noise in a WLAN system

Simulate as in calculate or on the bench?

For the bench test, any 802.11 source of spread spectrum transmission is considered noise if it's not correlated with the desired receive signal. I use an old Teletronics frequency hopping client radio as a noise source. The frequency hopping give it a nice flat frequency response as seen on the spectrum analyzer. I sometimes wish that it worked somewhat out of band but for the price, I can't expect much more.

For calculations, it's just the usual Boltzmans Constant noise floor times bandwidth thermal noise calculation.

I can't tell where you devined the 10dB noise figure (not noise factor), what type of modulation you're using, what you're working with or even what you're trying to accomplish. See:

a reasonable explanation of the calculations and techniques involved.

X-No-Archive Thank you for the information. I meant Noise Figure of 10dB. I read in different articles that NF of

5-10 dB is considered for systems based on IEEE 802.11 WLAN.

The problem is that if only thermal noise will be considered as the noise source in the system the noise power density will be around

10^(-20) and as a result SNR at locations even far from AP will be very high!

For example if we calculate the received power[dB] as transmit power minus path loss minus noise power, considering the 802.11a parameters (carrier frequency at 5 GHz and transmit power of 100 mW) with path loss exponent of 3 at distance of 20m from the AP the SNR will be around 20dB! Compared to reality this is too high!

Jeff Liebermann wrote:

Why?

Read again. 10dB is a tolerable number for the *SYSTEM* noise figure which includes everything between the antenna and the demodulator. The actual minimum acceptable noise figure is heavily dependent on the type of modulation, data speed, and desired BER (bit error rate). My guess(tm) is that the 10dB figure is for a BPSK 1-2 Mbit 802.11 system. This doesn't apply to an ODFM 802.11a system.

Well, I can't check your calculations without seeing them first. I'm also not going to try and reproduce your calculations without the slightest clue as to which system parameters you're using, and which assumptions you're making. However, I do have a guess. The usual mistake is to forget to include the 10dB minimum (FCC required) spread spectrum processing gain.

You forgot the antenna gains and coax losses.

You can subtract the noise power but getting a value of noise power is a bit tricky as it's affected by the noise figure of the receiver front end. If you're using the occupied bandwidth of the spread spectrum signal as your noise power bandwidth, you're doing it wrong. Think digital filter and demodulator bandwidth.

I suggest you switch to using the receiver sensitivity at a reference BER, which will yield a corresponding S/N ratio at the demodulator depending on modulation type and speed. Model the system from the transmitter to the receiver input seperately from the receiver input to the demodulator output.

5GHz or 5.7GHz?

Path loss is measured in dB and is easily calculated.

'm too lazy to work the numbers.

Path loss exponent is used to determine the variation in path loss caused by various edge obstructions and atmospherics and should not enter into your equation unless you're looking for a worst case analysis or doing interior reflective ray tracing. The value is also empirically derived somewhat of a guess. At this point, I have no clue as to why you introduced loss exponent.

The system noise figure and the minimum demodulated S/N ratio are not the same number. The S/N ratio is more commonly related to the modulation type and the BER.

I have a feeling this is your homework project. I don't mind helping but most schools have better resources available. If you would kindly disclose: 1. What you're trying to accomplish? 2. What you have to work with? I'll try to be more helpful.