Posts by SQ6QV

    I understand a liitle "how it's made" because of my many failed attemts to build something.
    I've got spectrum analyzer from the begining...and seeing DIY hamradio on spectrum analyzer is like seeing cheap fast food :D
    There is dirty everywhere, awfully dirty :D

    Still there is common that somone says "my switching power supply does not make noise at all"...
    He just didn't measured it...
    And those switching power supplies are powering our PAs and LNAs, gain blocks...and LDOs are not better.
    In my BladeRF I had oscillating (xtra low noise :D) LDO! It added +-250khz harmonics to signal output.
    And there comes phase noise..,Just count how many generators our signal "pass" from modulator to demodulator...

    But dirty signals on WB seems to be no problem to anyone...this is just "esthetics" problem I think.
    Some are more "pointy" and "eyes hurt" on them...but whats wrong if they are there just for our fun.

    Everything as Mike said....but...
    I've done a lot of experiments to have cleanest output possible and it is a problem more (&more) complicated than You might think.

    1st...Web monitor shows spectrum of your signal in Goonhilly, not power, not MER.
    To measure how good your MiniTiouner decodes your signal compared to power You need something to compare to.

    Only way (I think) is to transmit 1M or 2M signal and control its power to be exactly at same level as beacon and using same rolloff.
    then You have beacon MER for example 7.0db and Your signal MER 6.2 in my case.
    So my signal decodes 0.8db worse than beacon.

    I could write pages what I've tried to fix that gap and I still didn't discovered it yet. But I've made a progress.
    Mike's list is not complete.

    What I see on transponder most have problem with non linear amplifiers.

    Bitrate on QO-100 puts power over quality.
    QPSK and 8PSK works good with horribly overdriven amplifiers, and we (almost) all like to use more bitrate.
    Would You use Your amplifier at 20% of its maximum power capabilities?
    Or just solve 1db "signal quality problem" with that 6db more power available :D

    Second problem is modulator quality (SDR software & hardware). It is also not easy to make clean signal in digital domain, and as Mike said SDR hardware is not perfect (I think right words are "it is barely usable").


    Nice to hear any news in that matter!

    There is one feature that is easy to make, and we would be very happy about it - turning beacon off at 18:00 to 24:00 every day.

    Today I've sucessfully transmited 333K 32APSK signal which was received (only) by Renny PE1ASH with his 3m dish.

    I've done this test just to smile at:


    Users are encouraged to experiment with higher-order DVB-S2 modes at lower symbol rates (for example 333 KS 32APSK) to conserve bandwidth for other users.
    from: https://amsat-dl.org/en/p4-a-w…and-operating-guidelines/

    Power required for that signal was quite sick :) and reception on 0.9m dish was still impossible (mer 14 of mer 14.1 required).

    I think I'm the only one who achieved this "333K 32APSK" goal :)

    So please change guideliness: 32 APSK to 8PSK

    (as 333K 16APSK needs power comparable with 2M beacon)


    8PSK FEC 2/3 is really a maximum when there are so many stations that we care of bandwidth.
    -----
    Also It is worth to reconsider "overpower" message over 1M signals - making it absolute in scale (<7db) rather than relative to the beacon.

    I'm perhaps the only one who was decoded with 8PSK with this software (and it's no mystery why :D ).

    I've played with it some time and it works. I've used cheapest rtl-sdr.


    The problems are:

    1. With 2.4M sdr sample rate, only 1M or narrower signals are possible to decode.
    2. Locking is slow. We are used to MiniTiouners, and not patient.
    3. It needs much stronger signal.


    I've done some tests today and locking 8PSK needed D3 signal on MiniTiouner, and after lock it could go down to D1.

    It was 333K FEC 3/4 mer 12db+ signal


    It does not decode 16APSK, but I think technically it is possible, as two circles in constellation were visible.


    Screenshot by Rob PE1CME