Oscilloscope waveforms show real world electronic signals. How well the scope displays those waveforms determines its performance. Assuming two scopes have adequate base specifications – like bandwidth, sample rate and frequency response – does the better scope show thin or fat waveforms? Like most engineering questions, the answer is 'it depends'.
The update rate indicates how many waveforms per second can be acquired, processed and displayed. The higher the rate, the more quickly the scope will display the signal tested and its subtle details.
Modern scopes have update rates ranging from 1million waveform/s to one every few seconds. Update rate is impacted by a variety of settings, with acquisition memory depth having the biggest impact. Let's look at a quick example. The upper portion of fig 1 shows two scopes of equal bandwidth connected to the same 10MHz sine wave source. However, one shows a fatter waveform. This will cause differences in measurement values. Which is more accurate? One of the biggest differences between these scopes is update rate. One updates at 1m waveform/s while, using the same settings, the other updates 16,000 times slower.
How does this impact the displayed waveform? The lower portion of fig 1 shows what each scope displays if infinite persistence is turned on. Both scopes build the picture over time. After 10s, the scopes show the same wave shape with the same waveform thickness. In this case, the scope with the faster update rate showed a thicker wave shape that was more representative of what each should have displayed.
Scope noise impact
How accurate are scope measurements? Typically, they are extremely accurate from a horizontal timebase perspective, but substantially less accurate vertically. Why? One reason is measurement disturbance created by noise.
Oscilloscopes generate noise that is coupled on the signal under test. The oscilloscope's a/d converter can't discriminate between noise generated by the scope and the target signal. But a simple test can determine how much noise your scope adds to the signal. Fig 2 shows two oscilloscopes viewing the same 10MHz sine wave. One shows a much thicker waveform; is this because of a faster update rate? The answer is no. The two scopes have identical update rates and if we turn on infinite persistence, one would continue to show thick waveforms, while the other would show thin ones.
The difference is that one scope generates more noise and this causes a fatter signal.
Other sources of noise include active and passive probes. Active probes typically use the 50? signal paths for scope channels and these paths have lower noise than their 1M? signal path counterparts.
What's a quick way to check scope noise? Most vendors will characterise noise for specific models and include values on the datasheet. If not, ask for the information or find out yourself – it only takes a few seconds.
Disconnect all inputs from the front of the scope and set the scope to 50? input path – you can also run the test for the 1M? path. Turn on a decent amount of acquisition memory – 100kpt to 1Mpt will suffice – run the scope with infinite persistence and measure the height of the waveform. The thicker the waveform, the more internal noise is being produced. The scope will have unique noise qualities at each vertical setting. You can look at wave shape thickness or be more analytical and take an ac rms voltage measurement. Change the vertical settings to more a sensitive value, say 10mV/div instead of 100mV/div, and you will see noise increase as a percentage of full scale vertical values.
Scopes with lower noise show thinner waveforms if the initial signal is skinny and will produce better viewing and measurements. Changing your scope settings to reduce bandwidth eliminates some broadband noise, making signals thinner. Scope vendors have a number of methods – such as averaging, high resolution mode and bandwidth limiting – to mitigate inherent scope noise. Having low noise to start with reigns supreme as these noise mitigation settings work equally well, if not better, on scopes that start with low noise.
Target signals can be low noise or noisy. It's sometimes difficult to determine whether the noise is coming from the target or from the scope itself. When the scope's a/d converter digitises signals, it can't discriminate between signal noise and scope noise; it simply stores the output and displays the resulting values. Is the fat wave shape representative of your test signal or your scope? There are several methods to answer this. First, do a quick assessment of your scope's internal noise (see above). Expect this much deviation to be added at any sample point. Turn on infinite persistence to see if the wave shape gets fatter or stays thin.
Infinite persistence will also show the impact of scope noise on the target signal. Doing a quick test with a known wave shape and seeing how much the waveform changes between normal display and infinite persistence modes provides a quick overview of noise and update rate for that scope. A noisy scope with a slow update rate will initially show a thin waveform, building to a thick waveform when infinite persistence is turned on. A noisy scope with a fast update rate will show a thick waveform almost instantly – whether or not the signal under test is skinny or fat. A low noise scope with a slow update rate will show a skinny signal initially and, if infinite persistence is enabled, will stay skinny or grow fat if the target signal has noise. A low noise scope with a fast update rate will show the target signal correctly initially and, when infinite persistence is turned on, the waveform will stay the same thickness.
Averaging mode typically makes waveforms thinner by reducing noise. Here, the scope takes successive acquisitions and averages each associated captured point across the acquisition. This technique lowers overall scope noise by averaging out noise values over multiple acquisitions. Tradeoffs include the fact that averaging will also average target signal values and averaging works exclusively on repetitive signals.
High resolution mode can also reduce noise and represent signals more closely. This mode works for repetitive signals and for single shot captures. In high resolution mode, the scope averages adjacent samples and can reduce overall scope noise. One tradeoff here is the scope must average samples, so the resulting averaged sample points occur less frequently than the initial samples were acquired. This reduces effective sample rate and overall bandwidth.
Still wondering if skinny or fat waveforms are better? You are now equipped to choose an oscilloscope that will reproduce more faithfully your target signal's wave shape. Or, if you already have settled on a particular oscilloscope, you can determine how representative the scope's thin or fat wave shapes are of the signal under test.
Joel Woodward is senior oscilloscope product marketing manager with Agilent Technologies.