Calculate power in RF emissions
You can use a spreadsheet to calculate a signal's total power.
Mark Briggs, Elliott Laboratories, Sunnyvale, CA -- Test & Measurement World, 5/1/2002
When you measure RF emissions with a spectrum analyzer, you may need to measure the total power contained in a spread-spectrum signal. FCC part 15, subpart E (Ref. 1) requires these power measurements for Unlicensed National Information Infrastructure (U-NII) radios used in wireless LANs. Some spectrum analyzers can calculate the total power over a frequency range for you, but most won't.
If your spectrum analyzer doesn't offer this feature, you can save data from the analyzer's frequency scan and import it into Microsoft Excel to calculate a signal's total power. I've written an Excel macro that imports data and calculates total power. You can download a spreadsheet that contains the macro. The spreadsheet comes with a sample data file so you can see how it works.
Before you can use my Excel spreadsheet, you'll need to save the test result data in a comma-separated variable (CSV) file. You can write your own program to do this, or the spectrum analyzer's manufacturer may provide off-the-shelf software that you can use. For example, Benchlink software from Agilent Technologies (Santa Clara, CA) controls the company's instruments and lets you store data in CSV files. You can easily import CSV files into Excel.
Once the data is in Excel, you can use my macro code to calculate a signal's power. You can modify the macro's source code, written in Visual Basic for Applications (VBA), to suit any measurement task.
Where's the signal?A U-NII radio transmits a spread-spectrum signal that covers a portion of the frequencies from 5.15 GHz–5.35 GHz or 5.725 GHz–5.825 GHz. But, since a typical spectrum analyzer has a maximum resolution bandwidth of 1 MHz, you must calculate the total power in a U-NII radio signal by summing the power for each 1-MHz band within the signal's frequency spectrum.
|
|
| Figure 1. This U-NII signal consumes 84 MHz centered around 5.77560 GHz. |
To calculate a signal's power, you must first determine its bandwidth and center frequency. Start by adjusting the spectrum analyzer's reference level to compensate for any signal loss introduced by cables or attenuators. Set the spectrum analyzer's resolution bandwidth to 1 MHz. Set the video bandwidth according to the FCC's guidelines. The guidelines are still in development so check with the FCC (www.fcc.gov) for the latest version. Then, adjust the frequency span on the analyzer until you see the –26 dB points on either side of the signal's peak. You'll then know the signal's bandwidth and center frequency. Figure 1 shows a spectrum from a U-NII signal with the spectrum analyzer's vertical scale set to 5 dB/division. This signal consumes 84 MHz of the 100 MHz between 5.725 GHz and 5.825 GHz.
The spectrum analyzer will provide the screen data to your software in either 401 or 601 frequency-amplitude pairs, depending on the analyzer you use. If you simply sum the power at each data point, you'll get a result that's much too high because the data points are less than 1 MHz apart. You need to add the power in adjacent 1-MHz bands to get the correct total power.
Before you can proceed with the measurement, set the spectrum analyzer's vertical units to watts because the power calculation requires a linear summation. Don't use dBm or dB(W) units because they're logarithmic.
|
|
| Figure 2. A comma-separated variable file created with Benchlink displays pertinent information about a frequency scan. |
From the signal's start frequency (cell A12), its stop frequency (the last-filled cell in column A), and the analyzer's resolution bandwidth (cell B3), the macro can calculate how many data points it needs to extract (the frequency span divided by the resolution bandwidth) and how often to extract them (the total number of data points divided by the number of data points to extract). In the scan in Figure 1, for example, the raw data contains 601 points from which the macro extracts 84 points (84-MHz spectrum bandwidth/1-MHz resolution bandwidth) at seven-point intervals (601 actual points/84 required points).
After extracting the points, the macro opens a new Excel file and places the extracted points' frequency and power (in watts) data into two columns. The macro adds the power levels of the data points to get the total power contained in the signal and places the results in cells at the end of the data columns. In practice, I've found that the results you get with the macro agree with those you get from a spectrum analyzer that performs this function.
To run the macro, download the Excel file and open it. Excel will ask you if you want to enable macros, so enable them. From Excel's Tools menu, select Macros and run the "get_data" macro. The macro asks you for the location of the file containing the raw data (CSV file). Next, you'll see a box that tells you how many points the macro extracted for the power calculation.
You can adapt the macro's VBA code to other measurement tasks. For example, you can change the code to provide band-edge power measurements. To get those measurements, calculate the power spectral density (per MHz) from a CSV file containing data taken with resolution bandwidths of 100 kHz or less, rather than 1 MHz. The finer resolution lets you measure the power at the edges of a signal's spectrum.
For more informationHurst, William, Telecommunication Certification Bodies (TCB) Tutorial and Workshop, Session 6, "Unlicensed Radio Service Equipment: U-NII Devices," July 10–12, 2001. NIST, Gaithersburg, MD. ts.nist.gov/ts/htdocs/210/216/nist-tcb-workshop-july-2001/s6-unii/unii-presentation.ppt.
Walkenbach, John, Microsoft Excel 2000 Power Programming with VBA, Hungry Minds, New York, NY. 1999.
| Reference | ||
|
