08.02.2011
A collaboration of the working groups ”Acceleration” and ”Terahertz Optics” of PTB realized the first experimental determination of the combined electro-optic delay of PIN photo diodes with integrated preamplifier. The results from these measurements confirm the estimates of measurement uncertainty for the primary phase calibration of accelerometers especially for high frequencies.
PTB's primary national standards for the phase calibration of accelerometer sensitivity rely on the precise determination of the time delay of the transducer's electrical response to the mechanical input acceleration. Since the input acceleration is in turn determined by laser interferometry, any (unknown) delay caused within the laser interferometric measurement system will result in systematic deviations of the phase response calibration results. In order to achieve reliable information about the signal delay inherent to the typically applied Laser Doppler vibrometers, PTB set up a facility for the primary phase calibration of such devices in 2009 [1, 2]. This facility uses high speed PIN photo diodes as a reference in a modified Michelson interferometer. The major source of uncertainty in the facility was so far the still unknown delay caused by these PIN diodes and their attached pre-amplifiers.
In a joint approach of the working groups ”Acceleration” and ”Terahertz Optics” it was now possible to determine the opto-electronic conversion delay of the PIN diodes with an uncertainty of less than 80 ps [3].
The set-up (c.f. Fig. 1) makes use of a mode locked femtosecond laser (FSL) which runs with a repetition rate of 76 MHz providing pulses of 200 fs duration. The Laser beam of the FSL is split by a beam splitter in order to have one beam line providing the trigger and the second beam line providing the input for either a photoconductive switch (PCS) as the reference or the PIN diode under investigation (DUT).
Figure 1: Set-up for the determination of the phase delay by a substitution method. The photoconductive switch (upper part) is the reference and is subsequently replaced by the device under test (lower part).
As the inherent delay of the PCS was well known from elaborate investigations in the field of terahertz-optics [4] it could pose a high speed, high resolution timing reference in this setup. By comparing the pulse delay (relative to the trigger) of the PCS measurement with that of the measurement by the PIN-diodes (c.f. Fig. 2) a time delay of about 3.10(8) ns could finally be attributed to the DUTs.
Figure 2: Combined plot of the FSL pulse measured via the PCS and the FSL pulse measured with either of the photo diodes under test on the same time scale. The time values given in the picture denote the delay after triggering of the UFSO.
Considering the technical specification giving 1 ns as typical rise time this result might indicate that the rise time alone is insufficient to characterise the phase response. Considering the application of acceleration measurement, however, the delay of 3.1 ns converts to a phase delay of 0,022° at 20 kHz vibration frequency. This is still negligible considering current measurement uncertainties in the field, which are beyond 0,1°.
[1] Scientific news from Division 1: 
Measuring set-up for the primary phase calibration of laser Doppler vibrometers developed
[2] Blume, F.; Täubner, A.; Göbel, U.; Bruns, Th.: Primary Phase Calibration of Laser-Vibrometers with a Single Laser Source, Metrologia, 46 (2009), 489–495
[3] Bruns, Th.; Blume, F.; Baaske, K.; Bieler, M.: "Optoelectronic Phase Delay Measurement for a Modified Michelson Interferometer", IMEKO 2010: TC3, TC5 and TC22 Conferences, Pattaya, 21-25, November, 2010, Download:
http://www.imeko.org/publications/tc22-2010/IMEKO-TC22-2010-NP-008.pdf
[4] Bieler, M.; Spitzer, M.; Pierz, K.; Siegner, U.: Improved Optoelectronic Technique for the Time-Domain Characterization of Sampling Oscilloscopes, IEEE Trans. Instrum. Meas., 58 (2009), 1065–1071
Contact person:
Thomas Bruns, FB 1.7, AG 1.71, E-Mail: thomas.bruns@ptb.de