Data processing: measuring – calibrating – or testing – Measurement system in a specific environment – Electrical signal parameter measurement system
Reexamination Certificate
2002-05-06
2004-06-08
Shah, Kamini (Department: 2863)
Data processing: measuring, calibrating, or testing
Measurement system in a specific environment
Electrical signal parameter measurement system
Reexamination Certificate
active
06748335
ABSTRACT:
FIELD OF THE INVENTION
The subject invention concerns, in general, the field of multi-channel Digital Storage Oscilloscopes (DSOs) having long record length capability, and concerns in particular, an architecture for DSOs that provides improved digital signal processing of long record length waveform data.
BACKGROUND OF THE INVENTION
The long-record-length features of modern long-record-length oscilloscopes are generally found to be very difficult and cumbersome to control. Even when the available memory is apportioned to active channels, the amount of data collected is difficult to use effectively (referred-to herein as a relatively long data record).
That is, when one has collected 8 Mbytes of data from four channels in such a deep memory oscilloscope, how does one then use and interpret that data? For example, assume a display of four rows of data (one for each active channel), and further assume that the user wanted to scroll through the entire record looking for a particular event that caused a problem in the system under test. For such a visual scan, a scrolling-rate of about 500 points per second is quite reasonable. That is, a particular point on each waveform would move across the screen from right to left in about 1.0 second. Unfortunately, at this rate it would take approximately 4.375 hours for the user to view the entire data record.
The fact that many oscilloscopes include a printer might lead one to think that the solution to this problem would be to merely print out the entire record. For such a print out, a resolution of 300 points per inch is quite reasonable. Unfortunately, if the user were to print out such a relatively long record on paper in four rows at 300 points per inch (approximately 118 points per cm), the printer would use 0.421 miles (0.6736 km) of paper. These two examples highlight the difficulty in dealing with large amounts of data. It simply is not practical for the user to visually inspect all of the collected data for the anomalies that the user must find.
Modern DSOs attempt to solve this problem by waiting for a trigger event to occur, and then acquiring in memory a frame of waveform data surrounding the event. The frame is then processed by waveform math software, measurement software, and display system software. All of this post-processing creates extremely long periods of “dead time”, in which the DSO is incapable of acquiring and storing additional waveform samples. As a result, the anomaly that the user is searching for may occur, and be missed.
More recent DSOs have attempted to reduce the “dead time” by physically positioning Digital Signal Processing (DSP) ICs close to the acquisition memory to convert acquired waveform data to display data more efficiently. This arrangement is sometimes referred to as a “FastAcq” mode of operation. Use of FastAcq circuitry has greatly reduced the “dead time” between triggers, and increased the number of samples per second that are displayed. Unfortunately, the data frames processed by the FastAcq circuitry are not retained, and are therefore unavailable for additional processing. Moreover, cycle-to-cycle measurements (for jitter measurement) are adversely affected by the use of FastAcq circuitry because the time relationship between successive triggers is not maintained.
Another disadvantage of many current DSO architectures is a “bottleneck” that exists because they transfer all of the data from acquisition memory to main memory for processing and display over a relatively slow (i.e., typically 30 Mb/sec.) data bus.
In order to address this transfer-rate issue, Agilent Technologies, Inc. of Palo Alto, Calif., has recently introduced Infiniium MegaZoom deep-memory oscilloscopes employing a custom ASIC that optimizes the sample rate for a given sweep speed and sends only the waveform data needed for a particular front panel setting. MegaZoom provides a waveform update rate that is approximately twenty-five times greater than conventional deep memory oscilloscopes.
Wavemaster™ oscilloscopes with X-Stream™ technology, manufactured by LeCroy Corporation of Chestnut Ridge, N.Y. provide an alternative solution to the transfer-rate problem. These oscilloscopes employ a silicon-germanium (SiGe) digitizer and a high-speed streaming bus to transfer data from an analog to digital converter (ADC) through an acquisition memory and into a memory cache for extraction of information by software routines.
However, what is needed is an oscilloscope having the capability to repeatedly “loop through” the four-channel relatively-long data record in order to detect predetermined anomalies and produce a lively and active display.
SUMMARY OF THE INVENTION
A real time multi-channel digital storage oscilloscope acquires a relatively long data record for each channel in an acquisition memory and processes the data of the relatively long data record to search for predetermined events. Upon detection of such a predetermined event, circuitry generates an event detect signal, and data comprising an acquisition frame surrounding the event is applied to a waveform processing and display system. The relatively long data record can be replayed in order to perform additional searches throughout the data record using different search criteria, thereby permitting multiple waveforms to be displayed simultaneously, each being captured as a result of a different user-defined event. A screen display may be programmed to display a different kind of event such as Runt signal, Overshoot, or Pulsewidth Violation in each waveform, or to display multiple occurrences of the same kind of event such as Runt signal in each waveform. The multiple waveforms of the screen display may be derived from a single channel or from different channels.
REFERENCES:
patent: 4975636 (1990-12-01), Desautels
patent: 5255365 (1993-10-01), Hungerbuhler
patent: 5268851 (1993-12-01), Zivny
patent: 5438531 (1995-08-01), Shank
patent: 5898420 (1999-04-01), Timm
patent: 5929838 (1999-07-01), Hall
patent: 5978742 (1999-11-01), Pickerd
patent: 5986637 (1999-11-01), Etheridge et al.
patent: 5999163 (1999-12-01), Ivers et al.
patent: 6151010 (2000-11-01), Miller et al.
patent: 6195617 (2001-02-01), Miller
patent: 6219094 (2001-04-01), Gerlach et al.
patent: 6222521 (2001-04-01), Ivers et al.
patent: 6271773 (2001-08-01), Kobayashi
patent: 6615369 (2003-09-01), Beck et al.
patent: 2001/0001850 (2001-05-01), Miller
patent: 2003/0220753 (2003-11-01), Pickerd et al.
Gray Francis I.
Lenihan Thomas F.
Pretlow Demetrius R.
Shah Kamini
Tektronix Inc.
LandOfFree
Acquisition system for a multi-channel relatively long... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Acquisition system for a multi-channel relatively long..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Acquisition system for a multi-channel relatively long... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3364382