Posts by OE3GBB

    Hi George,

    Sorry, but I am not the right person to discuss that topic, I am just a tester.
    As far I have understood, the sampling of the voice in codec2 is 20ms and 40ms for lower bitrate modes. So 50 Baud is a pretty good match for any type of modem. To get the whole coded data over the channel, David has used several carrieres in parallel and used OFDM to get them as close as possible. Any other modems would be maybe possible and there might be some information about the theory on or
    In my view the easiest approach to get a best result regarding audio quality without neural network is to fill the 2700 kHz bandwith with as many carriers of 50 Baud. Using differential coding helps to reduce the BER.



    Hi George,

    very interesting and good work!

    From the experiments with the original FreeDV mode2020 we know that OFDM and QPSK is not working over the satellite. The reason probably are the frequency / phase changes by LOs and doppler. So the quick work around by David Rowe was to change QPSK into DQPSK. With stable LOs on TX and RX this is now working without problems over QO-100. Maybe you also want to try differential QPSK?

    73's de Gerhard

    Today I have successfully modified the Opticum Twin LNB LTP-04H with S/N19-01-0xxxx.

    I left the crystal on the backside. I have cut the trace beween "2" and RDA3567. I have cut the meander trace "3" and soldered 470 nH in between. At "4" I soldered 100nF to cut the DC path to the PLL. The reason is that I am using a L-C-L LPF consisting of 2x 220nH and 100pF instead of the usual series resonance filter.

    At "1" I soldered 220nH. From the other side of this inductance I soldered 100pF to ground "5" and another 220 nH, building a LPF. The LPF was connected by a thin wire to the 100nF capacitor at "4".

    The right F-Connector was used as WB connection of the AMSAT-DL converter with LO 24 MHz. Using just 1 meter of coax, the signal level on 24 MHz was sufficient to lock the PLL. Did no further level tests for the LO at the moment.

    The cover has to be grinded between "1" and "5" not to have a short circuit.

    Test with a little 10.5 GHz radar source was successful.
    Tomorrow I will hook the LNB to my POTY and I hope I can report sufficient sensitivity and LO stability afterwards.

    73's Gerhard

    No, it is all different supplies. I was thinking therefore it could be a ground loop. Will try to feed the Plutos with HF-transformers next and avoid a DC-connection.

    I have found a solution. It was quite a simple error of my setup. I did have both Plutos at the same IP adress! Grrrrrr!!!

    Tnx, so the two vias are the connection to the crystal and the two "pads" are the capacitors. I think is should be possible to leave the crystal, cut one trace and use the other "pad" = capacitor as part of a voltage divider / LPF. I did not find a data sheet of the chip, so I don't know which side is the output of the oscillator.

    Opticum Twin LNB LTP-04H and S/N19-01-0xxxx is using RDA3567 and the crystal is on the backside. I think it is possible to cut one of the traces to the crystal right behind the Chip and before the via. LO input should be on the other trace, having a pad for soldering. Will try to modify and test.

    No, it is all different supplies. I was thinking therefore it could be a ground loop. Will try to feed the Plutos with HF-transformers next and avoid a DC-connection.

    Hi Thomas, I have the same problem. I do have a double unit and do want to combine them for double power. The right amplifier is doing well, but the left one is oszillating. Did you find a solution yet?

    An interesting problem came up yesterday, when I was using two Plutos in parallel from the same LO. One pluto for NB and the other with Evaristes firmware for DATV. Both connected via LAN to PC and LO connected by -3dB splitter (in this case LO from the 40 MHz output of the AMSAT-DL-Converter V3). When I switch on the second Pluto the first one is stopping. It seems that the LO signal is dropping because the decoupling by my splitter is not sufficent. Any ideas? Will have to use a distribution amplifier for frequency standards?

    The Opticum Twin LNB LTP-04H up to the S/N: 18-08-0 xxxxx is in any case easy to modify.

    For the S/N with 18-12-0xxxxx and 19-01-0xxxxx we are currently investigating whether a modification is possible with reasonable effort.

    At the moment we still have a larger quantity of the LNB LTP-04H from the 18-08-0 series, which are now being modified. We ask for some patience here

    Hello, I have a model 19-01-0xx. Any news about modification?

    73's Gerhard

    Hi Rasto,

    many thanks. Now I am beginning to understand. Before knowing your paper I was thinking that it is sufficent to choose a feed where the -10dB angle of the feed equals the SA of the dish. But it is much more complicated :-). The approach to find a feed for a basic paraboloid and considering AET is clear now. But you also showed that without simulation there is no way to find the least 1/10 of dB Gain. It also shows that the whole setup offset-dish and Helix feed is quite forgiving.
    Thanks again.
    73's Gerhard

    Hi Rasto,

    thank you for all this very interesting informations. It made me get deeper interested in parabolic theory and made me read some technical literature about. So I now understand that for offset dishes the values for f/D and SA in the data sheets do not correlate. I allways thought that we should choose the feed to get -10 dB (or -12dB) for the actual SA. But according to your paper that is only true for H=0. Do you have an easy to understand explanation, why we have to first find the f/D for a virtual dish having same SA, and than the find the optimal gain for that new f/D? Why is it not sufficient just to design the feed for the SA of the dish?

    Thanks for your help!

    Hi George,

    I checked your GNURadio graph. Congratulations! I just know a little about it. Tried it a few years ago, but no success. Maybe we could discuss a repeater setup on 70cm for FreeDV using Pluto?