Receiver sensitivity calculation - WiMax
in [DSP]
|
Prev: New Tech Competition
Next: What would be the fastest dial-up access speed if phone systems had 144 dB of dynamic range?
From: lkc on 16 Nov 2009 11:55 Hi, the receiver minimum sensitivity level Rss is given as (in IEEE 802.16e): Rss=-114+ SNRrx - 10log(R) + 10log( Fs x Nused / Nfft) + ImpLoss + NF where SNRrx is the receiver SNR R is the repetition factor Fs is the sampling frequency ImpLoss is the implementation loss NF is the receiver noise figure Nused is the number of used carriers (include data, pilots and dc carriers) Nfft is the number of FFT points. From the equation, with lesser number of used carriers, Rss improves due to the factor: 10log( Fs x Nused / Nfft), or due to the lower bandwidth used. However, does SNRrx improve too with lesser number of carriers used? From my simulation, with Nfft =128, Nused=100, there is an SNR improvement of about 1dB (corresponding to 10*log10(128/100)=1.07dB) as compared to the case where all 128 subcarriers are used. Is my simulation correct? That is, does Rss improve due to 2 factors when lesser number of carriers are used? (1) Due to the SNRrx and (2) Due to lower bandwidth Regards, lkc
From: Eric Jacobsen on 17 Nov 2009 00:26 On 11/16/2009 9:55 AM, lkc wrote: > Hi, > the receiver minimum sensitivity level Rss is given as (in IEEE 802.16e): > Rss=-114+ SNRrx - 10log(R) + 10log( Fs x Nused / Nfft) + ImpLoss + NF > where > SNRrx is the receiver SNR > R is the repetition factor > Fs is the sampling frequency > ImpLoss is the implementation loss > NF is the receiver noise figure > Nused is the number of used carriers (include data, pilots and dc > carriers) > Nfft is the number of FFT points. > > From the equation, with lesser number of used carriers, Rss improves due > to the factor: 10log( Fs x Nused / Nfft), or due to the lower bandwidth > used. However, does SNRrx improve too with lesser number of carriers used? > From my simulation, with Nfft =128, Nused=100, there is an SNR improvement > of about 1dB (corresponding to 10*log10(128/100)=1.07dB) as compared to the > case where all 128 subcarriers are used. Is my simulation correct? > > That is, does Rss improve due to 2 factors when lesser number of carriers > are used? (1) Due to the SNRrx and (2) Due to lower bandwidth > > > Regards, > lkc > If the transmit power is held constant then, yes, the Rx SNR for a given channel will improve as the occupied bandwidth decreases due to power concentration. -- Eric Jacobsen Minister of Algorithms Abineau Communications http://www.abineau.com
From: lkc on 17 Nov 2009 20:35 >On 11/16/2009 9:55 AM, lkc wrote: >> Hi, >> the receiver minimum sensitivity level Rss is given as (in IEEE 802.16e): >> Rss=-114+ SNRrx - 10log(R) + 10log( Fs x Nused / Nfft) + ImpLoss + NF >> where >> SNRrx is the receiver SNR >> R is the repetition factor >> Fs is the sampling frequency >> ImpLoss is the implementation loss >> NF is the receiver noise figure >> Nused is the number of used carriers (include data, pilots and dc >> carriers) >> Nfft is the number of FFT points. >> >> From the equation, with lesser number of used carriers, Rss improves due >> to the factor: 10log( Fs x Nused / Nfft), or due to the lower bandwidth >> used. However, does SNRrx improve too with lesser number of carriers used? >> From my simulation, with Nfft =128, Nused=100, there is an SNR improvement >> of about 1dB (corresponding to 10*log10(128/100)=1.07dB) as compared to the >> case where all 128 subcarriers are used. Is my simulation correct? >> >> That is, does Rss improve due to 2 factors when lesser number of carriers >> are used? (1) Due to the SNRrx and (2) Due to lower bandwidth >> >> >> Regards, >> lkc >> > >If the transmit power is held constant then, yes, the Rx SNR for a given >channel will improve as the occupied bandwidth decreases due to power >concentration. > >-- >Eric Jacobsen >Minister of Algorithms >Abineau Communications >http://www.abineau.com > Then wouldn't it be better to always use lesser subcarriers as this will bring about better receiver sensitivity, some sort of double "advantage", with both the noise floor/thermal noise and the snr smaller? What is the catch here? Say we were to compare the performance of 2 systems with similar data rate. Should the receiver sensitivity comparison (holding transmit power same for both, simulation in snr) be independent of the BER comparison (holding energy per data bit at the receiver to be the same, simulation in Eb/No)? 2 different aspects?
From: Eric Jacobsen on 18 Nov 2009 01:08
On 11/17/2009 6:35 PM, lkc wrote: >> On 11/16/2009 9:55 AM, lkc wrote: >>> Hi, >>> the receiver minimum sensitivity level Rss is given as (in IEEE > 802.16e): >>> Rss=-114+ SNRrx - 10log(R) + 10log( Fs x Nused / Nfft) + ImpLoss + NF >>> where >>> SNRrx is the receiver SNR >>> R is the repetition factor >>> Fs is the sampling frequency >>> ImpLoss is the implementation loss >>> NF is the receiver noise figure >>> Nused is the number of used carriers (include data, pilots and dc >>> carriers) >>> Nfft is the number of FFT points. >>> >>> From the equation, with lesser number of used carriers, Rss improves > due >>> to the factor: 10log( Fs x Nused / Nfft), or due to the lower > bandwidth >>> used. However, does SNRrx improve too with lesser number of carriers > used? >>> From my simulation, with Nfft =128, Nused=100, there is an SNR > improvement >>> of about 1dB (corresponding to 10*log10(128/100)=1.07dB) as compared to > the >>> case where all 128 subcarriers are used. Is my simulation correct? >>> >>> That is, does Rss improve due to 2 factors when lesser number of > carriers >>> are used? (1) Due to the SNRrx and (2) Due to lower bandwidth >>> >>> >>> Regards, >>> lkc >>> >> If the transmit power is held constant then, yes, the Rx SNR for a given > >> channel will improve as the occupied bandwidth decreases due to power >> concentration. >> >> -- >> Eric Jacobsen >> Minister of Algorithms >> Abineau Communications >> http://www.abineau.com >> > > Then wouldn't it be better to always use lesser subcarriers as this will > bring about better receiver sensitivity, some sort of double "advantage", > with both the noise floor/thermal noise and the snr smaller? What is the > catch here? > > Say we were to compare the performance of 2 systems with similar data > rate. Should the receiver sensitivity comparison (holding transmit power > same for both, simulation in snr) be independent of the BER comparison > (holding energy per data bit at the receiver to be the same, simulation in > Eb/No)? 2 different aspects? Reducing the bandwidth reduces the amount of data that can be carried. Examine the capacity formula closely, and it becomes apparent that it is advantageous to occupy as much bandwidth as possible from an ultimate capacity perspective. It is always possible, however, to trade rate for SNR (it is often said as "trading rate for range"). This is what reducing the bandwidth does if the Tx power is kept the same; the SNR goes up, but the transmission rate goes down. -- Eric Jacobsen Minister of Algorithms Abineau Communications http://www.abineau.com |