Posts by DH2VA

    to make it simple for everybody: the S2-3200 is virtually useless for us. DVB-S2 only between 10MSym/s and 30 Msym/s and even DVB-S only down to 2 Msym/s.


    If you are using a PC anyways, get yourself a Minitiouner or Minitiouner-Express. Much more flexibility and wider userbase in the QO-100 community.

    DG7YEO please don't confuse AZ/EL (left/right/up/down) with skew. The latter is rotating the LNB in its mount like a cloc. The reason for doing so is that horizontal and vertical is defined by the satellite's local coordinate system and as it is not straight south from us, we have to rotate our LNBs a bit (13deg for central europe).

    Please don't ask me about direction.. I always explain it wrong and after all the alignment is straight forward and non-ambiguous.

    DG7YEO Have you corrected the LNB skew? It should not be fully upright but rotated a bit. The little things can be decisive now.

    Skew can be optimized to switch to 18V (H) and *minimize* the CW beacon. Then the WB beacon should be optimum (minimizing CW beacon give clearer result).

    the 5mm aluminium mounted flat underneath the PCB might be doing a better job at heatsinking than your (admittedly beautifully made) brass origami ;)

    om0aao if you really want to use SDR-console, please switch on the FFT statistics. This will include the FFT Bin size in the display. see here


    signal history shows =- 75.8 dBm which is matching the y-axis at the left side for the peak signal. Note the Resolution which is 19 Hz, this is the FFT bin size.


    Now I set the receiver in noise. The signal history shows -103.6 but if you look left, this is hardly the average noise power. The average is probably something like -111 dBm. So -75.8-(-111) = 35.xx dB in a 19 Hz bandwidth (this is the noise BW). 19 Hz is (almost) 13 dB more than 1Hz so this is then something around 48 dB in 1 Hz. Which for me is close enough to the 50-53 dB (1 Hz) noted in this thread. After all, this was eyeballed.. :)

    om0aao please check on my post about the SDR-Console readings.. They are pretty hard to interpret as the signal history is displaying peak and not average.. and the Noise BW of the FFT is not straight forward to see. So it's hard to compare.

    Or check on the sliders for noise and signal measurement.. it should be constrained to sensible numbers. DB8TF's screenshot was almost perfect (except for center frequency).

    I believe this is a 'feature' of the program. No signal peak, no SNR measurement. But with the correct settings, the readings in the constant-on periods are fairly reproducible.

    I am by no means an expert on Andy's software but make sure the sampling rate of the virtual audio cable is correct, also press 'peak to CF' and then another box with metrics should come up in the bottom right corner (your center frequency is way off). Still digging through the manual myself but numbers are converging now.

    got 54 dB SNR now..

    "But how can this be, you said 28dB before?"

    different software(algorithm) and/or settings.. and I think here lies the problem in our wildy varying numbers.

    We all know we should quote SNR related to receive BW. So most in the community quote it for a 2500 Hz SSB BW (as like for JT modes and many others). Others (like the amateur-DSN people) use 1 Hz as normalized BW and this is fine as well. The conversion factor is simply 10*log(2500)= 34dB.

    Our problem here now is that the different SDR packages we use have wildly different algorithms. Take SDR-Console as example.

    - no way of setting FFT bin size manually

    - signal history seems to display peak values, not mean

    - no direct SNR value given and no explanation of underlying parameters (noise BW?)

    SDR# has a SNR meter but they depend on the FFT bin size and sampling rate so no easy measurement possible as well.

    Andy Talbot G4JNT has a fine tool for doing such measurements:

    http://www.g4jnt.com/SignalToNoiseMon.zip


    When I let SDR-Console do the drift correction and route the audio via virtual cable to Andy's software I get 54 dB in 1 Hz for the CW beacon in an 88cm dish.

    At least that is my own goal. So after some brief testing, I canceled the idea of using a helix next to the LNB in my 88cm. I still have an AO-40 Mode-S G3RUH dish .. its 60cm with the genuine G3RUH patch feed optimised for it should be superior to the 88cm. But once set up, I can do easy comparisons.

    hello everybody I tried 3.5 m of h155 cable with two bnc that I loss 0.37 @2400

    Roberto

    With all due respect, I doubt this value as no datasheet of the various H155 variants suggests an attenuation of <0.1dB/m @2400 MHz. BNC is not optimal as well on 13cm, but this is another issue. My point is that the H155 cable is a 5-6mm OD cable and even a 1/2 inch cellflex coaxial cable (LCF12-50) only reaches 0.11dB/m.

    That said, 1dB difference is barely noticable in daily life unless you have QRO amplifiers or cable in front of the LNA :)

    Well, you can even connect the PA to a leaky dummy load and need even more. Or 100W into dipole.. at the end of the day, EIRP counts.