Beiträge von G3WDG

    Hi Achim


    As expected, WB beacon dropped out here on my 76cm dish, which gives an MER of 6dB under clear conditions. Sun noise was measured yesterday as 5.8dB with this dish.


    At 1012z MER started to fluctuate between 6 and 5, with sun at 146.4/24.3 deg az/el. Satellite is 147.7/25.3.


    MER continued to degrade and I lost video at 1016:30 with sun at 147.5/24.7


    Video back at 1024:31 with sun at 149.6/25.3.


    MER back to 6dB at 1030 with sun at 151/25.7.


    73


    Charlie

    Hi Matteo


    Three different software packages.


    Spectravue for sun noise measurements, with homemade SDR based on Softrock. I'm not sure if this will work with RTL based SDR's. Hopefully someone else on the forum can advise.


    Minitournier for measuring MER of the WB beacon.


    RSP-SAS Spectrum Analyser in conjunction with RSP1A SDR for comparing the WB spectrum above noise.


    73


    Charlie

    I just confirmed the measurements using Sun noise. With helix fitted sun noise was 4.5dB, without helix present it was 5.8dB. Correcting for S+N/N, this equates according to my sums to a drop of 1.9dB in system receive performance caused by the helix.


    73


    Charlie

    One issue that I have been thinking about with helix feeds is blockage from the helix structure in front of the feed pipe.


    I have a simple feed here with 22mm water pipe protruding a few mm from the reflector of a 2.25turn G3RUH feed. The feeds works great on my big dish, but have been testing it today on a 76cm prime focus dish with approx 0.4 f/D.


    Two tests so far, measuring the spectrum and MER of the WB beacon. Both tests show significant degradation when the helix is present (I unsoldered it from the feed while the feed was mounted in the dish, changing nothing else). The helix's reflector plate was left in place.


    For marginal systems, particularly on the WB transponder, this would suggest that while simple to construct, a helix feed may be inferior to a patch. For the NB transponder I doubt if there is any discernible difference (from blockage) as the transponder noise floor is the limiting factor for most of us.


    Upper screenshots are with helix present, lower two are with it removed. Spectrum is at 5dB/div.


    There wasn't any measurable difference in beacon MER with the helix terminated with 50 ohms, or left open circuit.


    There could be advantage in moving the open end of the feed further into the helix to reduce blockage, but as I understand it that would move the phase centres of the 2.4 and 10GHz feeds apart, leading to loss in gain for one of them. So some tradeoff may be possible.


    Lots to experiment with these days!







    73


    Charlie

    For the NB transponder, the noise floor is so high that I would not expect to see much difference in S/N of downlink signals. The situation with the WB transponder may be more noticeable as the transponder noise floor is at a lower level, so the sun may well cause a noticeable degradation.


    I'm testing a 76cm dish (for /P operation) and see about 5.5dB sun noise with it, measured using Spectravue continuum mode. The cold sky reference was at the same elevation as the sun. Clear blue sky conditions.


    73


    Charlie

    I agree it is probably best for now to have FT8 located at dial frequency .600. This will place all FT8 'channels' in the correct segment, and would occupy a few kHz above .600. .610 dial frequency could then perhaps be a calling frequency for other narrow band JT modes such as JT65 and QRA64, with activity spreading (upwards?) as necessary.


    Busy commissioning a 76cm dish with dual band feed. Hope to back on the air shortly.


    73


    Charlie

    I think at around the 1.7 to 1.8 time there was a change to the way the frequency table was stored. If you end up with an empty frequency table, the fix is to reset the frequency table:



    This should be a one time operation.


    Edit: Forgot to mention that some of the modes (FT8 and MSK144) changed to 77 bit messages in v 2.0, meaning that earlier versions of WSJT-X are now incompatible with the latest version.


    Also, the JT65 decoder in v2 is considerably more tolerant of drift.


    73


    Charlie

    DD4YR


    I did some further tests here just now and was getting FT8 decodes reliably down to -20dB. This seems to be about the threshold for successful decoding here.


    This is about 10dB below what I can see on the BATC WebSDR waterfall. SIgnals decoding at -10dB are just visible.


    73


    Charlie

    Hi Robert


    Yes - I have one PC connected to the TX and another connected to the RX and can decode myself with no problems. In the first round of tests (with the locked LO but poor signal quality) I got some FT8 decodes down to -20dB. I'll try some more tests tomorrow morning (I may have to find some more attenuators to put in the TX path :-). DT generally between 0.4 and 0.5s.


    I usually test with dial frequency mid band in the NB digital modes segment, QRM permitting (have not had any QRM yet).


    73


    Charlie

    The drifty xtal oscillator chain has now been replaced by the original 10MHz locked synthesiser, with a redesigned loop filter. It now has a much cleaner spectrum.


    After switching back to the locked LO, I found that FT8 was then very reliable.


    73


    Charlie

    DD4YR I was waiting for a confirmation email from that site but never got it (also checked spam folder). I chatted to Jean-Pierre on batc wb chat and he said I had been registered OK for two weeks. I then tried logging in and it worked.


    73


    Charlie


    PS I am making progress learning MMSSTV....

    Yes, agreed you need a very stable link to be able to work with FT8. My unlocked LO is marginally good enough for FT8 (and JT65A). I decoded you today a number of times with JT65A. My RX is GPS locked here.


    I found today that with my slightly drifty LO, I can work with JT9F fast with 5 and 10s periods. I received myself via the BATC Web SDR for these tests. It seems very stable. Also tests showed JT65C was better than JT65A.


    Also tested JT9E (fast), MSK144 (wideband), and ISCAT B. All seem fine with 5 sec periods.


    Given that in most cases S/N margins are good, we may not need to use modes with 60s periods, and use of these will be compromised if frequency stability is going to be problematic for many stations.


    I managed an easy QSO today with OZ1CTZ using JT9F. I also managed a QSO on FT8 but this was more difficult.


    Of course as technology improves, FT8 will still be interesting as you can pack so many users into a very small bandwidth.


    73


    Charlie

    g3wdg1@gmail.com

    Hi Digi folks


    Following change of LO here to a clean one that drifts more, I am trying JT9F 10s periods and it is working very well. Drift is probably too much for FT8 as it is also problematic with JT65A. 10s periods are good too as it progresses QSOs faster!


    73


    Charlie

    Thanks, Robert


    Slight problem here - my 10MHz locked LO for the upconverter is dirty and I changed this afternoon to an old free running xtal chain. SSB signal now much better but I think the stability is now not good enough for FT8. Tomorrow I fit a crystal heater and see if that makes it good enough.


    I can receive OK, so happy to try to decode you. With the present LO, JT65 should be OK.


    73


    Charlie

    Achim


    I have measured TX power here now. To get a comfortable but not excessive signal back, I am running about 200mW into 1dB loss to the feed with 3m 0.3 f/D dish dish and 2.25t helix feed.


    Transponder noise is now about 15dB - seems less than yesterday, but I may not have the 10GHz feed exactly at the focus.


    Testing now with FT8 running 200mW is showing decodes at +7dB S/N


    73


    Charlie