Menu Bar

Main Window

Trash Theme Home News Downloads Tutorials WebLinks Login to Electromaniacs

  Menu Bar
Main Menu
Live Chat
News Feeds
How To
Online tools
Electronics lessons
Download Zone
Contact Us
Forum Wrapper
Books store
About us
Members: 3887
News: 247
Web Links: 21
Visitors: 3549566
Your IP
You are connecting to this site from:
Home arrow Blog arrow Measuring RF parameters in GSM/EDGE power amplifiers
Measuring RF parameters in GSM/EDGE power amplifiers PDF Print E-mail
Written by Administrator   
Article Index
Measuring RF parameters in GSM/EDGE power amplifiers
Page 2

Measuring RF parameters in GSM/EDGE power amplifiers

GSM/EDGE RF power amplifiers used in mobile communications in North America, Europe and elsewhere must conform to a strict ETSI standard in order to obtain type approval by the end user, the service provider and the public in general. In this article, a method for conducting RF lab measurements is presented using commercially available test equipment and following common RF practices. Usually, this involves testing in four bands and two modes.

Global system mobile (GSM) or enhanced data rate for GSM evolution (EDGE) modulated signal delivered by the mobile station (MS) uses the random access channel (RACH) link to initially communicate with the base transceiver station (BTS). Once the BTS to MS instructions are decoded via the slow associated control channel (SACCH), the RF output power level in the MS is dynamically set. The purpose is to provide the minimum necessary output power[1], in such a way that the BTS receives signals from many MS at approximately the same received signal strength (RSS). The end effect is that overall parameters such as talk time and standby time are vastly improved, with minimum interference. Hence, it is paramount that the RF power amplifier (PA) be tested and evaluated at all required power levels.

High efficiency is accomplished by operating the amplifier in class D for GSM (non-linear mode) and class AB2 in EDGE (quasi-linear mode).


The RF input and RF output of the amplifier device mounted in a typical evaluation PC board are connectorized with SMA-type miniature coaxial connectors. These are part of the 50 Ω impedance system together with the microstrip traces on the board. Hence, it is important that all RF connections are low-loss 50 Ω coaxial type.

To minimize any source of error or uncertainty during the measurement, it is important that the RF coaxial contacts are clean. Use laboratory-grade alcohol to remove any contaminant. Due to “skin effect,” most if not all, of the RF current flows near the surface of the conductor.

A GSM/EDGE modulated signal from a suitable generator is applied to the RF amplifier evaluation board via the bench setup as shown in Figure 1. The RF output under a 50 Ω load is monitored via an RF power meter, spectrum analyzer and an optional vector signal analyzer.

The spectrum analyzer displays the transmitted spectrum. Due to the pulsed nature of the GSM signal, it has two main components: the modulation and switching (or transient) spectrum. These two must be measured separately and carefully measured following[3] and compared to[2] for a pass/fail criteria. While the vector signal analyzer gives the value of the modulation quality in error vector magnitude (EVM) and phase error.

The transmitted EDGE spectrum contains additional energy due to AM/AM and AM/PM conversion by virtue of the quasi-linear class AB2 amplification (spectral re-growth). The figure of merit is usually adjacent-channel power ratio (ACPR), given in [dBc] at an offset frequency that corresponds to the first RF adjacent and first RF alternate channels.

The RF PA is dc biased under typical operating conditions as given in the data sheet for the particular device under test (DUT). Typical values are: Vcc1 = Vcc2 = +3.5 Vdc; Vreg = +2.7 Vdc.

The minimum acceptable levels are given in references 2 and 3. Also, Table 1 provides a summary of the relevant portions of the ETSI standard and its application to GSM/EDGE RF PAs


Test setup

A detailed block diagram is shown in Figure 1.

The input section has:

  • RF digital signal generator;
  • bandpass filter;
  • isolator;
  • directional coupler; and
  • RF power sensor.

The output section has:

  • RF load attenuator;
  • directional coupler;
  • RF power sensor;
  • spectrum analyzer; and
  • vector signal analyzer.

It is important that the dual-channel RF power meter and sensors be zeroed and calibrated and set to the correct operating frequency before any measurements take place.

Then, a bench calibration is performed, to determine the input and output insertion losses, entered as offsets into the power meter. Hence, channel 1 (assigned to the input) measures exactly the input RF power and channel 2 (assigned to the output) measures the output power. The gain of the device in dB is determined immediately by taking the difference of these readings in dBm.

Ramp control signal

The Vramp control signal is generated by the arbitrary function generator. In order to maintain a bounded spectrum during RF power increase or decrease, a gradual change is required on Vramp. This is usually implemented with a ¼ period sin2(x) or sin3(x) function applied to the rising edge. In similar fashion a ¼ period cos2(x) or cos3(x) is applied to the falling edge. In general, the cubic function will provide a gentler rise and fall than the square function, however, the latter has a faster time response.

A proper ramping profile can be generated by considering the following:

  • The maximum amplitude (typically 2 V to 2.5 V).

  • The pulse duration is 577 µs for one active slot (i.e., slot 0) and 12.5% duty cycle.

  • The repetition rate is 4615 µs or approximately 217 Hz.

  • The rise and fall times of the pulse are such, that they conform with Annex B in[2].

  • In general, more than one ramping profile is required to satisfy[2] over RF power output, battery voltage and temperature.

As a starting point, rise and fall times of the ramp can be approximated to those in annex B[2], that is approximately 14 µs to 16 µs and same time frame rate of 4615 µs.

To communicate with the ArbFuncGen a GPIB/IEEE 488.2 controller card (available, for example, from National Instruments or equivalent) and GPIB control software (such as NI 488.2) installed in the PC is needed. Then install the waveform creation software (such as Tektronix AXW-100 for model AFG-310 function generator or an Agilent 34811A for model HP 33120A function generator). Interconnecting GPIB cables are required between the computer and function generator.

In developing the Vramp profiles, it is useful to have an oscilloscope connected to the output of the arbitrary function generator. Final verification can be made in the RF output power with a spectrum analyzer in the zero-span (time domain) mode.

A typical display of the frequency spectrum as obtained by the spectrum analyzer with a GSM/EDGE personality is shown in Figure 2 (GMSK at 824 MHz) and Figure 3 (EDGE at 1910 MHz).

Required test equipment includes:

  • RF digital signal generator GMSK and EDGE modulation capable in the bands of interest;

  • spectrum analyzer with GSM personality;

  • vector signal analyzer with GSM personality;

  • printer for the above units, if required;

  • RF power meter, dual channel with sensors calibrated to the band of interest;

  • arbitrary function generator, for Vramp generation;

  • oscilloscope, four channels with 10x voltage probes and 1x current probes;

  • directional couplers, 10 dB coupling (input) and 20 dB coupling (output);

  • bandpass filter and isolator for the bands of interest;

  • load attenuator 20 dB, 5 W;

  • dc power supplies, regulated 0 to 5 Vdc, 2 A output (two);

  • DVM for accurate voltage measurements (two); and interconnecting RF and dc cables.

Next >
Sponsored Links

  Privacy Policy