Abstrad--Scalar measurements at 1.3 pm of high-speed ( > 20 GHz -3-dB optical bandwidth) photodetector transfer function magnitudes by swept frequency and short pulse (<3-ps FWHM) response techniques are presented and shown to be in excellent agreement to beyond 30 GHz. Scalar deconvolution is used to obtain photodetector response from the pulse response measurements with effects of the measurement apparatus removed.

No abstract

No abstract

Analysis of signals in Hilbert space is used for narrowband communications signals in both linear and exponential modulation formats (i.e. QAM, QPSK, FSK etc). Less well known perhaps is the use of the analytic signal in broadband, non-linear modulation analysis. Inter-symbol interference (ISI), Deterministic Jitter, phase noise, and AM-PM conversion are examples of non-linear system impairments whose measurement and analysis are simplified using analytic signal techniques in Hilbert space. Synchronous sampling allows for direct demodulation and extraction of baseband Inphase and Quadrature channels. Its application to measurement of both random and deterministic jitter, phase noise, and AM-PM effects is discussed, as well as calibration methods and other accuracy enhancement techniques. Synchronous sampling is contrasted to conventional waveform recorder timebases: real-time sampling, random equivalent-time sampling, and triggered sweep. Synchronous Sampling uses a phase locked timebase with delay control and carrier extraction to provide a rotating reference frame in Hilbert space, from which the signals are analyzed in three dimensions: I and Q vs. Time or Modulation Frequency.

100 GHz bandwidth equivalent-time sampler modules have been built and tested.These modules include strobe generation circuitry, sampling diodes, blow-by compensation, and IF amplification. The modules have been characterized for risetime, swept frequency response, timing jitter, linearity, dynamic range, and input referred noise. High-speed pulse response measurements yield a deconvolved 10%-90% risetime of approximately 3 ps. This corresponds to an estimated bandwidth of 113 GHz. Long-duration (100 ns) pulse measurements showed clean responses as well. Swept frequency measurements corroborate the bandwidth estimate, showing little to no change in response out to 65 GHz (limit of measurement system). Timing jitter measurements yielded a value of approximately 175 fs. Linearity measurements showed uncorrected linearity to within a few percent and corrected linearity of approximately 0.1%. This was over a dynamic range of 2 Vpp. Noise measurements gave a typical input referred noise of 2.5 mVms, approximately 0.7 % of the dynamic range, very close to the calculated quantum limit.