Trigger Gating: Circulating Loop and Burst Data Analysis with the Agilent 86100A Infiniium DCA Wide Bandwidth Oscilloscope

Transcription

1 Copyright 2000 Agilent Technologies, Inc. Trigger Gating: Circulating and Burst Data Analysis with the Agilent 86100A Infiniium DCA Wide Bandwidth

2 Trigger Gating: Circulating and Burst Data Analysis with the Agilent 86100A Infiniium DCA Wide Bandwidth There are experiments incorporating wide-bandwidth oscilloscopes that require select portions of a waveform to be viewed while other portions of the waveform are intentionally ignored. Examples include circulating loops, or data that occurs in bursts. In the specific example of the circulating loop, signal propagation through an extremely long length of fiber, typically in excess of 1000 km, is simulated with multiple circulations through a shorter length of fiber. For example, a 9000 km trans-pacific fiber link can be simulated by routing a signal 18 times through a loop of fiber 500 km in length. When an oscilloscope is used to view such a signal, the instrument should only sample the elements that have propagated the correct number of circulations. Thus the oscilloscope needs to be synchronized with the signals used to control the loading and circulation within the loop. Wide-bandwidth oscilloscopes typically require a signal synchronous to the data as a timing reference. For each trigger edge/event, one and only one data point is sampled. To acquire a waveform that is composed of 1000 data points, 1000 trigger edge/events must be responded to by the oscilloscope. The Agilent Infiniium DCA has an external BNC connector port called the trigger gate. This port responds to TTL compatible signals. With the trigger gate feature enabled, a high voltage at the port will enable the oscilloscope to respond to trigger edges and thus acquire data. When the signal at the port is low, the oscilloscope will not respond to triggers even if they are presented to the instrument. They are simply ignored and waveform data is not captured. An example of the control signals used in a circulating loop experiment is shown in figure 1. The basic process is to configure the switching to load the loop with data, and then close the loop so that the data continuously propagates around the loop. Once the data has traversed the loop the correct number of times, the oscilloscope is gated to allow it to respond to the always-present trigger signal and acquire only this portion of data. It is important to note that the signals used to control the loop switches as well as the signal used to gate the oscilloscope are defined and provided by the user. The oscilloscope itself does not generate or control any of these signals.

3 Transmitter Transmitter Load state state Transmitter switch switch Trigger gate Figure 1: Timing diagram The exact timing required for loading the loop and circulating within the loop is determined by the time to propagate through the loop and the number of circulations required. It should also be noted that it is ideal to have the duration of the trigger gate enable be less than the round trip time through the loop. This is necessary to guarantee that any signals acquired are temporally distant and within the time between the switching transients as well as to allow for the delay required to enable and disable the oscilloscope gating. In the example mentioned above, the 500 km loop has a roundtrip time of about 2.5 ms. The trigger gate was set to about 1.5 ms in duration to avoid the switching transients. The gating pulse is enabled only after 18 roundtrips through the loop, so the gating function, loop switch, and transmitter switch have periods of 450 ms.

4 Figure 2: Measurement results: Output waveform of loop after 18 circulations If an experiment is to be conducted which requires a greater level of precision in controlling when the gate is enabled and disabled, it becomes important to consider the rate at which the instrument responds to both the enable and disable states of the gating signal. The instrument will be able to respond to a trigger 100 ns after the gate is enabled (low to high transition). When the gating signal makes the high to low transition, eventually the oscilloscope will not accept and respond to trigger signals. In the sampling process, the oscilloscope is triggered and a sample is taken at a time equal to the delay setting (this is a minimum of 24 nanoseconds and is identical to the time position on the display where the sample will be located). Once a sample is taken, the oscilloscope must rearm. This takes approximately 27 microseconds. In the process of rearming, the status of the trigger gate port is checked. Thus if the gate goes from a high to a low just after the status is checked, over 27 microseconds will lapse before the instrument will no longer accept triggers. (In this worst case scenario, only one sample will be taken before the gate is enabled again). The maximum time that can elapse from the time a trigger is accepted, the gating signal drops from a high to a low state, and a data sample is taken is defined by t DISABLE.

5 t DISABLE = 27 us + trigger period + maximum time position on the instrument screen Signal under test Valid data Section of signal actually sampled Gating signal 100 ns < t DISABLE > t DISABLE It is important to make sure that these timing issues are considered to guarantee that data is not acquired outside of the valid data region. It is also important to note that the gating signal has no affect at all on the timing relationship between when the scope is triggered and when the data is sampled. Trigger gating only controls when the oscilloscope will accept a trigger signal. A gating feature exists in the Agilent option 100 oscilloscope mainframe. Some significant differences exist in how this feature has been implemented in the Agilent Trigger gating is a standard feature for all mainframes. In the implementation of trigger gating, the instrument was susceptible to false triggering at the gating transients. These were manifested as vertical streaks across the displayed waveform. This problem has been eliminated in the Agilent 86100A. Note: Carl Davidson of Tyco Submarine performed definition and verification of the experiment example shown above