DVB-S -v- DVB-S2
DVB-S -v- DVB-S2
Following the release of Minitioune 0.5a which receives DVB-S2, I thought I would do some tests to compare it with DVB-S. According to the lectures at CAT16, DVB-S2 should give a 2dB advantage of DVB-S. These tests were conducted using a DATV Express running v1.23 software, 146.5MHz, H264 codec, 333ks/s, across the shack using aerials maintained in the same position. There is a slight difference in power output between the various modes, which I measured and corrected for, so that power output was consistent (-23dBm). The results received on Minitioune 0.5a were;
DVB-S MER 22dB.
DVB-S2 QPSK MER 15dB.
DVB-S2 8PSK MER 14dB.
The theoretical advantage does not appear. Perhaps those that understand the theory of this can explain why or tell me what I have done wrong?
73 Shaun.
DVB-S MER 22dB.
DVB-S2 QPSK MER 15dB.
DVB-S2 8PSK MER 14dB.
The theoretical advantage does not appear. Perhaps those that understand the theory of this can explain why or tell me what I have done wrong?
73 Shaun.
Re: DVB-S -v- DVB-S2
Shaun,
There is an issue with the Express DVB-S2 constellation so the MER figures you recorded are slightly misleading.
To get 32APSK to work with my Satlink I reduced the BPSK preamble reference symbols by 10% this has had the
side effect of making the MER appear very bad.
If I set the preamble to the correct level the MER measured on an old version of Minitioune jumps from 15 dB
to 44 dB. The newer version of Minitioune is less generous but I am not running that at the moment so I can't
give you the exact figures.
The DVB-S2 advantage was calculated firstly using mathematics then by computer simulation and finally using
a real world chipset by the DVB Consortium and the figures are correct both in theory and in practice.
The very latest version of the DATV-Express software contains a Gaussian noise generator so that I can simulate
a particular S/N and with that I get results (within 0.4 dB) of those quoted by the DVB Consortium.
Why my Satlink does not work when the preamble is set to the correct level I have not been able to figure out yet.
- Charles G4GUO
There is an issue with the Express DVB-S2 constellation so the MER figures you recorded are slightly misleading.
To get 32APSK to work with my Satlink I reduced the BPSK preamble reference symbols by 10% this has had the
side effect of making the MER appear very bad.
If I set the preamble to the correct level the MER measured on an old version of Minitioune jumps from 15 dB
to 44 dB. The newer version of Minitioune is less generous but I am not running that at the moment so I can't
give you the exact figures.
The DVB-S2 advantage was calculated firstly using mathematics then by computer simulation and finally using
a real world chipset by the DVB Consortium and the figures are correct both in theory and in practice.
The very latest version of the DATV-Express software contains a Gaussian noise generator so that I can simulate
a particular S/N and with that I get results (within 0.4 dB) of those quoted by the DVB Consortium.
Why my Satlink does not work when the preamble is set to the correct level I have not been able to figure out yet.
- Charles G4GUO
Re: DVB-S -v- DVB-S2
Charles,
Thanks for that explanation. Besides the MER results, the constellations were definitely tighter with DVB-S than S2. However, perhaps the best test will be to get a more distant signal near to the decoding threshold and see which holds up best - something to try next I'm out portable.
73 Shaun.
Thanks for that explanation. Besides the MER results, the constellations were definitely tighter with DVB-S than S2. However, perhaps the best test will be to get a more distant signal near to the decoding threshold and see which holds up best - something to try next I'm out portable.
73 Shaun.
Re: DVB-S -v- DVB-S2
The other issue is that there is an output power difference between DVB-S and S2 modes with Express
so you need to factor that into account.
To be honest the only way you can scientifically compare the two is with an RF channel simulator
there are too many variables in the real world
- Charles
so you need to factor that into account.
To be honest the only way you can scientifically compare the two is with an RF channel simulator
there are too many variables in the real world
- Charles
Re: DVB-S -v- DVB-S2
Isn't the MER related to the symbol error rate? In which case I would not expect to see a difference.(for the same modulation) 8PSK will be 3dB worse than QPSK.
Mike
Mike
Re: DVB-S -v- DVB-S2
MER(dB) = 10 log10(Psignal/Perror)
Where Psignal is the RMS power of the Signal vector
and Perror is the RMS power of the Error vector
So for both QPSK and 8PSK let us assume the signal vector power has magnitude of 1.0
Then for the same error vector the MER will be the same for both QPSK and 8PSK
As will the signal to noise ratio
but not the BER and EbNo
People might find this link useful
http://dsp.vscht.cz/konference_matlab/M ... Kaller.pdf
Forget the maths and look at the pretty colours
- Charles
Where Psignal is the RMS power of the Signal vector
and Perror is the RMS power of the Error vector
So for both QPSK and 8PSK let us assume the signal vector power has magnitude of 1.0
Then for the same error vector the MER will be the same for both QPSK and 8PSK
As will the signal to noise ratio
but not the BER and EbNo
People might find this link useful
http://dsp.vscht.cz/konference_matlab/M ... Kaller.pdf
Forget the maths and look at the pretty colours
- Charles