SPECIAL SECTION FEBRUARY 2001

DA Table of Contents

Synchronizing Computer-Based
Measurement Systems

Time-interleaving high-speed A/D converters can dramatically multiply the sampling speed of a data acquisition system, but it becomes a tricky and complex chore at high sampling speeds.

Kosta Ilic and Lokesh Duraiappah,
National Instruments

Photo 1. Synchronization of PXI-based motion control and data acquisition enables an integrated platform for robotic applications.

n a synchronized computer-based measurement system, components are connected and programmed to take measurements at precisely the same time. In this context, "component" typically refers to a measurement device (e.g., a digitizer), but it can also refer to a functional portion of a device, such as the general-purpose counter/timer on a multifunction data acquisition (DA) board. With tight integration of these components and effective synchronization, you can customize solutions that would be prohibitively complex and costly to create with traditional stand-alone instrumentation (see Photo 1).

Building Blocks for Synchronization
When creating synchronized measurement systems, you typically designate one device as a master and one or more as slaves. The master device generates a signal or signals used to control all the measurement devices in the system. The slave devices receive control signals from the master device.

To synchronize a system, you have to understand the basics of timing and triggering signals (see the sidebar "Clocks and Triggers"). Latencies and timing uncertainties arising during the orchestration of multiple measurement components are significant challenges in synchronization, especially for high-speed systems. These issues, often overlooked during the initial system design, limit overall speed and accuracy.

The main objective of synchronization is either to correlate measurements or to precisely control process execution. In most cases, you're interested in correlation in terms of time (occasionally you may think of it in different terms, such as position). When performing this type of synchronization, you should take into account the system's sampling rate to determine how tightly to correlate the sample clock of your measurement system.

For example, if your sample clock is 1 Msps (the equivalent of 1 µs, or 1000 ns), you need not synchronize measurements to within nanoseconds. The objective is to create a system of devices synchronized to submicrosecond accuracy. Precise timing of measurements is a prerequisite of the measurement system.

Thus, the synchronization of measurements across multiple devices implies that you must synchronize all sample clocks to within the uncertainty of the period of the sampling clocks. To this end, you must also examine the distribution of triggers.

To implement synchronization, you'll usually follow one of three schemes. These include synchronization at the trigger level, at the trigger and sample clock level, or at the trigger and reference clock level.

Scheme 1: Synchronization at the Trigger Level
The first synchronization scheme is based on sharing trigger signals among measurement devices. The master device distributes these signals to all slave devices to start operations (see Figure 1A). This scheme is appropriate when you're synchronizing short- duration measurements or, in some cases, dissimilar measurement types.

Figures 1. The fundamental schemes for synchronization are based on sharing different signals among measurement devices. Synchronization at the trigger level (A), synchronization at the trigger and sample clock level (B), and synchronization at the trigger and reference clock level (C) illustrate the general schemes between master and slave measurement devices.

When using this scheme, you should be aware of several issues, including delay, skew, and drift. Delay refers to the interval between triggering the master and triggering the slaves. The trigger signal should arrive at each slave device with as little delay as possible. To reduce delay, minimize the signal path length from the master to the slave devices. Measurements that require low sampling rates can tolerate a degree of slack in delay specifications, but at high sampling rates, delay can affect measurement integrity.

Skew is the degree of nonsynchronism in the triggering of the various slave devices. Significant skew causes loss of total system synchronization. To minimize the problem, make sure the signal cables from the master to the slaves are of equal length.

Drift of clocks used to time different devices for long-duration measurements are likely to cause loss of synchronization. When clocks operate independently, small errors accumulate and, over time, may exceed your sampling time. You can minimize drift by using measurement devices with high-quality clocks, or you can eliminate this effect by using one of the following two schemes.

Scheme 2: Synchronization at the Trigger and Sample Clock Level
The master device can control operation of the measurement system by exporting both trigger signals and sample clock signals to the slave devices. For example, a system consisting of multiple digitizers and analog output sources has a common sample clock from an appointed master. The master sample clock controls A/D and D/A conversion timing on all devices (see Figure 1B).

The delay and skew issues associated with the first scheme apply here as well-in this case, for both trigger and sample clock signals. The major advantage of this approach over the previous scheme is that a common sample clock controls all devices. With a common sample clock, all waveforms are sampled at the same time, and drift is eliminated. This resolves the central issue of synchronized measurements.

This scheme is appropriate in systems where delay of the sample clock is small æompared with the sampling rate. For example, if you have established that propagation delay is 10 ns and sampling rate is 1 Msps (1000 ns), you can use this scheme. If the sampling rate is 500 Msps (2 ns), the approach is inappropriate.

Scheme 3: Synchronization at the Trigger and Reference Clock Level
You can synchronize your system by sharing a trigger and reference clock between multiple measurement devices. In this scheme, the master measurement device or a dedicated high-precision clock can supply the reference clock. This approach is particularly valuable for high-speed operations, where the sample clock is derived from the reference clock by multiplying the reference clock to obtain high rates.

For example, you can derive a 100 Msps sample clock by multiplying a 10 MHz reference clock tenfold. Scheme 2 requires direct feed of the sample clock to each device. Line integrity and propagation delays make this difficult to accomplish for high-speed sampling clocks across cables or trigger buses. A better approach is for measurement devices to share a common reference clock for generation of all sample clocks.

The method usually used with Scheme 3 to synchronize and generate sampling clocks is phase-lock looping, or PLL (see Figure 2).

Figure 2. Reference clocks can be used to synchronize and generate high-speed clocks via phase-lock looping (PLL) to deliver high-speed synchronized measurements needed in many applications.

The PLL method basically monitors the phase of the reference clock and produces a high-speed sampling clock phase locked to the reference clock (see Figure 1C).

To achieve high measurement timing accuracy, consider using frequency sources based on rubidium or an oven-controlled crystal oscillator (OCXO). The accuracy of these devices can be better than ±100 ppb. For example, an OCXO frequency source with ±100 ppb accuracy yields a 10 MHz clock with ±1 Hz uncertainty.

Another important property of your reference clock is multiple output capability for multiple instrument synchronization. The reference clock from either the master instrument or a precision frequency source should be capable of being driven to multiple destinations without any loss of signal integrity.

The delay and skew issues described in Scheme 1 are applicable here as well and are particularly important for high-speed digitization, given that delay and skew can in this case easily approach and exceed sampling rate, invalidating synchronization of the measurement.

Connecting Signals
Measurement devices usually come with one of three options for connecting synchronization signals: user-supplied cabling, proprietary vendor-defined cabling, and connections integrated with the measurement platform.

User-Supplied Cabling. User-supplied cabling is available for both computer-based and stand-alone measurement devices. For example, you can often externally synchronize your function generator or digital storage oscilloscope with a reference frequency source.

When synchronizing your instrumentation, you must ensure that the cables from the frequency source to the other components of your measurement system are the same length to avoid skew. The same criterion applies in the distribution of your trigger signal from the master to the slave devices. As noted above, your frequency source should be able to distribute a common reference clock to multiple destinations. User-supplied cabling is the only synchronization option for traditional stand-alone instruments.

Proprietary Vendor-Defined Cabling. Some vendors of computer-based measurement devices (e.g., DA boards) address synchronization by providing a proprietary bus, which can be external or internal to the computer. Sampling clocks, reference clocks, and triggers are distributed through this bus. Dedicated high-speed digital buses are designed to facilitate system integration.

The physical bus interface is a multipin connector on the card, and signals are shared through a cable. You can serially chain two, three, four, or five boards together, achieving synchronization of several I/O channels. Another attractive feature of these trigger buses is built-in switching, so you can route signals to and from the bus on the fly through software programming. This eases the burden of manually configuring timing and triggering signal distribution on your boards.

Connections Integrated with the Measurement Platform. Some of the computer-based measurement devices are implemented in industrial form factors, such as VME/VXI and CompactPCI/PXI. VME/ VXI and CompactPCI/PXI address test and measurement, telecommunications, defense, industrial research, and other markets. VXI and PXI extended VME and CompactPCI by adding timing and triggering buses to the form factors. This greatly simplifies synchronization of multiple devices.

VXI and PXI are open standards, and many companies make products for both of them. VXI is traditionally used in large test and measurement applications. Though relatively new to the market, PXI is gaining acceptance because of its smaller footprint, portability, high throughput, and lower costs, made possible through use of standard commercial technologies spawned by the large PC industry (see the sidebar "PXI-Platform for Synchronized Measurements").

Applications
Synchronization plays a particularly important role in high channel count systems and systems that join different forms of input or output, among other applications. Consider a few examples.

High Channel Count Systems. By integrating computer-based measurement modules, you can create compact high channel count systems. Coming up with equivalent systems with traditional instruments would be complicated and require substantial space. Some examples of high channel count systems include 12-channel 100 MHz digitizers, 6-channel 40 MHz arbitrary waveform generators, and systems with hundreds of lines of digital I/O. Synchronization is essential in these types of systems for simultaneous snapshots of data from different channels or lines.

A/D Converter Evaluation. You can create a computer-based measurement system for A/D converter (ADC) evaluation with two arbitrary waveform generators and one digital input module. The waveform generators provide analog signal and clock for the ADC, and the digital input module monitors the ADC's outputs.

If you synchronize generation of analog waveforms and acquisition of digital data, you can compare digital response with analog stimulus. By taking advantage of a PC's number-crunching power, you can perform advanced analysis and presentation of the results. Using similar techniques to test other classes of circuits with analog and digital signals, you can often reuse test systems created during development in later production tests and field service.

Analog Electronic Component Evaluation. Analog electronic component evaluation and manufacturing tests often involve measurements of both transient and steady-state parameters. With the computer-based measurement approach, you can use an arbitrary waveform generator to stimulate the component under test and a high-speed digitizer to acquire a transient response. A high-resolution digitizer acquires the steady-state response, and a computer-based counter/timer can supply a time delay function. If you program the synchronized system so digitizers start acquisition with the appropriate delay relative to the start of analog stimulus, you can obtain high-quality measurements and optimize test time.

Manufacturing Round Objects. When manufacturing or testing round objects, you often need to control part of the process by monitoring the object as it rotates. For example, if you manufacture high-quality metal disks, your test system would consist of a single rotating disk, a sensor that can detect defects, and a small, electronically controlled paint sprayer to mark areas with defects (see Figure 3). After the test system marks the defects, you can take corrective action.

Figure 3. Detection of defects on circular objects can easily be accomplished by synchronizing analog input measurements with analog output control using an optical encoder to pace the operation.

In the test system, the sensor and the paint sprayer are mounted to slide across the radius of the disk, and the paint sprayer is offset with respect to the sensor so that it hovers over the position where the sensor was four rotations earlier. You can incorporate a multifunction DA card and use an optical encoder to time the acquisition of the data from the sensor, set up the computer to analyze data, and generate a waveform to control the paint sprayer.

This is an example of a computer-based measurement system in which position is the key element for pacing different synchronized operations. Note that time is not essential in this case. The same approach can be used in similar situations, for disks and cylinders and for systems in which optical encoders report linear movement. In this example, synchronization makes it possible to tie control of the paint sprayer to the output of the sensor, with the appropriate position-based delay.

CAN Automotive System Validation. The Controller Area Network (CAN) standard defines a bus for robust communication between subsystems in modern automobiles (see Sensors October 2000, "Getting Control Through CAN"). These subsystems include the transmission, engine control, antilock brake systems, lights, dashboards, power windows, audio/video control, and power steering. In-vehicle networks contain sensors and actuators. To test these networks, you need to collect data synchronized to the operation of the CAN-based subsystems.

For example, to test a transmission subsystem, which contains temperature and fluid pressure nodes, you can monitor the network with a CAN board and independently monitor temperature and pressure with a DA board connected to the sensors with appropriate signal conditioning (see Figure 4).

Figure 4. The CAN bus data are synchronized to analog input measurements performed by data acquisition boards.

Using the CAN board to pace the system and programming the DA board to acquire data whenever the CAN board receives a packet, you can validate data obtained through the CAN bus. Without tight synchronization of the DA board to the CAN bus, you would acquire data from the two measurement systems with an unknown and varied time delay, which would make validation of the system more difficult.

Conclusion
To synchronize a computer-based measurement system, timing and triggering details are critical keys to success. Precise synchronization requires proper distribution of clocks and triggers. The three main synchronization schemes and proper knowledge of the pros and cons of each and the capabilities of your measurement devices help you make the right decision in choosing your solution.

Authors' Note
For more information on synchronization, you can access the application notes AN 128, AN 129, and AN 162 from National Instruments.

Clocks and Triggers The names of signals used for synchronizing measurement devices are not standardized and may vary with device type and manufacturer. Here three terms are used in reference to high-speed digitizers: reference clock, sample clock, and triggers. The reference clock is a periodic signal used to pace the acquisition at a low level. A typical digitizer contains an onboard oscillator to produce a reference clock. In some measurement systems, the reference clock is distributed from a central place to multiple measurement devices. The sample clock is a signal that controls the timing of A/D conversions. It's most often a periodic signal derived from the reference clock. For example, when you use a digitizer with a 10 MHz reference clock to acquire data at 1 Msps, the reference clock signal is divided by 10 to produce the sample clock. Alternatively, in high-speed systems, the sample clock is obtained by multiplying the reference clock. Trigger signals control acquisition of data at the highest level. External events, or triggers, are the main methods for initiating an acquisition. Triggers come in different forms, including analog, digital, and software.

PXI-Platform for Synchronized Measurements Electrically, VXI and PXI extended VME and CompactPCI by adding local buses and synchronization features. For synchronized measurements, key elements built into PXI and VXI are the reference clock, trigger bus, and star trigger bus. The PXI features described below broadly apply to VXI as well. Reference Clock. The PXI backplane provides a built-in common reference clock for synchronization of multiple modules in a measurement or control system. Each peripheral slot features a 10 MHz TTL clock. Equal-length traces from the clock to each peripheral slot yield skews of <1 ns between slots. The accuracy of the 10 MHz clock is usually ±25 ppm (dependent on individual chassis vendors), making it a relatively reliable clock for synchronization applications that rely on phase lock looping (PLL) methods. If you need a more accurate reference clock, you can insert a PXI counter/timer device with an oven-controlled crystal oscillator (OCXO) clock source into the second slot of the chassis. The slot's OCXO 10 MHz clock can be driven onto the PXI backplane clock lines in lieu of the PXI backplane clock. Then, the whole PXI chassis can inherit the OCXO clock stability. Trigger Bus. The PXI eight-line trigger bus provides intermodule synchronization and communications. Trigger or clock transmission can use the trigger bus lines. Triggers can be passed from one module to any number of modules, so you can distribute digital trigger signals from master to slave measurement devices. With variable frequency sampling clock transmission, multiple modules can share a time base that is not a derivative of the 10 MHz reference clock. For example, four data acquisition (DA) modules using a 44.1 Ksps CD audio sampling rate can share a clock that is a multiple of the 44.1 KHz or the direct 44.1 Ksps clock. For high-speed synchronization, the propagation delay and skew between slots can reach up to a maximum of 10 ns on a single PXI backplane. Star Trigger for Ultra-High-Speed Synchronization. The star trigger bus has an independent trigger line for each slot oriented in a star configuration from a special star trigger slot (defined as slot 2 in any PXI chassis). The trigger can provide an independent dedicated line for each of up to 13 peripheral slots on a single PXI backplane. The PXI star line lengths are matched in propagation delay to within 1 ns from the star trigger slot. This feature addresses high-speed synchronization where you can distribute start/stop trigger signals from the master measurement module in the star trigger slot with low delay and skew. For more information on PXI go to the PXI System Alliance and National Instruments.

DA Table of Contents

Home | Contact Us | Advertise © 2009 Questex Media Group, Inc.. All rights reserved. Reproduction in whole or in part is prohibited. Please send any technical comments or questions to our webmaster.