Michael S. Heutmaker

End-to-end test strategy for wireless systems

This paper proposes an end-to-end test strategy for wireless systems described by using wireless microcell as a vehicle. It presents a framework for resting digital and radio frequency (RF) boards in a wireless system and provides an integrated solution. The test strategy is based on two concepts: using boundary-scan for digital testing, and using boundary-scan to provide a gateway for DSP-based functional testing of RF circuits. The test signals are generated within the system and the test control and verification are also provided by the system leading to system self-test.

The error vector and power amplifier distortion

The RMS value of the error vector magnitude (EVM/sub RMS/) is a useful system-level figure of merit for the accuracy of digitally-modulated signals. At the circuit level, the nonlinearity of a power amplifier is often given by the amplitude distortion (AM-AM) and phase distortion (AM-PM). When the amplifier input is undistorted, there is a simple relationship between the amplifier distortion and the error vector at the output: the error vector is the difference of the measured signal and a reference (ideal) signal, while the AM-AM and AM-PM are found from the ratio of the measured and reference signals. This relationship enables the EVM time series to be predicted from the AM-AM and AM-PM. In the frequency domain the spectral regrowth at the amplifier output is equal to a portion of the error vector power spectrum. A frequency-domain distortion power ratio may be estimated from the appropriate time-domain average of the EVM. These relationships are validated experimentally for a class AB amplifier and /spl pi//4 DQPSK signals.

Timing jitter and pump-induced amplitude modulation in the colliding-pulse mode-locked (CPM) laser

The timing jitter and spurious amplitude modulation of colliding-pulse mode-locked (CPM) lasers were measured. The absolute jitter (the jitter of the laser alone) varied between 5 and 10 ps RMS in a 50-500-Hz bandwidth. The smallest measured relative jitter (timing fluctuations between the CPM and a radio-frequency (RF) synthesizer synchronized to the CPM) was 1.8 ps RMS in a 2-Hz to 1-kHz bandwidth. Separate from the jitter, spurious modulation in the CW pump laser mixes with the CPM pulse train to produce a set of discrete amplitude-modulated sidebands in the power spectrum of the CPM output. The frequencies of these sidebands change with cavity length, and the sidebands can be eliminated by operating the pump laser in a single longitudinal mode. >

Electro‐optic sampling of high‐speed silicon integrated circuits using a GaAs probe tip

Electro‐optic sampling is performed on a high‐speed silicon multiplexer integrated circuit using a gain‐switched semiconductor laser and an external probe tip fabricated from GaAs. An approximate electrostatic model is used to calculate the dependence of the electro‐optic modulation on the height of the probe tip above the circuit, the geometry of the circuit, and the dielectric constants of the probe tip material and the passivation layer (if any) on the circuit. The measured variation of the electro‐optic modulation with probe tip height is in good agreement with the prediction of the model.

Amplitude noise reduction of 50 dB in colliding-pulse mode-locking dye lasers

Spurious sidebands observed in the noise spectrum of colliding-pulse mode-locking dye lasers result from coupling of the longitudinal mode beating modulations of the (argon-ion) pump laser with the dye laser. We achieve a 50-dB amplitude noise reduction at the mixing frequencies by inserting a single-frequency étalon in the argon-laser cavity. Furthermore, we find the amplitude noise of the colliding-pulse mode-locking laser to be entirely dominated by the (plasma discharge) noise of the argon laser.

Measurement of multigigahertz waveforms and propagation delays in modulation-doped field-effect transistors using phase-space absorption quenching

High‐speed waveforms (up to 20 GHz) in InGaAs/InAlAs modulation‐doped field‐effect transistors (FETs) are measured using 10–20 ps optical probe pulses via the quenching of the absorption in the quantum well gate channel due to Pauli exclusion. The technique is a noncontact probe of the charge density in the gate, and hence, of the logic state of the FET. This charge‐sensitive probing technique is combined with voltage‐sensitive electro‐optic sampling to study internal dynamics of the FET. A gate channel charging time of 11 ps and a gate to drain propagation delay of 15 ps are measured.

Using Digitally-Modulated Signals to Measure the Gain Compression and Phase Distortion of a Radio Frequency Amplifier

Radio frequency amplifier linearity test data obtained from a network analyzer cw power sweep does not always match the amplifiers response to a digitally modulated carrier (which is the signal present in a communications system). This paper describes a DSP-based RF test technique which extracts the gain compression and phase distortion characteristics of an amplifier from measurements performed with a digitally-modulated test stimulus. The technique utilizes the magnitude variations inherent in phase-shift keyed (PSK) modulation to sweep the input power to an amplifier, instead of using a power sweep from a network analyzer. By downconverting the modulated signal and digitizing in the time domain, the magnitude and phase of the modulation may be measured at the input and output of the amplifier. Using the modulation instead of a cw power sweep to drive the amplifier into compression might produce better correlation between test results and actual operation, at least for PSK modulation schemes. We have demonstrated this new technique experimentally on a commercial power amplifier, where we compare network analyzer power sweep data to DSP-based results using BPSK signals at two different bit rates. For this amplifier at 1 kbps the DSP-based measurement of saturation agrees with the network analyzer measurement and demonstrates the validity of the new technique. At 100 kbps, the amplifier saturation behavior during BPSK operation differs from that measured on the network analyzer in two repects: the nonlinearity of the amplifier decreases at the higher bit rate, and the saturation shows significant time-domain hysteresis.

Measurement of stray capacitance due to solder flux residue on radio frequency circuit boards

A microstrip directional coupler is used as a test circuit to measure the stray capacitance due to solder flux residue at frequencies up to 4.8 GHz. Comparison of experimental data with simulation results and coupled line theory provides an estimate of 25/spl plusmn/15 femtofarads for the stray capacitance from one flux residue. This measurement technique may be used to experimentally compare the RF effects of different solder fluxes. When combined with information about conductor spacings, residue properties, and the amount of residue, this measurement could validate estimates of the stray capacitance due to flux residue in a variety of circuit geometries.

Electrooptic probing of MMIC devices with a semiconductor laser using a novel method for phase referencing

Measurements were made of both the phase and amplitude frequency response of GaAs MMICs (monolithic microwave integrated circuits) over a range of 2 to 18 GHz using electrooptic sampling with a gain-switched semiconductor laser. A novel phase-referencing technique allows accurate frequency scans of the phase response by eliminating problems associated with phase changes between the synthesizers driving the circuit and the laser. Using coated optics and differential detection, a voltage resolution of 0.5 mV/ square root Hz, within 4 dB of the shot-noise limit, has been achieved.<<ETX>>

Measurement of phase and amplitude response of a GaAs MMIC by electrooptic sampling

A gain-switched injection laser was used to measure the amplitude and phase response of GaAs MMICs (monolithic microwave integrated circuits) and transmission lines by electrooptic sampling, up to 18 GHz. A phase referencing method that enables the phase response to be measured accurately has been devised, and the voltage gain and phase shift of a MMIC low-noise amplifier were measured over the 5-15 GHz range.<<ETX>>