Internally triggered equivalent-time sampling system for signals having a predetermined data rate

[0002] Equivalent-time sampling systems are used to characterize and display high-speed data signals. An equivalent-time sampling system disclosed by Agoston in U.S. Pat. No. 4,812,769 relies on a trigger signal that is synchronous with the data signal to provide trigger events used for acquiring samples of the data signal. The timing of the acquired samples is progressively varied relative to the trigger events until the number of samples is sufficient to reconstruct the waveform of the data signal on a display. However, a synchronous trigger signal may not be available in optical fiber networks or other communication networks, making it difficult to characterize high-speed data signals within these networks using this type of equivalent-time sampling system. In U.S. Pat. No. 4,678,345, Agoston discloses an equivalent-time pseudorandom sampling system that does not rely on a synchronous trigger signal. Trigger events in this system are provided by the data signal. Time intervals between the trigger events on the data signal establish the relative timing of acquired samples. It is difficult to isolate characteristics of the data signal such as jitter or other timing instabilities using this type of sampling system since timing of sample acquisitions is derived directly from the data signal itself. For example, jitter in the data signal may be masked or enhanced by the equivalent-time pseudorandom sampling system, depending on how the system processes the trigger events. Accordingly, there is a need for a sampling system capable of characterizing a high-speed data signal that does not rely on a synchronous trigger signal and that does not derive the timing of sample acquisitions directly from the data signal.

[0003] According to the preferred embodiment of the present invention, an internally triggered equivalent-time sampling system characterizes high-speed data signals having a predetermined nominal data rate by using an internal oscillator that runs independent of a synchronous trigger signal. The sampling system acquires samples of the data signal at a strobe rate, set by the oscillator, which is lower than the nominal data rate of the data signal. The samples, the strobe rate, and the nominal data rate are communicated to a quality optimizer which compensates for static rate deviations between the nominal data rate and the actual data rate of the data signal to establish timing correspondence between the acquired samples. The timing correspondence establishes the time placement of the acquired samples when characterizing the data signal on a display or other output device. When the oscillator has high timing stability relative to the data signal, jitter and other timing instabilities of the data signal are isolated so that attributes of the data signal are accurately characterized by the sampling system.

[0004] FIG. 1 shows an internally triggered equivalent-time sampling system constructed according to the preferred embodiment of the present invention; and

[0005] FIG. 2 shows a compensation scheme used by a quality optimizer included within the internally triggered equivalent-time sampling system of FIG. 1.

[0006] FIG. 1 shows the internally triggered equivalent-time sampling system 10 constructed according to the preferred embodiment of the present invention. Typically, the sampling system 10 is incorporated within an oscilloscope or other measurement instrument. A high-speed data signal 11 is applied to a sampler 12 that is driven by a strobe signal 13 provided by an oscillator 16. When the data signal 11 is provided by an optical fiber network or other communication network, a trigger signal, synchronous with the data signal 11, is not always available to the sampling system 10. The oscillator 16 does not rely upon the presence of such a synchronous trigger signal and runs independent of the applied data signal 11.

[0007] The sampler 12 acquires samples of the applied data signal 11 at time intervals determined by a strobe rate R_Sof the strobe signal 13. An analog-to-digital (A/D) converter 18 coupled to the output of the sampler 12 digitizes the samples to produce sample points 15. A quality optimizer 20 receives the sample points 15 directly from the A/D converter 18, or alternatively, from a memory 19, such as a FIFO, as shown. The quality optimizer 20 is implemented within a central processing unit CPU (not shown) associated with the sampling system 10 or is implemented using a digital signal processor or other type of processor. The quality optimizer 20 derives a timing correspondence between sample points 15 and drives a display or other output device (not shown) used to reconstruct segments of the data signal 11.

[0008] The data signal 11 applied to the sampling system 10 has a nominal data rate R_Ethat is communicated to the quality optimizer 20. In one example, the data signal 11 is supplied by a communication network, such as a SONET (synchronous optical network) OC48 which has a nominal data rate R_Eof 2.48832 Gigabits per second, and the nominal data rate R_Eis the standard data rate specified within the network. Generally, the nominal data rate R_Eis a predetermined rate that is provided to the quality optimizer 20 by a user of the sampling system 10, or alternatively, is preset within the quality optimizer 20.

[0009] Typically, the nominal data rate R_Eof the data signal 11 is substantially higher than the maximum operating rate of the sampler 12, and the strobe rate R_Sis substantially lower than the nominal data rate R_E. The strobe rate R_Sset by the oscillator 16 is communicated to the quality optimizer 20 and is chosen so that integer multiples of the strobe rate R_Sare offset from the nominal data rate R_E. The timing resolution of the acquired sample points 15 increases as the offset decreases. The offset assures that the timing of acquisitions of the sample points 15 progressively varies relative to cycles of the data signal 11 to provide for equivalent-time sampling of the data signal 11. However, it is difficult to reconstruct segments of the waveform of the data signal 11 using known equivalent-time sampling techniques because the data signal 11 has an actual data rate R_Dthat differs from the nominal data rate R_Eby an unknown rate deviation. This difference between the actual data rate R_Dand the nominal data rate R_Etypically has a static rate deviation and time-varying rate deviations. The static rate deviation includes fixed rate shifts or rate offsets between the actual data rate R_Dand the nominal data rate R_E, whereas the time-varying rate deviation includes jitter and other timing instabilities of the data signal 11. When characterizing the data signal 11, it is often desirable to isolate and display jitter and the other timing instabilities, while compensating to suppress the effects of static rate deviations.

[0010] The quality optimizer 20 compensates for the static rate deviations between the actual data rate R_Dand the nominal data rate R_Eto derive the timing correspondence between the sample points 15. When the timing stability of the oscillator 16 is sufficiently high, the strobe signal 13 has high timing stability relative to the timing stability of the data signal 11, enabling the jitter and other timing instabilities of the data signal 11 to be isolated so that attributes of the data signal 11 can be accurately characterized by the sampling system 10.

[0011] FIG. 2 shows a scheme 30 used by the quality optimizer 20 to compensate for static rate deviations between the nominal data rate R_Eand the actual data rate R_Dof the data signal 11 and to derive timing correspondence between the sample points 15. In step 32, the sample points 15, the nominal data rate R_Eand the strobe rate R_Sare communicated to the quality optimizer 20. When the static rate deviation between the nominal data rate R_Eand the actual data rate R_Dis too large, the acquired sample points 15 are insufficiently spaced to accurately reconstruct segments of the data signal 11. Generally, the accuracy and timing resolution of the reconstructed segments of the data signal 11 are improved as the static rate deviation, or rate offset, between the nominal data rate R_Eand the actual data rate R_Ddecreases. Thus, the more closely the nominal data rate R_Eis set to the actual data rate R_D, the greater the achievable accuracy and timing resolution.

[0012] In steps 34-40, the timing correspondence between the acquired sample points 15 is established based on analysis of the acquired sample points 15, the nominal data rate R_E, the strobe rate R_Sand an estimate R′_Dof the actual data rate R_D. In step 34, the acquired sample points 15 are analyzed to locate successive signal events within the acquired samples 15. The signal events are predefined characteristics within the sample points 15 such as predetermined amplitude levels, or rising or falling edge transitions detectable within the sample points 15. In this example the successive signal events are a rising edge followed by a falling edge transition detectable within the sample points, or alternatively, the successive signal events are a falling edge followed by a rising edge transition detectable within the sample points. Interpolation between data points 15, binary searching or other searching techniques are used to accurately locate the successive signal events when the events occur between sample points 15. While signal events may occur at various spacings within the acquired sample points 15, for example when the data signal 11 is a random or pseudorandom bit stream, the successive signal events correspond to the closest spaced signal events within the sample points 15 and are sought in step 34.

[0013] Once the successive signal events are located within the acquired sample points 15, a number M, which includes both the integer number of sample points counted between the successive signal events and a fractional portion resulting from the interpolation or searching between sample points 15 for the successive signal events, is determined in step 36. As an alternative to counting the number of sample points 15 between successive signal events and searching between sample points for the successive signal events, M is determined by locating non-successive signal events within the sample points 15, counting the integer number and fractional portion of sample points between the non-successive events and dividing this result by the equivalent number of successive signal events occurring between the non-successive signal events.

[0014] Based on the determined number M, the nominal data rate R_Eand the strobe rate R_S, the static rate deviation between the nominal data rate R_Eand the actual data rate R_Dis compensated. In step 38, the compensation is achieved by performing an estimate R′_Dof the actual data rate R_Daccording to the relationship:

[0015] R′_D={M*FLOOR(R_E/R_S)+1}/{M* MOD(R_E/R_S)+(M/R_E)*FLOOR(R_E/R_S)}

[0016] where FLOOR indicates the integer portion of the expression in parentheses following the FLOOR and where MOD indicates the fractional portion of the ratio in parentheses following the MOD. The timing correspondence between sample points 15 is then determined in step 40. The timing correspondence involves determining the time placement of acquired sample points 15. This time placement is used to reconstruct segments of the data signal 11 on a display or output device. In this example, the time placement of the Nth acquired sample point 15 equals MOD(N *R′_D/R_S).

[0017] While the preferred embodiment of the present invention has been illustrated in detail, it should be apparent that modifications and adaptations to this embodiment may occur to one skilled in the art without departing from the scope of the present invention as set forth in the following claims.