234 E. M. Jacobson et al. of how different DAQs attenuate data outside of the bandwidth of interest. This study also attempts to investigate how system slew rate is calculated and how it relates to other system specifications, like sample rate and anti-alias filter properties. 22.2 Background Below is a list of data acquisition system requirements for adequate shock data collection, from MIL-STD-810G change 1 Method 517 (Table 22.1). A sample rate of ten times the maximum frequency of interest is for amplitude accuracy when computing the absolute maximum shock response spectrum. Most accelerometers can accurately measure up to 10 kHz, so a sample rate of 100 kS/s would satisfy this sample rate requirement. Signal to noise ratio (SNR) is generally listed on the specification sheet and can also be verified experimentally. But how can the slew rate requirement be evaluated? Slew rate can be defined as the maximum change in voltage that a DAQ can detect without serious anomalies. These anomalies occur when a change in voltage is not detected and can show as clipping or a zero-offset in the collected pyroshock data [1]. It should be noted that other errors associated with pyroshock testing, such as accelerometer errors, also present themselves as clipping and zero-shift [3]. Slew rate specifications can be found for the system’s Analog to Digital Converter (ADC) chip, but the slew rate of interest is of the entire system, not the ADC chip. Slew rate can be defined as the change in the detected voltage (Vout) with respect to time [4]: Vout =V ∗sin(ωt) dVout dt =V ∗ω∗cos(ωt) Slew rate is generally presented in volts per microsecond (V/μs). The gain-bandwidth product (V*ω) is a constant and can become saturated at high frequencies and high voltages [4]. This is also known as slew rate saturation and is the value of interest in this research. The maximum slew rate capabilities of a system can be found in a few ways. Smith [5] uses a sine sweep across a large frequency range and identifies discontinuities or anomalies, calculating the slew rate from that frequency. This method is attempted in this research. Another method, mentioned in IEEE-1057 section 9.3 [6] is to record the DAQ response to a step input, increasing the amplitude from 10% full scale. When the system does not detect a change proportional to the step response increase, the slew rate has been saturated. The slew rate of the system is equivalent to the largest recorded rate of change. A similar method is attempted in this research. Although slew rate requirements are mentioned in [1], no test method is described. However, examples of contaminated data are presented in Annex A. Most modern DAQs implement a sigma-delta ADC, which internally oversamples at a high rate (~9 MS/s), then uses an analog AAF for that internal sample rate. The ADC uses a one-bit comparator to toggle between a change in the voltage (the least significant bit). The digitized data is then downsampled (including a digital AAF) with the user-defined sample rate and reconstructed over the entire dynamic range (Fig. 22.1). This type of ADC is cost-effective and allows for a wide frequency and amplitude range while still providing alias protection. Three protocols were selected (or designed) to test a DAQs ability to detect abrupt changes, characterize the AAF, and compare calculations (slew rate, alias-free bandwidth, dynamic range) to the supplied specifications and test requirements. Table 22.1 Minimum data acquisition system requirements for adequate shock data [1] Data Acquisition System Item Recommended Value Sample Rate >10*Fmax Signal to Noise Ratio (SNR) ≥60 dB SlewRate ≥0.5*Vmax /μs without distortion
RkJQdWJsaXNoZXIy MTMzNzEzMQ==