not practical to do in real time, especially at low SNR. SDR Angel, Lean DVB, Matlab https://uk.mathworks.com/help/wireless- ... coder.html and other packages use complex algorithms to solve this as best they can, buy they are not perfect. The result being the SNR required is higher than the theoretical minimum and the difference is the implementation loss, which can be less than 1dB or several dB.

[G4NZV]

Agree with Mike's summary of limitations of this technique, they agree with the caveats in my original posting.

Thinking about this a different way; if this could be thought of as a technique to receive 'any signals' not just DVBS/DVBT, its purpose is to frequency shift a junk (in this example ~2.5MHz bandwidth) of spectrum from one frequency to another. It will never be as good as a demodremod 'demodulate the wanted signal then immediately remodulate on a different frequency' which will remove the noise and other artefacts present. BUT if the capability to receive a 'strong cleanish signal', without the use of local oscillators and mixers etc. (filters still advised especially on transmit) then this technique could be reliably used.

As always with engineering on an amateur budget compromises in absolutely performance are made, the aim here is to try and increase activity on bands outside the range of minituner at the expensive of ultimate strong signal, ultimate low SNR performance by the use of SDR radios (Pluto, Lime, USRP, SDRplay(rx), RTL(rx) etc.) which already maybe present in the shack along with a PC or Pi etc.

I also considered the use of a processor only to solve this problem around 5 years ago and came to the same conclusion; an FPGA card with fast immediate access to large memory is essential to perform the LDPC lookups. Now with better processing and memory performance and/or newer FPGA cards with onboard memory then a solution might be possible.

Steve G4NZV