Hi,

I have used Noise Source to calculate the Noise figure of the FCOMMS2 board. I have modified the Balun and the matching to adjust to the band of interest (2GHz). Using Nosie source I have captured the FFT plots (I,Q data in binary format) under Hot and Cold conditions and have calculated the Noise figure using Y factor. I want someone form ADI to review if my calculations are correct. I have used Scilab to perform the calculations. I have attached the code. I want to ensure I am converting the sum of the IQ binary data into equivalent dB value correctly. Please let me know.

Steps:

1. Read the CSV files for Hot and Cold

2. Sample Rate: 30.72MHz, Band width: 20MHz

3. Validated the quality of the data by ensuring the mean is 0 for both I & Q data

4. Performed Sqrt (I*I+Q*Q) for each bin and added all of the bins

5. Divided the result in step 4 with sample size and took real value.

6. Converted the real value in Step 5 to dB level by using the formula 20*log(real value)

7. Performed the same steps for the other CSV file (Hot and Cold)

8. NF = ENR-10*log(10^((dB_hot-dB_cold)/10)-1)

Yes your calculation is correct.