HP E42 manual Model 5328A Theory of Operation Resolution, Time Interval Averaging

Model 5328A

Theory of Operation

4-24.Resolution

4-25. The resolution of the measurement is determined by the frequency of the counted clock (e.g., a 10 MHz clock provides 100 ns resolution). The elements within the time interval counter (input amplifier, main gate, DCA’s) must operate at speeds consistent with the clock frequency, otherwise the instrument’s resolution would be meaningless. The 5328A counts a 10 MHz clock.

4-26. Clock frequencies of 1, 10, 100 MHz, and other 10n frequencies, are preferred since the accumulated count, with the appropriate placement of decimal point, gives a direct readout of time interval. This explains why the conventional time interval counter is at present limited to 10 nanoseconds, a clock frequency of 100 MHz. 1 GHz is beyond reach and a clock frequency of 200 MHz would require some arithmetic processing of the accumulated count in the DCA’s to enable time to be displayed directly.

4-27.Time Interval Averaging

4-28. This technique is based on the fact that if the ±1 count error is truly random it can be reduced by averaging a number of measurements. The words “truly random” are significant. For time interval averaging to work, the time interval must (1) be repetitive, and (2) have a repetition frequency which is a synchronous to the instrument’s clock. Under these conditions the resolution of the measurement is:

where N = number of time intervals averaged

4-29. With averaging, resolution of a time interval measurement is limited only by the noise inherent in the instrument. Ten picosecond resolution can be obtained with the 5328A. Most time interval averaging suffers one severe limitation; the minimum measurable time interval is limited to the period of the clock. This limitation is removed by circuits known as synchro- nizers which are used in the 5328A to measure intervals as short as 100 picosecond.

4-30. The 5328A synchronizers operate as shown in Figure 4-5.The top waveshape shows a repetitive time interval which is asynchronous to the square wave clock. When these signals are applied to the main gate, an output similar to the third waveform results (no synchronizers). Note that much of this output results in transitions of shorter duration than the clock pulses. DCA’s designed to count at the clock frequency are unable to accept pulses of shorter duration than the clock. The counts accumulated in the DCA’s will therefore approximate those shown in the fourth trace — the exact number of counts is indeterminant since the number of short duration pulses actually counted by the DCA’s cannot be known. Since the time interval to be measured is slightly greater than the clock period, the fourth waveshape shows that the average answer will be in error, having been biased, usually low, because of the DCA’s requirement of having a full clock pulse to be counted.

4-31. This problem is alleviated by the synchronizers which are designed to detect leading edges of the clock pulses that occur while the gate is open. The waveshape applied to the DCA’s, when synchronizers are used, is shown by the fifth waveform. The leading edges are detected and reconstructed, such that the pulses applied to the DCA’s are of the same duration as the clock.

4-32. Synchronizers are a necessary part of time interval averaging; without them the aver- aged answered is biased. In addition, it may easily be seen that with synchronizers involved, time intervals of much less than the period of the clock can be measured, This technique is only as good as the synchronizers, however. The 5328A high-speed synchronizers enable intervals as small as 100 picosecond to be measured.

4-4