Rigol's UltraVision technology relies on the interaction between specialized hardware and intelligent software. The A/D converted data is handled by a hardware-based sampling controller. Large memory is directly connected to the controller, so data storage occurs without requiring CPU intervention. Each CPU action increases the so-called "blind time" and reduces the waveform capture rate on the opposite side. Similarly, the waveform displayed on the screen is generated by a special waveform tracer. This, too, is accomplished without increasing the CPU workload.
The main functions are hardware-based.
Other functions, such as triggering, pass/fail decision, and waveform recording, are implemented in hardware, so the CPU is only minimally involved in handling these functions. All of this serves to maximize the waveform acquisition speed. To achieve optimal results from this high waveform repetition rate, the instrument incorporates a display with multiple intensity levels. Multiple waveforms are displayed on the screen, one on top of the other. Depending on the frequency of occurrence, a different intensity is selected. This means that signals that appear infrequently are displayed with lower intensity. There are 256 levels available, and as a result of this functionality, it is possible to see at a glance how frequently errors or anomalies occur.
Another extremely important part of UltraVision's circuitry is its enormous data memory. This allows it to maintain the maximum sampling rate, especially with larger time/division values. This is crucial when acquiring longer time periods without losing signal detail.
In addition, Rigol offers a function called Record, which is also part of UltraVision technology. This function combines intelligent triggering with efficient data storage. The acquired signal is stored in unique memory segments. The advantage is that only rarely occurring, important, and interesting events are stored in the available memory, and not all (uninteresting) data in general.
But why should we consider the signal capture speed? The reason lies in the technology of digital oscilloscopes. When all oscilloscopes were analog, this specification was virtually unknown or insignificant. This is because analog oscilloscopes have no data handling or storage, or any other post-processing. The input signal is used directly for the vertical deflection of the electron beam. There is only a very brief period during which no curve is displayed on the CRT screen: when the electron beam switches from right to left. As the world has moved into the digital age, oscilloscopes have also become digital. This shift has brought with it a vast array of possibilities, such as simple data storage, mathematical calculations, and much more, but these advantages also come with some drawbacks. One of these is the so-called "blind time," which is the time comparable to the previous beam return time, but the cause is different. In digital oscilloscopes, the input signal is conditioned and converted from analog to digital in near real-time, but the data must then be handled and processed. While this is happening, no more data can be captured at the input. The time elapsed before new data can be stored at the input is defined as "blind time." This time interval is so long that it cannot be ignored. The time during which "nothing is visible" can reach 99.9%. In other words, this means that you can only see for 0.1% of the time. All oscilloscope manufacturers try to reduce "blind time" and increase the signal capture speed with different techniques, such as using higher-performance processors. Rigol's solution to "blind time" is the UltraVision circuit set.
The advantages of a higher waveform capture rate will be explained in detail below. As an example, consider a 1 MHz square wave with infrequent interference. The error occurs 10 times per second.
Let's first take an oscilloscope with a capture rate of 1,000 waveforms per second. We set a time base of 500 ns/div. This results in a total display time (14 screen divisions) of 7 ms. Assuming no blind time, we could say that we see interference for every 14,285 "screen captures" on average; equivalent to a time of 14,285 * 7 ms = 100 ms (on average).
But let's not forget that this is the theoretical value without blind time. This means achieving a capture rate of approximately 143,000 waveforms/s. With a capture rate of only 1,000 waveforms/s, and without taking probability into account, we obtain an average time of 14.3 seconds until we see an error on the screen.
This simple calculation makes it easy to understand the effect that a higher waveform update rate will have on the detection time of a rarely occurring error. A signal capture rate ten times higher theoretically leads to error detection ten times faster.
Errors can be found and analyzed more quickly.
In real-world laboratory environments, a higher waveform acquisition speed offers two important advantages for faster and more accurate operation. First, it's essential to be able to detect anomalies or errors. One of the worst things is that errors that occur are rarely overlooked and can cause problems in the final product. Second, it's crucial to significantly reduce testing or measurement time without compromising the first issue mentioned earlier. This is important because, as everyone knows, time is money.
These two points can be achieved by using oscilloscopes with a higher waveform capture rate. But a high refresh rate isn't everything. Why not? Imagine a waveform acquisition rate of 50,000 waveforms/s and a TFT display refresh rate of 50/s (50Hz). This results in thousands of waveforms per display refresh, which will appear overlapping on the screen. If all the waveforms are displayed with the same color or intensity, it won't be possible to distinguish rare events from regular ones.
This is where the 256-level intensity display comes into play. Thanks to it, signals that occur infrequently will be displayed with less intensity than signals that occur frequently. Now it's possible to differentiate between the two and take full advantage of the high waveform capture speed.
But Rigol's UltraVision is capable of much more. All oscilloscopes incorporating this technology come standard with deep data memory. Why is deep memory important?
We can all agree that if only short or fast signals, or disturbances, are being visualized, the key lies in a high sampling rate and adequate bandwidth (the key word being rise time). But if the signal needs to be monitored over a longer period and there are other signal peaks or dips to analyze, deep memory is absolutely essential. To make this clearer, let's first look at the mathematical relationship between sampling rate, time base, and memory:
The Time/div and number of divisions factors can be combined in the total time displayed.
The formula is thus obtained:
A larger data memory allows for more accurate measurements.
In this formula, the data memory is a fixed parameter, while the other two can be changed. The total displayed time is affected by the application. Therefore, the sampling rate is the parameter that will be adjusted within the formula. This graph can be obtained from this relationship.
What can be deduced from this graph? In short, it can be said that the more memory available, the longer the maximum sampling rate can be maintained.
The following example is artificially created to highlight the advantage of more memory. The goal is to generate the need to capture data over a long period at the highest possible horizontal resolution. Our test signal will be a pulse train with 8 bursts. The interval between bursts will be 8 ms. The following graph shows the shape of each burst.
The goal is to capture all eight bursts in a single shot, which requires a minimum total display time of 64 ms. Depending on the available time/div setting and the number of divisions, we choose 5 ms/div. With the 14 divisions on our oscilloscope, the result is a total display time of 70 ms.
In the following steps, we will perform the acquisition twice, but with different memory depths. In the first measurement, the memory depth will be 1.4 MPts, and in the second acquisition, a memory depth of 140 MPts will be used. After performing both measurements, we will have a detailed view of the single-pulse shape of the pulse train.
The result: in the measurement with the smallest memory, the sampling rate is adjusted according to the formula above.
(In the image below the screen, the sampling rate is 12.5 MS/s, depending on the oscilloscope used.) If you look at the acquired and enlarged burst, you can see the lost information. This is because at the lower sampling rate, the time between sampling points is longer, and detailed information can no longer be obtained.
The measurement view with maximum data memory indicates that the maximum sampling rate is only reduced by a factor of two (down to 2GS/s). The deep burst view demonstrates that its original shape can be reproduced.
The Rigol oscilloscope offers a recording function in addition to the deep memory. This function is useful when analyzing signals and/or serial buses.
What lies behind the recording function? At first glance, it's simply an additional way to store signals. But it actually has much more potential. Combined with the trigger functions available on the oscilloscope, it becomes a very powerful analysis tool. Furthermore, it provides optimized memory management. Therefore, only the important parts of the signal are stored, and not, say, irrelevant signals that waste memory.
The example is again the pulse train discussed earlier. With the appropriate triggering parameters, we now fire on the burst. The time base can be adjusted to a smaller value. Only the data related to the triggering event will be stored in memory. The time between each pulse (8ms), during which nothing happens, will not be saved.
An additional application that could be interesting is the combination of the register function and serial bus decoding. For example, let's consider an I2C bus with more than three participants. The goal is to analyze only one slave (read and write). Rigol offers a simple solution for this task: the serial bus trigger and register function.
The serial bus trigger will be configured to fire at the address of the participant to be monitored. The logging function will then be activated, and only messages originating from or destined for the selected address will be stored. After logging is complete, each stored message can be reviewed to verify that everything is working correctly. Furthermore, saved messages can only be played back automatically after a delay, allowing for further viewing.
Rigol, as an instrument manufacturer, also works on the development and distribution of the LXI (LAN eXtensions for Instrumentation) standard. This standard includes specifications and solutions that enable the integration of instruments and measurement systems with LANs. Rigol is a strategic member of the LXI Consortium, whose mission is to promote and advance the standard. "Strategic member" status is the highest possible level of membership in LXI and entitles members to a seat on the board of directors of this non-profit organization.
