Quick ACPR Analysis Performs Necessary PA Simulations

Traditionally, power amplifiers for communication systems have been designed to meet specifications like 1-dB gain-compression output power or third-order intercept point. Such tests can be made with one or two sinusoidal input tones. In reality, however, these power amplifiers are usually required to meet a certain adjacent-channel-power-ratio (ACPR) specification or satisfy a spectral mask at a particular output power. These results may not be obtained by using sinusoidal input signals.

To complete these simulations more accurately, one may use the Circuit Envelope simulator in the Advanced Design System (ADS) RF and microwave design software. It may be time consuming to obtain swept-power results, however. At each input power level, the modulated input signal must be swept long enough to attain sufficient spectral resolution. Such resolution is needed to accurately calculate output power and ACPR.

A faster method exists for simulating the ACPR and output power of an amplifier versus input power (with a modulated input signal.) A one-tone, swept-power, harmonic-balance simulation, which typically requires only seconds to run, is required to generate the amplifier's gain-and phase-versusinput-power curves. With the magnitude and phase of the modulated signal, one can then calculate ACPR and output power versus input power. This ACPR calculation is performed using equations on the data display. As a result, not a lot of additional time is required to calculate the ACPR.

The calculation of ACPR or output power is done in several steps:

  1. Generate curve-fit expressions for output voltage and output phase versus input voltage from the one-tone, swept-power, harmonic-balance simulation. The top two plots in the figure show how well the curve-fit expression matches the harmonic-balance simulation.
  2. Obtain the modulated input signal versus time—either from some other simulation or from a signal generator.
  3. Calculate the magnitude and phase of the output-signal voltage versus time from the curve-fit expressions as well as the magnitude and phase of the modulated input signal versus time. This computation may be repeated for a range of input-signal amplitudes by applying a scale factor with a range of values to the input signal.
  4. Compute the spectrum, output power, and ACPR of the output signal for each value of the scale factor. The bottom plot in the figure shows the output power in the main channel and the upper and lower ACPRs versus the available source power.

This technique of computing ACPR from a one-tone, swept-power, harmonic-balance simulation does not include any memory effects in the active devices or from the bias networks. Nor does it include any frequency-response variation of the circuit within the bandwidth of the main and adjacent channels. The circuit is characterized by a static, single-frequency, harmonic-balance simulation. To include these effects, one should use a standard Envelope simulation of the transistor-level amplifier. In doing so, results that previously required at least tens of minutes should now require only seconds. As a result, designers can experiment much more with their designs while improving performance.

There are two versions of the data display used to compute the ACPR and output power versus input power. One has all of the equations visible on the data display page while the other uses a custom AEL function that implements the same equations. The computation uses matrix math to determine the coefficients of a fifth-order polynomial that fits the Vout versus Vin and phase out versus Vin data of the simulated amplifier. To get the output amplitude and the output phase shift, the magnitude of the modulated input signal is then applied to both of these curve-fit polynomials.

The markers on the Vout-versus-Vin curve of the amplifier can be moved to determine the range of data to be used in the curve-fitting calculation. A range of scale factors can be specified to scale the magnitude of the modulated input signal. The custom AEL function lets the user know if the maximum scaled input-signal amplitude is beyond the maximum input-signal amplitude used in the one-tone simulation of the amplifier. If this occurs, the user may want to re-run the one-tone simulation with a higher maximum input-signal power.

The ACPR calculation does require that the user specify the frequency limits of the main channel and the upper and lower adjacent channels relative to the center frequency of the main channel. These limits differ for each communication standard. Users of Agilent's Advanced Design System may download an example file showing the technique discussed here from the Agilent EEsof EDA Knowledge Center (www.agilent.com/find/eesof-knowledgecenter).

Agilent Technologies, 1400 Fountaingrove Parkway, Santa Rosa, CA 95403; Internet: www.agilent.com.

TAGS: Content
Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.