Design Article
Oscilloscope memory depth: when bigger is not always better
Richard Markley, Agilent Technologies
3/7/2012 1:55 PM EST
Example
Let’s look at an example where segmented memory might be advantageous. In Figure 3 you can see two radar bursts separated by a long period of idle time in between. In a traditional deep memory oscilloscope, we are digitizing the bursts and the idle time. As you can see in Figure 3, the sample rate for the scope (which typically samples at 5GS/s) is only 625MSa/s, and that is with us just capturing two of the pulses! What would happen if we wanted to capture 100 of the pulses? The sample rate would drop to less than 10MS/s and the pulses would no longer be identifiable because they are severely under sampled.
If we wanted to capture those 100 pulses and all the idle time in between them at a sample rate of 5GS/s, we would need an oscilloscope with 2.5 gigapoints of memory (2,500,000,000). No scope on the market today offers that deep of memory. With segmented memory, we are able to digitize just the portion of the waveform that we care about (the burst itself) and ignore all of the idle time in between bursts. Figure 4 shows the first of 100 RF bursts captured using segmented memory. Note the sample rate was 5GS/s, and each segment was time stamped so you know exactly when it happened in relation to the initial trigger. Figure 5 shows the 100th burst and its time stamp (396.001ms). The oscilloscope allows you to walk through each of the segments and analyze them (including decoding of each segment’s packets if you were using segmented memory with a serial bus).
In the end, it often pays to look beyond the banner spec a manufacturer presents in the datasheet. While a datasheet with a large number for acquisition memory can be tempting, you should definitely consider how you will be using the scope. In some cases, the deepest memory possible will be the best option. But in many cases, a scope designed to handle deep memory is going to be a better option and create less frustration from sluggishness or odd operating modes.
What is update rate?
Update rate (sometimes called “dead time”) is how fast an oscilloscope can trigger, process the data it has captured, and then display it to the screen of the oscilloscope. The faster the update rate, or the shorter the dead time, the more likely you are to catch an infrequent event. People often associate fast update rates with analog oscilloscopes from years ago. Fortunately, new oscilloscope architectures allow for even faster update rates than the fastest analog scopes of yesterday.
Let’s look at an example where segmented memory might be advantageous. In Figure 3 you can see two radar bursts separated by a long period of idle time in between. In a traditional deep memory oscilloscope, we are digitizing the bursts and the idle time. As you can see in Figure 3, the sample rate for the scope (which typically samples at 5GS/s) is only 625MSa/s, and that is with us just capturing two of the pulses! What would happen if we wanted to capture 100 of the pulses? The sample rate would drop to less than 10MS/s and the pulses would no longer be identifiable because they are severely under sampled.
If we wanted to capture those 100 pulses and all the idle time in between them at a sample rate of 5GS/s, we would need an oscilloscope with 2.5 gigapoints of memory (2,500,000,000). No scope on the market today offers that deep of memory. With segmented memory, we are able to digitize just the portion of the waveform that we care about (the burst itself) and ignore all of the idle time in between bursts. Figure 4 shows the first of 100 RF bursts captured using segmented memory. Note the sample rate was 5GS/s, and each segment was time stamped so you know exactly when it happened in relation to the initial trigger. Figure 5 shows the 100th burst and its time stamp (396.001ms). The oscilloscope allows you to walk through each of the segments and analyze them (including decoding of each segment’s packets if you were using segmented memory with a serial bus).
Figure 3: Two RF pulses spread out in time. Notice the lower sample rate due to the oscilloscope digitizing the pulses and the idle time between them.
Figure 4: First of 100 RF pulses captured using segmented memory. Note the 5GS/s sample rate.
Figure 5: Pulse 100 of 100 RF pulses captured using segmented memory. Note the 5GS/s sample rate and the time stamp (396.001ms).
In the end, it often pays to look beyond the banner spec a manufacturer presents in the datasheet. While a datasheet with a large number for acquisition memory can be tempting, you should definitely consider how you will be using the scope. In some cases, the deepest memory possible will be the best option. But in many cases, a scope designed to handle deep memory is going to be a better option and create less frustration from sluggishness or odd operating modes.
What is update rate?
Update rate (sometimes called “dead time”) is how fast an oscilloscope can trigger, process the data it has captured, and then display it to the screen of the oscilloscope. The faster the update rate, or the shorter the dead time, the more likely you are to catch an infrequent event. People often associate fast update rates with analog oscilloscopes from years ago. Fortunately, new oscilloscope architectures allow for even faster update rates than the fastest analog scopes of yesterday.
Figure 6: Oscilloscope dead time can hide rare events. A faster update rate (the inverse of dead time) can help increase your chances of seeing those infrequent events.
About the Author:
In 2001 Richard joined Agilent Technologies in Colorado Springs, Colorado as a technical support engineer for Intel Front Side Bus Logic Analyzer solutions. Since then, he has held individual contributor and management positions in Logic Analyzers and Oscilloscopes focused on business development, product marketing and product planning.
Richard is currently a Strategic Product Planner for Agilent’s High Volume Oscilloscopes, including the 2000 and 3000 X-Series. His responsibilities include championing customer requirements, developing sustainable competitive advantages and partnering with R&D to optimize Agilent’s roadmap and strategy in the High Volume Oscilloscope segment. Prior to being product planner for the 2000/3000 X-Series, Richard was product planner for the Infiniium 9000 Series.
Richard holds a bachelor’s degree in Electrical Engineering from Kansas State University and a Masters of Business Administration from the University of Colorado. When not studying the latest technology, Richard enjoys automotive auto crossing, playing golf and spending time with his family.
In 2001 Richard joined Agilent Technologies in Colorado Springs, Colorado as a technical support engineer for Intel Front Side Bus Logic Analyzer solutions. Since then, he has held individual contributor and management positions in Logic Analyzers and Oscilloscopes focused on business development, product marketing and product planning.
Richard is currently a Strategic Product Planner for Agilent’s High Volume Oscilloscopes, including the 2000 and 3000 X-Series. His responsibilities include championing customer requirements, developing sustainable competitive advantages and partnering with R&D to optimize Agilent’s roadmap and strategy in the High Volume Oscilloscope segment. Prior to being product planner for the 2000/3000 X-Series, Richard was product planner for the Infiniium 9000 Series.
Richard holds a bachelor’s degree in Electrical Engineering from Kansas State University and a Masters of Business Administration from the University of Colorado. When not studying the latest technology, Richard enjoys automotive auto crossing, playing golf and spending time with his family.
|
|
|
|
|
Navigate to related information
Frank Eory
3/7/2012 5:07 PM EST
Thanks for this explanation of how sampling scopes work and how architecture affects performance. Very informative.
Sign in to Reply
http://www.lulu.com/spotlight/poconoarmchairreview
3/7/2012 11:43 PM EST
This would make a cool video.
Sign in to Reply
t.alex
3/8/2012 10:16 AM EST
In fact Agilent scopes have been my favourite for long time.
Sign in to Reply
agk
3/9/2012 1:21 AM EST
I used two scopes for the same application with 1 megabytes and 10 megabytes memory while making a buying decision for about 10 numbers of scope from two popular companies. The input to the scopes were the composite video signal from the DTV set top box. i found that the scope with 10 mb memory performed very well. I was able to see the Vertical and horizontal blanking periods easily by using the delay time base.The larger memory scope i was able to analyze the all the sync pulses,color burst,front porch and back porch.Any way the larger memory is always helpful in conjunction with the higher sampling rates.
Sign in to Reply
Wing traces in the sky
3/14/2012 1:48 AM EDT
What's the waveform update rate when you set the scope memory depth or record length to 10M ?
The maximum memory depth of scope available today is 2Gpts per channel,which is good for single shot analysis , but not for fast waveform update rate .
Sign in to Reply
docdivakar
3/14/2012 8:34 AM EDT
Thanks for an informative article.
@Rich Krajewski: good call, it would be nice to see a Youtube video version of this!
MP Divakar
Sign in to Reply
Michael.Lauterbach
3/21/2012 11:34 AM EDT
To those of us who know oscilloscope architectures this piece is a thinly veiled attack by Agilent on Tektronix. I suppose from the writer's point of view that is worthwhile. But he ignores a few major points:
1. Most scopes these days have a fast update mode for when the user simply wants to view the data.
2. In some cases viewing the signal is enough but very often the user needs measurements. In this case the waveform needs to go to the CPU. The LeCroy X-Stream (extremely fast, streaming) architechture makes measurements much faster than other architectures.
Sign in to Reply