topic Re: BER in Power Spectrum in LabVIEW

BER in Power Spectrum

tintin99 — Mon, 04 Jun 2012 09:46:24 GMT

I have a question about BER and SNR and I would really appreciate your help.

I have a Power Spectrum(live with the presence of interference) and I would like to know in what parts of the spectrum I can achieve a target BER (lets say 10E-3)

My suggestion was to first determine the signal to noist ratio from the graph (for BPSK ) and then find the available positions in the spectrum that can give me this SNR

so lets say for 10E-3 BER, the En/No= 9 dB (from the graph)

so Eb (dB) = 9 (dB)+ No(dB) or Eb(dBm) = 39 (dBm)+ No (dBm)

now my questions are?

1- instead of BER can I focus on SNR and are my calculations correct?

2- noise floor in my power spectrum is observed to be from -110 dBm to -130 dBm so it is not a constant value .what should I do in this situation to find Eb?

3- what about interference? because of WLAN the received power might be -50dBm so in this situation what the appropriate SNR should be to achieve that BER?

is it correct to say Eb should be= -50+39 (dBm)

I thik finding the avaiable postions in the spectrum (especially 2.4 GHz) to achieve a certain BER or SNR should not be as easy as I think.

Re: BER in Power Spectrum

nyc_(is_out_of_here) — Mon, 04 Jun 2012 13:26:12 GMT

@tintin99 wrote:

Hi

I have a question about BER and SNR and I would really appreciate your help.

I have a Power Spectrum(live with the presence of interference) and I would like to know in what parts of the spectrum I can achieve a target BER (lets say 10E-3)

My suggestion was to first determine the signal to noist ratio from the graph (for BPSK ) and then find the available positions in the spectrum that can give me this SNR

so lets say for 10E-3 BER, the En/No= 9 dB (from the graph)

so Eb (dB) = 9 (dB)+ No(dB) or Eb(dBm) = 39 (dBm)+ No (dBm)

now my questions are?

1- instead of BER can I focus on SNR and are my calculations correct?

2- noise floor in my power spectrum is observed to be from -110 dBm to -130 dBm so it is not a constant value .what should I do in this situation to find Eb?

3- what about interference? because of WLAN the received power might be -50dBm so in this situation what the appropriate SNR should be to achieve that BER?

is it correct to say Eb should be= -50+39 (dBm)

I thik finding the avaiable postions in the spectrum (especially 2.4 GHz) to achieve a certain BER or SNR should not be as easy as I think.

This is LabVIEW forum. Not a how-to-solve-your-RF-problem forum.

Re: BER in Power Spectrum

tintin99 — Mon, 04 Jun 2012 14:26:28 GMT

you are right my question is not about LabVIEW programming although I would like to use Spectral measurement and Modulation toolkit in the end to find the available positions in the spectrum to achieve a target BER . Is there any section or threads for RF problems like this in this forum?

also

I would really appreciate any help regarding these questions

Re: BER in Power Spectrum

Yupeng_Jia — Mon, 04 Jun 2012 17:52:02 GMT

Hi tintin99,

My personal opinions are as follows:

1- instead of BER can I focus on SNR and are my calculations correct?

BER is an inverse proportion to SNR. You can find threshold SNR via target BER if you know the channel statistics, modulation scheme, etc.

2- noise floor in my power spectrum is observed to be from -110 dBm to -130 dBm so it is not a constant value .what should I do in this situation to find Eb?

Note that Eb/No is different from SNR. If your input is complex signals, you can find Eb/N0 as follows

Eb/N0 (dB)= SNR(dB) + Tm/Ts(dB) – m(dB)

Tm is the signal’s symbol period and Ts is the sampling period, and m is the bit number per symbol.

Depending on your specific application, you can pick either peak noise floor or average noise floor.

3- what about interference? because of WLAN the received power might be -50dBm so in this situation what the appropriate SNR should be to achieve that BER? is it correct to say Eb should be= -50+39 (dBm)

Theoretically it is correct, but you need to consider the difference between Eb/N0 and SNR as discussed above.

I would suggest you posting at edaboard.com or other discussion forums to seek for answers of communications concepts. If you came across problems using NI RF products, please try posting at RF Measurement Devices.

Re: BER in Power Spectrum

tintin99 — Mon, 04 Jun 2012 21:05:13 GMT

Hi Yupeng

Many thanks for your reply. So lets say En/No is 9 dB from the graph for a target BER and symbol rate is 25 MS/s and sampling rate is 100 MS/s and we are using QPSK then 9 = SNR(dB) +10 log 4 - 10 log 2 --> SNR= 6 dB or P (signal) should be = 6dB(36 dBm) +noise floor+interference to achieve that BER. Is it right?

Is there any method to measure noise floor just by using spectrum analyser? Although as you said I can use the peak noise floor as well.

and also many thanks for your suggestion about communication forums.

Re: BER in Power Spectrum

Yupeng_Jia — Mon, 04 Jun 2012 23:26:13 GMT

hi tintin99, the calculation sounds for me.

For the second question, I'm not sure which application you're considering, so it's hard to answer. If you are measuring the dynanic performance, I would go for the peak noise floor. To help us provide accurate answer, you need to provide more information about your application. For example, NI-6115. You will find the following link useful : Noise Floor of NI-6115.

In case you're only interested in theoretically calculating the noise floor, please check radio receiver noise floor. According to the definition, you can calculate the noise floor as follows:

Noise floor = -174 + NF + 10 log Bandwidth

Where NF is the noise figure.

I hope my answers help. Again, for questions related to communiations concepts, please try the discussion forum I mentioned above. For NI RF measurement device, please post at RF Measurement Devices