Voice Rx level sensitivity testing

  • Hi,

    A user service for the NB seems feasible to be done that will transmit audio files that will contain the Tx dB level (if calibrated can be how dB is down from the beacons ). This will be useful for someone to test his Rx sensitivity. A similar approach has been discussed before using CW. It can be implemented using GNURadio.

  • DB2OS is there any objection to operate the voice beacon at 27/6 from 5:00 UTC to 17:00 for example and open a vote that users can enter the maximum number entered, in order to have a comparison of received sensitivity? Each Tx holds 1-2 min. Repetition details can determined. If this service is to be included in the QO100 services any personal help is available.

  • Hi George,

    Many people might remember the very popular ZRO test on OSCAR-10, also discussed here: proposal for beacon Hi

    We thought about the possibility implementing this on the lower CW beacon a while ago.

    We just discussed your proposal in our regular internal QO100 team meeting and the first opinion was that traditionally Wednesday = Experimental Day would be in fact a good day for such a test.

    Since you need to be attending the station when it runs all day, we understand that you might prefer a weekend which is also helpful to receive more reports from a wider audience.

    So from our perspective June 27th from 05:00 UTC to 17:00 UTC would be fine.

    Depending on the acceptance and results, we could even later decide if we want to install this as a permanent service with Uplink from Bochum or if it will be say a monthly event, with some sort of Awards for the stations which received the "lowest level".

    I think it could be very helpful for users to understand what their receive performance really is, as we see people promoting 30cm dish beside the recommended size of 60cm or more...

    73s Peter

  • Hi,

    A common QO100 NB user question is: what my Rx sensitivity? What is the minimum signal I can receive? This is not always easy to be measured. This need has to cover this test.

    The approach is to transmit numbers as voice 0 to 30, each one with a variable attenuation equal to the number transmitted. So a bigger number possible to be read, is mean a more sensitive Rx system exists.

    Also every one is happy with its system. But it could be very useful for all of us to compare our Rx sensitivity with others. So a Vote will be opened to get relative results.

    The test is planned to be transmitted at Sunday 27/6/21 from 05:00-17:00 UTC at 10.489.940. It consists from a Voice ID and 0-30 level transmissions (update) in 3 dB steps, repeated 3 times and then a final voice ID is to be send. Each sequence will start every 10 minutes and holds about (update) 2:25 minutes.

  • What is your receive SNR measured with a 50 Hz CW filter centered at the end beacon? 3

    1. less than 25 dB (0) 0%
    2. Between 25 but less than 28 dB (0) 0%
    3. Between 28 but less than 31 dB (2) 67%
    4. Between 31 but less than 34 dB (1) 33%
    5. Greater than 34 dB (0) 0%

    What you are discussing is related to the link budget, there is a RX and a TX part, the RX part concerns your receive sensitivity at 10 GHz and the TX is the other way around.

    Let's do something everyone can do, and that is to look at a unique signal of the satellite from both end beacons where it is assumed that they always operate on the same power level.

    What you do is to test your signal to noise ratio at the begin or end beacon of the NB transponder, so you measure with a 50 Hz wide CW filter the dBm's of a continuous wave transmitted by the CW beacon, and one repeats that measurement 5 kHz below the beacon in a region where you only see the noise floor.

    In the end you get two dBm values and you subtract them, in my case I find 28 dB and this is a value corresponding to my receive sensitivity. The former receiver performance where I had a semi-rigid problem between the feed and the downconverter and a smaller dish it was 20 dB which is in my opinion poor receive sensitivity, that problem is gone so I can hear others better.

    Assume the value of 28 dB as an artibrary reference, assume a factor 2 (or 3dB) in the sensitivity test and complete the following poll, this is a QRM free test everyone can do.

  • Hallo Ernst,

    interesting idea. Would you please give a short description of your equipment and method to make the results comparable.


    You only need SDRconsole and a SDR, go to view -> signal history and you can get an the situation as shown, best guess the peak dBm values in both 50Hz CW decoders and subtract them, this morning I got approximately 29 dB, meaning also that there is nothing attenuating the satellite signal (like rain water on the canopy of the antenna feed). Signal history in sdrconsole even has the possibility to download the dBm values in a csv file that you can import in a spreadsheet. I focus on peak values of what I see in SDR console.

    You will see that the SNR that you find is mostly unaffected by settings of the SDR, even the bandwidth of the demodulator should not matter too much in my opinion. Important is that you get the maximal signal of the CW beacon for the moment that it is continuous and the real noise floor somewhat but not too far away from the beacon.

    Martin DM4IM did this test also I learned during a QSO, he got 34 dB but his antenna is taller than my TDS88 from Triax which I rate at 90cm diameter, his is 150cm, the difference in the antenna gain is like 4 dB (see the 31-May article on my blog), the rest of the differences are due to downconverter or LNB differences or simply the antenna efficiency. I think these are interesting tests to compare. It says in my opinion everything about receiver sensitivity.

    I know that SDRplay's rspduo results in realistic dBm values, used a HP stepper attenuation device, a HP power meter and a signal source to calibrate the dBm scale in SDRconsole, it is fine in my opinion.

  • Hallo Ernst,

    thanks for your informations. Unfortunately the Console ist not the best instruments to do such tests. To compare SNRs of a specific station between stations at different locations, there are some more parameters which has to be kept comparable. The zoom value for example has an enormous influence on the level meter. When I do the test, you described, I find 35 dB peak of the idle carrier of the CW beacon. But we know, this is much to much. To get reliable results other level meters with other instruments have to be used.
    The Console Level Meter is good for optimizing the own dish or LNB position. From former threads we know, the transponder signals will not override 20 dB SNR, also with larger dishes. If you can see the transponder noise floor clearly, your receiving system is ok. There is nothing left to do. SNRs of the stations you receive, depends on their transmit power only.
    From my unimportant point of view, when turning a dish to the sat, the increase of the noise floor is a valid criteria to optimize position and checking the own receiver. And, this absolutely independent from other parameters. We talk about 3 KHz bandwith.
    Please don't mind me, discussing it a bit controversly.


  • Hallo Ernst,

    at least I agree with you :-). The Console levels do not show the real SNR but are rather comparable under the discussed circumstances. May be, Simon will find the motivation to implement another DSP-routines one day. Probably just like PowerSDR. It's level measurement works perfect.
    In summary we are more than satisfied to have a program like the Console. All properties together give excellent satellite operation possibilities. Not to forget the Pluto. Who watched - as i did - over the last two years the deveplopment of this program has to bring a highly estimation to Simon.

    By the way, a friend, DL4JAL, desinged an USB-Control-Unit with Audio in- and output, PTT-Switch and a generator for CW. More details can be found on his website. All the functions are summarized to one USB-Port, controlled by a PIC.


  • Hi Ernst,

    thanks for your investigation into SDRconsole and finding a work around for the "real" SNR.

    The general principles has been discussed in another thread here in the forum where Achim DH2VA also explained the effects of small and large dishes and where the optimum is. Generally speaking, if you can clearly see the transponder noise there is only little to add...

    Maybe we need a new thread with a summary of all individual threads with investigations and conclusions as below:

    However the "Voice ZRO" beacon proposed by George is mainly a simple test for those who do not use a "PC", but a traditional converter and RX/TX.

    73s Peter

  • What is the maximum number that you can hear? 8

    1. 0 (0) 0%
    2. 3 (0) 0%
    3. 6 (0) 0%
    4. 9 (0) 0%
    5. 12 (0) 0%
    6. 15 (2) 25%
    7. 18 (4) 50%
    8. 21 (2) 25%
    9. 24 (0) 0%
    10. 27 (0) 0%
    11. 30 (0) 0%


    In order to measure each user, its Rx sensitivity, without any instruments, numbers is transmitted as voice from 0 to 30, each one with a variable attenuation equal to the number transmitted. So a bigger number possible to be read, is mean a more sensitive Rx system exists.

    The test is to be transmitted at Sunday 27/6/21 from 05:00-17:00 UTC at 10.489.940. It consists from a Voice ID and 0-30 level transmission in 3 dB steps, repeated 3 times and then a final voice ID is to be send. Each sequence will start every 10 minutes and holds about 2:25 minutes.

    A vote exists to compare the results.

  • 2600 Hz SSB filter in SDR console over end beacon remains steady at 28 dB, just export the values to an excel sheet and calculate the dBm difference within excel. It doesn't matter really how you do this. For instance take a mean dB value through the values at the lowest and the highest level and subtract them. I take the noise measurement a few kHz away from the beacon, and the signal measurement over the beacon. I find no convincing difference with the earlier described CW-U 50 Hz method.

  • Results of the ZRO beacon test, I submitted 18 but want to remark that not all pulses are sent at a predictable strength, for a next test maybe try a 1000 tone and ask people how many tones they heard?

    Hi Ernst,

    You are my mind ! The WAV files have been normalized but I have thought that a more accurate test will be to say the level and then a tone or CW Vs for example is transmitted. Update: for award some unknown ham radio words can be transmitted for example.

  • Well, i can tell that i hear level 24 , even 27, but only that it is there, just a bit over the noise. If i was meant to understand level 24, i would fail. So I voted 21.

    So to really determine how deep you can hear in the noise, random words do the trick. Or maybe a string of 5 figures in cw, like the old ZRO test on AO-10. 73, martin