From 2 PM Friday, Jan 20 - 10:00 PM CST Monday, Jan 23, will be undergoing system upgrades that may result in temporary service interruption. We appreciate your patience as we improve our online experience.

Transient Recording Using PC-Based Instrumentation

Publish Date: Mar 22, 2016 | 39 Ratings | 4.56 out of 5 | Print


Transient signals have irregular properties, and for this reason, transient recorders have traditionally been a special breed of instrumentation, supporting many special requirements including high accuracy and large data storage. Transient recorders, just as any device that is designed for a specific niche purpose, are typically expensive and are not expandable as channel count and analysis needs change. As PCs and PC-based instrumentation have evolved in all-around performance, they cover many, if not all, of the formerly specialized requirements of transient recorders. Today's general-purpose PC-based data acquisition systems meet and/or exceed the capabilities of specialized transient recorders.

Table of Contents

  1. Classification of Transient Signals
  2. Anatomy of a Transient Signal
  3. Important Specifications for Transient Recorder Hardware
  4. Transient Signal Analysis Techniques
  5. PC-based Transient Recorder Hardware
  6. PC-based Transient Recorder Analysis and Presentation

1. Classification of Transient Signals

Analog signals are divided into two broad classes, they are:

  • Random
  • Deterministic

Figure 1: Waveform Types

A random signal is one which never repeats and has a flat frequency structure. It cannot be characterized by a simple, well-defined mathematical equation and their future values cannot be predicted. Deterministic signals on the other hand can be characterized by a mathematical equation and their future values can be predicted.

Random signals can be further classified into stationary and non-stationary signals. Stationary random signal refers to stable statistical properties of the mean value and auto spectrum over time. Non-stationary random signals are those whose statistical properties change significantly over the observation window.

Deterministic signals are further classified into periodic and non-periodic signals. Periodic signals, in the simplest form are sinusoidal, that is, consisting of a single sine wave (with a single frequency component). Complex periodic signals are those signals that repeat over and over again, thus resulting in a frequency spectrum with multiple sinusoidal components.

Back to Top

2. Anatomy of a Transient Signal

Figure 2: Transient signal created by hammer excitation

A transient signal is an analog signal which lasts for a very short duration of time and normally occurs only once or very infrequently. Transient signals fall under the Deterministic-Non-Periodic waveform type. Theoretically their value is zero before and after the signal.

Figure 3: Transient signal created by impact

Signal shown in Figure-3 is a transient signal created by impact, e.g. slamming a door. The signal is continuous with the maximum amplitude at 0 Hz and decreasing with increasing frequency. Time T is the duration of the transient signal which depends on strength of the excitation which in turn depends on the mass and firmness of the impacting structures.

Figure4: Frequency range of a transient signal

The frequency spectrum has a periodic structure with zero force at n/T intervals where n is an integer and T is the duration of the transient signal. The frequency range where the transient signal can be analyzed accurately is from 0 Hz to a frequency f after which the signal magnitude decays by 10 to 20 dB.

Figure 5: Rise time of a transient signal

The bandwidth and sampling rate required to capture a transient signal depends on the fastest phenomenon of the rise time. A transient signal in figure 5 has two rise times 0.1 micro sec and 0.5 micro. If the rise time of a transient signal is known then the bandwidth required to capture the signal accurately can be calculated using the following equation

Bandwidth = k/fastest phenomenon of the rise time

where k is a constant between 0.35 to 0.45. If the bandwidth of the PC based transient recorder is < 1GHz then the value of constant k is typically 0.35. If the bandwidth of the transient recorder is >1GHz then k varies between 0.4 to 0.45.

In our case k = 0.35.

Bandwidth = 0.35/0.1 micro sec
= 3.5 MHz

It is recommended that the bandwidth of the PC based transient recorder you use to capture the transient signal be 3 to 5 times the maximum frequency component of the transient signal for minimal amplitude error. The minimum sampling rate required to reconstruct this transient signal completely is twice the maximum frequency component of the signal.

Transient Signals Around You

50 Hz – 20KHz Meteorology, Audio
20KHz – 30MHz SONAR, Military, Aerospace, Defense,
30MHz – 100MHz RADAR, LIDAR, Research
100 MHz and more Research, Communications

There are a wide range of industries in which transient recorders are used. Some applications are:

In meteorology studies, for instance, lightning strike detection is important. Here the randomness of the lightning strike makes it imperative that the measurement system is capable of monitoring the thunderstorm for all lightning strikes that occur. The storm may span minutes or hours and the lightning strikes can occur seconds or minutes apart. A transient recorder with time stamping capability can be used for this application.

Other applications are Wireless Signal Recording, Ultrasound, Imaging, RADAR, LIDAR, Shock Wave Testing, Disk Drive Testing, RF Receivers, CCD Testing, Crash Testing and Material Testing.

Back to Top

3. Important Specifications for Transient Recorder Hardware

Sampling Rate: The minimum sampling rate required to determine the frequency content of a signal is twice the maximum frequency component of the signal. For an example if the maximum frequency component of the signal is 100 MHz a sampling rate of at least 200 MSample/second is required to accurately determine the frequency content. However, for most transient applications, the shape of the incoming signal is as important as the frequency content.

To reconstruct the shape of the signal the sampling rate should be 5 to 10 times the maximum frequency component of the signal. For an example of the maximum frequency component of the signal is 100 MHz a sampling rate of atleast 500 MS/s to 1000MS/s is required to replicate the shape of the signal. High sampling rate is directly proportional to how identical the replicated signal is to the original signal in the time domain.

Simultaneous Sampling: In transient recording applications that require more than one channel, it is important to preserve phase relationships. Phase relationships are preserved by simultaneously sampling all of the input channels. There are several architectures that can be used to accomplish this simultaneous sampling, but a multi-ADC architecture has many advantages .

Bandwidth: Bandwidth describes the frequency range in which the input signal can pass through the analog front end with minimal amplitude loss - from the tip of the probe or test fixture to the input of the ADC. Bandwidth is specified as the frequency at which a sinusoidal input signal is attenuated to 70.7% of its original amplitude, also known as the -3 dB point. The following figure shows the typical input response for a 100 MHz bandwidth circuit.

Figure 6: Input Response for a 100MHz bandwidth circuit

For example, if you input a 1 V, 100 MHz sine wave into a circuit with a bandwidth of 100 MHz, the signal will be attenuated by the circuit’s analog input path and the sampled waveform will have amplitude of approximately 0.7 V.

Figure 7: Sine wave sampled by a 100MHz bandwidth circuit

It is recommended that the bandwidth of your PC based transient recorder be 3 to 5 times the maximum frequency component of interest in the measured signal to capture the signal with minimal amplitude error (bandwidth required = (3 to 5)*frequency of interest). The theoretical amplitude error of a measured signal can be calculated from the ratio of the circuit’s bandwidth in relation to the input signal frequency (R).

For example, the error in amplitude when measuring a 50 MHz sinusoidal signal with a 100 MHz bandwidth high-speed digitizer, which yields a ratio of R=2, is approximately 10.5%.

Data Storage: Depending on how much data is collected per second and how long the data must be recorded, the data storage mechanism and choice of bus may vary dramatically. First, the user must consider the data path from the converter to the final storage place and determine where any bottlenecks exist. Depending on where, and if, a bottleneck exists, the data storage mechanism and bus may be chosen. Generally, input samples are broken up into two bytes. For instance, 12- and 16-bit samples represent 2 bytes (B) of memory; whereas, 18- and 24-bit samples represent 4 bytes of memory.

Table 1: Comparison of data storage methods

Data Storage Mechanism
Storage Rate*
>1 GB/s
512 MB
Conduant StreamStor
100 MB/s
1,000,000 MB
3,000 MB
Hard drive
1,000,000 MB

* - intended as guidelines

Table 2: Common bus technologies (from peripheral device to PC memory)

PCI Express (x1 slot) 200-250 Mbytes/sec
PCI/PXI 132 Mbytes/sec
IEEE 1394 (Firewire) 50 Mbytes/sec
PCMCIA 20 Mbytes/sec
Ethernet 100BaseTX 11 Mbytes/sec
GPIB 8 Mbytes/sec

*Subtract 10-20% for bus overhead

Using PXI/PCI as an example, most transient applications are within the bandwidth of the bus and the collected data is small enough to fit into PC memory (1 MB/s-50 MB/s and 3 GB of total data). These applications are very inexpensive because they take advantage of common PC technologies. However, if the dataset becomes too large or the rate becomes higher than the PCI/PXI bus can accept, other options must be considered. For instances where the data rate is higher than the PXI/PCI bus limit, either another bus must be considered or the data acquisition device must have sufficient on-board memory. For instances where the dataset becomes too large for PC memory, another data storage mechanism must be considered such as hard drive or Conduant StreamStor. The choice between Conduant StreamStor and hard drive is typically determined by whether the data storage mechanism can handle the data rate.

There are many other bus/storage mechanism combinations in which variables such as throughput, price, and storage capability must be weighed against each other to find the appropriate combination.

Resolution: The number of bits that the ADC uses to represent the analog signal is referred to as the resolution. Typically, the higher the resolution (number of bits), the more accurate is the measurement will be. An 8-bit ADC divides the vertical range of the input amplifier into 256 (2^8) discrete levels. With a vertical range of 10 V, the 8-bit ADC cannot resolve voltage differences smaller than 39 mV (10/256). In comparison, a 12-bit ADC with 4,096 discrete levels can resolve voltage differences as small as 2.4 mV.

Resolution is also defined as the smallest amount of input signal change that a PC based transient recorder or instrument can detect. This is important in recording transients such as a step input where the amplitude of the step and the subsequent ripple need adequate vertical resolution. If the resolution is not high enough, either the step size or ripple shape information will be lost during conversion from analog to digital.

Figure 8: Digitize sine wave

Figure 8 shows a sine wave and its corresponding digital image as obtained by an ideal 3-bit ADC. A 3-bit converter divides the analog range into 2^3, or 8 divisions. Each division is represented by a binary code between 000 and 111. Clearly, the digital representation is not a good representation of the original analog signal because information has been lost in the conversion. By increasing the resolution to 16 bits, however, the number of codes from the ADC increases from 8 to 65,536, and you can therefore obtain an extremely accurate digital representation of the analog signal if the rest of the analog input circuitry is properly designed. Hence based on the application it must be decided whether a very small signal should be detected in which case the PC based transient recorder should have a high resolution, capable of detecting and capturing all the information of the signal.

There are many other specifications that can effect the accuracy of the signal input circuitry such as spurious-free dynamic range (SPDR), total harmonic distortion (THD), and signal-to-noise ratio (SNR). These are especially useful when comparing devices with similar resolution.

Triggering: For most applications once a transient signal is recorded it would be followed by a set of actions. By setting up a trigger you can start acquiring the signal once the trigger condition is satisfied. In most applications it is necessary to capture the transient signal before and/or after a trigger occurs to analyze the behavior of the signal. In such instances you can use the pre-trigger or post-trigger feature to specify the number samples of the transient signal that need to be recorded. Some of the triggering modes available are: Edge, Hysteresis, Window, Video, Digital, Immediate, and Software.

Analog Triggering Modes

Analog Trigger Circuit: The analog trigger operates by comparing the analog input to an onboard threshold voltage. This threshold voltage is the trigger value, and can be set to any voltage within the input range. A hysteresis value associated with the trigger is used to create a trigger window the signal must pass through before the trigger is accepted.

If, for example, the trigger level is set to 2.0 V with a positive slope, a hysteresis is often added to reduce false triggering due to noise. In this example, if a hysteresis is set from 1.9 to 2.0 V, then the trigger will only activate once when passing 2.0 V with a positive slope. Even if the voltage drops to 1.95 V, and then rises above 2.0 V again, triggering will not occur again. Hysteresis requires that the voltage first drop below 1.9 V, and then again cross 2.0 V before triggering will be activated. This means that the trigger circuitry can tolerate a maximum noise voltage of just below 0.1 V peak-to-peak without any false triggers. Triggers can be generated on a rising-edge or falling-edge condition as illustrated in the following figures.

Edge Triggering: An edge trigger occurs when a signal crosses a trigger threshold you specify. The slope can be specified as either positive (on the rising edge) or negative (on the falling edge) to the trigger.

Figure 9: Positive and Negative and Edge Trigger Modes

Window Triggering: A window trigger occurs when a signal either enters or leaves a window you specify.

Figure 10: Entering and Leaving Window

Analog Edge with Hysteresis: Hysteresis provides buffer to prevent noise from causing false trigger. Hysteresis (window size) = high value - low value

Figure 11: Edge with Hysteresis

Digital Triggering

Digital triggering is much simpler than analog triggering, and is useful for applications where another device initiates operation of the application and the transient recorder is required to commence sampling. A digital trigger occurs on either a rising edge or falling edge of a digital signal.

Time stamp: For most applications which record transient signal, time stamp data is critical. The time at which the signal is recorded and the lapse between signals recorded is required. For an example in meteorology studies, lightning strike detection is important. The storm may span minutes or hours and the lightning strikes can occur seconds or minutes apart and hence a transient recorder with time stamping capability is necessary for this application.

See how to do triggering and timestamping using NI PC-based digitizers.

Back to Top

4. Transient Signal Analysis Techniques

Figure 12: Acquiring, analyzing and presenting a transient signal

Figure 12 shows the sequence of steps in acquiring, analyzing and displaying the transient signal. The signal is captured, scaled, amplified and converted to a digital signal. The captured signal can be analyzed in the time and/or frequency domain.
In the time domain, the representation often plots the signal value (commonly a voltage or current that represents some other measurement, such as temperature or strain) as a function of time. The other signal representation method is the frequency-domain view of the signal. This representation is typically based on a variation of the Fourier Transform and commonly plots the power of a signal as a function of frequency. By examining the frequency-domain view of a signal, you derive information about your signal that might not be immediately apparent from an examination of the time-domain representation and vice versa.

The most common frequency analysis tools are Fast Fourier Transform (FFT), Constant Percentage Bandwidth (CPB), Order Tracking and Band Pass Overall Levels. Each of these techniques has their own advantages and can be chosen based on your application.

Although frequency-domain representations such as the power spectrum of a signal often show useful information, the representations don’t show how the frequency content of a signal evolves over time. To see how the frequency content evolves over time you can apply the Joint Time-Frequency Analysis (JTFA).

JTFA is a set of transforms that maps a one-dimensional time domain signal into a two-dimensional representation of energy versus time and frequency. There are a number of different transforms available for JTFA. Each transform type shows a different time-frequency representation. The Short Time Fourier Transform (STFT) is the simplest JTFA transform and the easiest to compute. However the STFT technique suffers from an inherent coupling between time resolution and frequency resolution (increasing the first decreases the second, and vice versa). This coupling can skew the measurements that you can derive from the transform, such as average instantaneous frequency.

Other JTFA methods and transforms can yield a more precise estimate of the energy in a given Frequency-Time domain. Some options include:

  • Gabor spectrogram
  • Wavelet transform
  • Wigner distribution
  • Cohen class transforms

Many of the techniques desribed above require a continuous waveform and a integer number of waveform periods to produce accurate results, i.e. minimize spectral leakage. To improve the effectiveness of the above techniques, windowing is often required. Windowing is a technique used to shape the time portion of your measurement data, to minimize edge effects that result in spectral leakage in the FFT spectrum. By using Window Functions correctly, the spectral resolution of your frequency-domain result will increase.

Back to Top

5. PC-based Transient Recorder Hardware

There are a number of advantages in using PC-based transient recorder in the place of a stand alone transient recorder. Some of them are:

Higher processor speed: A strategic advantage of the PC-based transient recorder over the stand alone transient recorder is its ability to leverage the high processor speed of PC’s. As PC technology continue to evolve with high processing speed, a transient recorder that fully integrates these technologies becomes even more powerful. This advanced processor capability may be applied to post-processing tasks to save time, or can be applied to the transient data in real-time.

Faster throughput: Any recorded signal needs to be sent to a computer for signal processing and analyzing. The communication between a computer and a stand alone transient recorder is not only cumbersome but also slows down data transfer. On the other hand, PC based transient recorders are either slotted inside the computer or slotted in a chassis with a fast data bus. There is nominal dependence on a connection cable or interface required but the transfer of data is much faster compared to a stand alone recorder. The throughput of a device under test can therefore be much faster, up to 16X.

More data storage options: PC-based technology offers several levels of storage options. On-board memory is always an option regarless of the bus used, however it is cost prohibitive. The PCI bus pushed the envelope to 132 MB/s and PCI Express will push it past 250 MB/s. Because of the ever increasing bus speeds, the data storage can be moved into the PC, greatly reducing the overall cost per MB.

Smaller footprint: Transient recorders are commonly used in industries and research centers where spacing is an important criteria. Most industrial practices require the use of several machinery and equipment and the need to provide adequate real estate to house all the requisite apparatus is a concern that has taken on epic proportions. PC based transient recorder not only have the advantage of having smaller footprints but also allow different test equipment to be connected to the same PC.

PC based transient recorders available at National Instruments are:

Based on the application and speed, memory, accuracy, trigger, bandwidth, timestamp you can choose the specific hardware. You can also choose the bus architecture PCI, PXI, PCMCIA, USB and ISA based on your specific needs. A selected few under each category are listed below for your reference. For more details please click on the respective product links.

Figure 16: Transient Recorders at National Instruments

The data acquisition models below have the features required for transient recording. Match your application with the specifications below.

Max. on-board memory (samples)
# of channels
Max. sample rate per channel (MS/s)
Input resolution (bits)
Bandwidth (Hz)
512 M
150 M
256 M
100 M
32 M
100 M
16 M / 66.3k
15 M
16 M / 66.3k
0.01 to 100
8 to 21
100 M
Dynamic Signal Acquisition
0.49 fs
0.49 fs
S Series
64 M
7.2 M
S Series
65 M
1 M
S Series
32 M
1.3 M
M Series
1.7 M
M Series
725 k

*not simultaneously sampled
Table 4: Transient Recorders at National Instruments:

Back to Top

6. PC-based Transient Recorder Analysis and Presentation

The recorded transient signal can be analyzed using LabVIEW. State of the art LabVIEW toolkits that can be used to analyze these short duration signals are:

  • Sound and Vibration Toolkit
  • Sound and Vibration Measurement Suite
  • Advanced Signal Processing Toolkit

The Sound and Vibration Toolkit extends LabVIEW with functions and indicators for handling transient analysis, fractional-octave analysis, swept sine analysis, sound level measurements, frequency analysis, audio measurements, frequency response measurement, and several sound and vibration displays.

The Transient Analysis VIs in the Sound and Vibration Toolkit offer two different techniques for extracting information about transient signals. You can use the STFT (Short Time Fourier Transform) for signals in which the frequency content changes relatively slowly with time and the SRS (shock response spectrum) for shock signals. You can use the SVT Shock Response Spectrum VI to evaluate the severity of a shock signal. The results generated by the SRS are typically displayed on an XY graph.

The Sound and Vibration Measurement Suite includes order analysis VIs, for rotating machinery analysis and monitoring. These VIs employ Gabor Order Tracking, a patented algorithm based on joint time-frequency analysis (JTFA), as well as conventional re-sampling for online processing capability. Order analysis is a tool for examining dynamic signals generated by mechanical systems that include rotating or reciprocating components. As with frequency-domain analysis, you can think of order analysis as a signal scalpel that can dissect sound, vibration, and other dynamic signals into components that relate to physical elements of mechanical systems. Unlike the power spectrum and other frequency-domain analysis standards, order analysis works even when the signal source undergoes rotational speed variations or frequency/Doppler shifts. 

The Advanced Signal Processing Toolkit has tools such as wavelets and joint time-frequency analysis (JTFA) for the analysis of fast transients. With this toolkit, you can experiment with and develop using modern analysis techniques that include wavelets, super-resolution (model-based) spectral analysis, and joint time-frequency analysis (JTFA). With the Digital Filter Design component of this toolkit, you can interactively design and characterize finite impulse response (FIR) and infinite impulse response filters.

Back to Top

Bookmark & Share


Rate this document

Answered Your Question?
Yes No