Regardless how tempting, many test engineers still hesitate to abandon their trusted analog oscilloscope (AO) in favor of the potentially more powerful digital storage oscilloscope (DSO). Since DSOs have so much to offer, oscilloscope designers and manufacturers devote a great deal of energy searching for the reasons for this.
In a recent survey conducted by LeCroy, AO users were asked why they had not converted to DSOs. A few said price, a few said they didn’t need digital measurements, but the overwhelming reason was fear of a long learning curve.
Other test professionals who use both AOs and DSOs expressed concern about the response-time limitations and the inability to see vital signal characteristics or anomalies because of excessive blind times typical of DSOs. Fortunately, all these issues are being addressed by major oscilloscope manufacturers—but in a variety of ways that can make selecting a DSO a challenge.
After all, when you chose an AO, all you had to specify was the number of channels, bandwidth, sensitivity and measurement features. For a DSO, you have all those characteristics plus sampling rate, amplitude resolution, memory length and display-generation capabilities.
A very simple DSO with a 20-MHz sampling rate and a 1k-word record length may be adequate for many applications, but a 200-, 500- or 1,000-MHz sampling rate and 1 MB of memory may be needed for others. Also, minimal blind time—the period during which signal acquisition cannot take place—may be essential.
To understand the key AO vs DSO performance differences, let’s look at the architectures that overcome the signal fidelity, response and blind-time problems. The focus will also be on how DSOs can be endowed with the analog-look-and-feel characteristics that provide greater ease of use and shorten the learning curve.
Strength and Shortcomings
By their very nature, DSOs only collect samples of signals and may miss perturbations occurring between sampling intervals. More importantly, the samples must be quantized by an A/D converter, time-qualified, processed, stored, read out and reprocessed for display—a series of processes that can be very time-consuming. While these functions take place, the conventional DSO is temporarily unable to acquire any more data and, in effect, is blind to anything happening to the signal.
As a result of lengthy acquisition processing times, the update rates of conventional DSOs are usually below several hundred waveforms per second. Operational delays also may be encountered when the control circuitry of the DSO is required to simultaneously perform other nonacquisition-related tasks.
“For instance, some scope displays lag behind when the time base or volts/div setting is changed,” said Marianne McTigue, Product Marketing Engineer at Hewlett-Packard. “A slow user interface can be very frustrating when troubleshooting a circuit.” It is similarly irritating when the DSO response is slow during UUT adjustments.
On the positive side, waveform storage innately provides a host of advantages. These include capture and preservation of single transient events, extended signal analysis via post-processing software, record keeping and report-preparation ease.
Operational advantages are also many. “Using a single ‘snapshot’ key, a waveform may be held on the display and you may grab other new or stored waveforms for comparison,” said Skip Morey, Product Manager at Yokogawa Corp. of America. “The history key on our DL1540 allows you to store and access the previous 100 displays in memory.”
AOs, of course, don’t store waveforms. But in contrast to DSOs, several hundred thousand waveforms may be acquired per second, since the AO acquisition rate is limited only by the trigger holdoff and the beam-retrace time. As a result, AOs do not miss intermittent or infrequent anomalies like metastable conditions, noise, time jitter and dropouts.1 The fast acquisition update rate also assures that results of UUT adjustments are immediately observable on the screen.
With AOs, the signals containing several widely different frequency components, such as modulated RF signals, are readily observable with a single time-base setting. To obtain an equivalent presentation, the DSO user may have to use alternate time bases, peak-detect circuits or long memories.
AO beam intensity also conveys some valuable signal-speed-related information. The faster the signal is, the less time the beam has to excite the phosphor particles on the screen and the dimmer is the trace. As a result, AO display intensity can provide more insight into signal characteristics. Conversely, some transients may be so fast that they are not readily observable on the AO and only a DSO can be used to capture and display them.
Overcoming DSO Limitations
The easiest way to maintain the advantages of an AO is to include it in the DSO box, suggested Charles Holtom, Product Manager Oscilloscopes at Fluke. Fluke combines a DSO and an AO in one instrument, and each mode is available by touching a button. Other companies, such as B+K Precision, L G Precision and Tektronix, also equip some of their DSOs with an analog operational mode.
While these combined scopes are very appropriate for many applications, many situations demand different or more extensive DSO facilities. Conventional DSOs may be well-suited to fulfill some of these needs; and DSOs with high update rates, fast response times and extensive display-processing capabilities provide full AO-like performance. These latter scopes furnish true representations of the most complex or elusive signals.
The key ingredients responsible for this improved performance are multiprocessing and long memory. Using dedicated microprocessors for the control, acquisition and display-processing functions improves response times and update rates. Longer memory stores more captured data, increasing the likelihood of catching important events and signal aberrations.
For instance, the newest members of the Hewlett-Packard HP 54600 Series feature MegaZoom technology, an architecture that uses multiprocessing as well as long memories. Four processors are used in the signal path, simultaneously performing different functions.
One processor is dedicated to performing user-interface tasks and managing I/O ports. Another controls signal acquisition while a DSP executes acquisition memory functions and a waveform translator quickly draws the waveforms.
The memory is partitioned; and while samples flow into one area of the acquisition memory, they are being processed and organized for display in another. Signal inputs appear immediately on the screen, response to signal or setting changes is fast and 1 MB of memory behind each channel provides a substantial sample-rate and data-accumulation advantage.2
An extremely high data acquisition rate—exceeding 400,000 waveforms per second—is provided by the Tektronix TDS 700A series of DSOs which features the push-button-activated InstaVu™ display mode. Figure 1 is a block diagram depicting the architecture of these scopes.
Signals are sampled and digitized at rates up to 4 GS/s and applied to a custom demux chip which includes a DSP integrated with a high-speed display rasterizer. The rasterizer builds a display bit map accumulated from the multiple triggered acquisitions of the input signal.
Besides displaying many acquisitions as a single raster image, InstaVu achieves rapid acquisition rates by allowing the system to re-arm itself and acquire a new signal as soon as it has completed each acquisition. This almost continuous acquisition capability facilitates the rapid display of rare bus contentions, infrequent glitches, time jitter and metastable circuit behavior.1 Figure 2 provides an example of how InstaVu can display noise and signal disturbances which may not be as readily identifiable without that feature.
In addition to these DSOs, many others employ multiprocessing or feature long memories—often one million words deep and some even longer. For instance, the Fluke CombiScope uses three microprocessors and LeCroy provides DSOs with very deep memory.
Deep memories provide higher signal resolution at long time-base settings and improved timing precision leading to enhanced post-processing results.3 But for some applications, not only time but also amplitude assessments must be performed with AO-like resolution and precision.
“The 0.4% amplitude resolution of 8-bit DSOs is adequate for many electronic signals, but physical measurements such as shock, vibration or strain measurements generally require a higher dynamic range,” said Gary Schneider, Senior Product Manager at Nicolet. “Twelve-bit digitizers provide 16 times more resolution for the detailed assessment of transducer-generated and other low-level signals.”
Applications determine how long a memory your DSO should have and whether high resolution or high data acquisition rates should be of prime concern. Regardless of the features provided, ease of use is always welcome.
Analog-Like Ease of Use
After all samples have been stored in a deep memory, how do you display just the portion of the acquired signal that you are interested in? Many times, the scope takes care of that automatically when you have triggered it by the event of interest. At other times, you may want to examine the entire record and zoom in on a particular detail.
“But many current DSOs have a default display of only 500 points,” cautioned Dr. Michael Lauterbach, Director Product Management at LeCroy. “If you are not an advanced scope user, you may not know that a scope capturing 100,000 points may put only 500 of them on the screen, while the other 99,500 are off screen to the left or right. To prevent confusion, LeCroy scopes default to display the full data record on the screen. This gives more of an analog scope feel—you don’t have to worry about telling the DSO how much data to acquire and which percentage to put on the screen.”
The zooming process also has been simplified. AO-like pan and zoom controls are provided not only by MegaZoom in the HP 54600 Series but also by other companies. For instance, the Yokogawa DL1540 includes a zoom-control button and a rotary knob.
“The zoom button splits the screen and both the main sweep and the zoomed portion are displayed,” explained Mr. Morey. “The rotary knob automatically selects the magnification and lets you move the zoom area to bracket the event of interest. The zoomed portion of the waveform is then displayed across the width of the screen.”
The capability to control the scope by turning knobs and pushing buttons instead of using a series of menus is paramount to many users. While AOs offer direct access to all functions through hard-wired switches and controls, many DSOs have replaced these with keypads and rotary encoders. But to provide you with the interface you are comfortable with, manufacturers are minimizing dependence on multiplexed controls and menus.
“While some DSOs still require up to three keystrokes to change simple things like the trigger slope or the input coupling, other companies, such as Fluke, have another approach,” said Mr. Holtom. “For example, the PM 3380A input coupling, along with the triggerslope and the source, are directly accessible; and all major functions have dedicated controls.”
“Many four-channel DSOs require multiple keystrokes simply to change position,” also commented Mr. Schneider. “All Nicolet DSOs include individual knobs and buttons for each channel so you do not need menus or softkeys for the most common adjustments.”
To determine which functions should be directly controllable as well as to arrive at ergonomically optimized panel layouts, HP convened focus groups and performed in-depth utilization studies. Tektronix invested hundreds of hours observing customers use oscilloscopes and recording work patterns, which later provided the foundation for the intuitive user interface for the company’s TDS-series DSOs.4
“Functions used most often must be most easily accessible,” stated Mark Lombardi, Product Marketing Engineer at Hewlett-Packard. “The HP 54600 series addresses this with knobs that directly access vertical and horizontal, scaling and position, the trigger level and holdoff functions.
“Controls that affect waveform display are given the highest priority, but the next level of functionality is addressed through a combination of buttons and softkey menus,” Mr. Lombardi continued. “Keys are arranged in logical groupings and allow access to functionality without paging through multiple levels of menus. For example, measurements are accessed by pressing the volts, time or cursors button.
Similarly, LeCroy offers dedicated and labeled knobs for time/div, volts/div, trigger level and offset. “If you have ever used an AO, you can adapt easily to one of our DSOs—and without relying on menus,” commented Dr. Lauterbach.
Conclusion
In some situations, AOs are still preferable for price/performance reasons; but whenever waveform storage or post-acquisition processing is required, a DSO is a must. Fortunately, new DSO architectures and implementations have successfully overcome almost every functional or operational shortcoming that could have made an AO mandatory. But in any case, more and more of today’s DSOs have the analog look and feel that many users are most comfortable with.
References
1. Chhabria, V., and Finnerty, J., “Analog or Digital? Delving in the Measurement Dilemma and Delivering the Best of Both,” Tektronix, March 1996.
2. “MegaZoom Technology Improves Oscilloscope Performance and Lowers Cost,” Hewlett-Packard Backgrounder, May 1996.
3. Lauterbach, M., “Understanding Long Memory in a DSO”, EE-Evaluation Engineering, July 1995, pp. 17-22.
4. “The TDS Series User Interface, Designed With and for the User,” Tektronix, TDS User Interface Backgrounder, June 1991.
These companies provided information for this feature:
Fluke (800) 44-FLUKE
Hewlett-Packard (800) 452-4844
LeCroy (914) 578-6020
Nicolet Instrument Technologies (608) 276-5600
Tektronix (800) 426-2200
Yokogawa Corp. of America (800) 258-2552
Copyright 1996 Nelson Publishing Inc.
July 1996
|
|