Accurate measurements demand more than just high-quality data-acquisition boards or instruments. When a project requires obtaining data from hundreds of sensors, precise timing plays a key role in gathering reliable information.
Most engineers have used a common type of trigger arrangementóthe star trigger. In this configuration, a trigger signal connects from its source to each acquisition instrument so that all acquisitions start simultaneously. Tests of a programmable controller, for example, might involve using a logic analyzer to monitor digital signals and an oscilloscope to monitor power lines. A logic-level transition in the controller provides the trigger signal that goes to both instruments. This arrangement works well for simple measurements.
An instrumentation system also might employ one device, say a logic analyzer, to trigger other measuring instruments. Thus, when a specific digital pattern appears, the logic analyzer starts to acquire data and relays a trigger signal to other instruments.
This arrangement also works well, but it can introduce timing errors as sample rates increase. Sampling at 100 Msamples/sec leaves 10 nsec between samples, so a delay of only a few nanoseconds in a trigger signal can offset measurements significantly. Even using short cables for trigger connections may not overcome problems with delayed triggers.
A trigger also may reach instruments at different times, thus adding skew to the measurements. In addition, because the instruments use their own clocks, some clock-drift will occur to add more uncertainty to measurement times. Over a long acquisition period, slight differences in clock frequencies accumulate so measurements no longer align properly on the time axis.
Many instruments furnish a sample-clock output that other instruments can substitute for their internal sample-clock signal. Distributing a "standard" clock signal helps ensure all instruments take measurements at the proper times. Cables of the same length keep the distributed clock signals in phase. But some delay will occur between the clock source and "slave" instruments. As long as the delay remains short compared to the sampling period, this clocking technique works well. But, say the delay amounts to 10 nsec and the instruments sample at 200 Msamples/sec, or 5 nsec between samples. Clearly the delay overwhelms the sampling and causes problems.
To overcome the clock-delay problem, test equipment can connect to a precision central reference clock and derive internal clock signals and sample-acquisition frequencies from it. Some instruments may include a 10-MHz oven-controlled crystal oscillator (OCXO) that can serve as a central clock.
Or, engineers can buy rubidium clocks sold specifically for lab use. A precision rubidium clock will supply a 10-MHz output with a frequency stability of one part in 10-10 to one part in 10-12 .
That stable signal can connect to all data-acquisition instruments over equal length cables. Many instrument or data-acquisition systems include a phase-locked loop (PLL) that locks to the 10-MHz signal and then creates an accurate internal timebase. This configuration offers the benefits of a central clock, but without the need to overcome a delay in the clock signal.
But suppose data-acquisition systems must operate at considerable distances from a precision 10-MHz source. How do such systems stay in sync? The Global Positioning System (GPS), established by the U.S. Department of Defense, can offer help. Each GPS satellite includes several accurate clocks that transmit information to Earth stations. (That timing information provides the raw data that, when processed, leads to a location).
Lab-grade GPS receivers can acquire the timing information and use it to produce an accurate 1 pulse per second (1 PPS) signal. Signal jitter can affect the short-term stability of that signal, but over the long term, the 1 PPS signal tracks the stability of the rubidium clocks in each GPS satellite. Inserting the 1-PPS signal into a PLL that controls a local rubidium oscillator produces a clock signal accurate to within a few tens of nanoseconds. Keep in mind that most instrument systems need to know relative time; that is, they need to stay synchronized with a master clock. They don't need to know the absolute time; for example, universal coordinated time (UTC) 1326 hours, 15 seconds, and so on.
On Deck: A military version of a combined
GPS receiver and rubidium clock produces accurate 10 MHz, 1PPs, and IRIG-B
timing signals that ensure synchronization of distant equipment.
Data-acquisition systems in military and aerospace applications often operate with a timing resolution of 1 Ķsec. In such cases, standard Inter-Range Instrumentation Group (IRIG) signals can synchronize instruments. Like a master clock, an IRIG time-code generator produces coded signals that keep equipment synchronized. Test-system designers can buy IRIG generators and instruments that accept IRIG signals. The IRIG signals can provide data that recording systems can use to "time-stamp" data. This time information makes it easy to locate specific test data and correlate it with other test data. Currently, the IRIG-B standard, specified in IRIG Standard 200-98 finds the most widespread use. Some GPS receivers will provide an IRIG-B timing-signal output.