Article count:1393 Read by:2254602

Account Entry

[Prize] 2020 MXR mid-range oscilloscope "Cloud Marathon" - the second stop!

Latest update time:2020-05-20
    Reads:


Click on the mini program link below.


Enter the MXR mid-range oscilloscope and invite you to the cloud marathon - Beijing Station



*Note: The third prize of the Beijing Cloud Marathon event has been changed to " Digital Design Package Silk Screen Ruler "


Comprehensive measurement of real-time oscilloscope vertical system performance and sharing of MXR test results





First, let's take a look at the typical reference architecture of a modern digital real-time oscilloscope.


Figure 1. Typical reference architecture of a modern digital real-time oscilloscope


In the typical reference architecture diagram above, we can see the key indicators often mentioned in the industry, such as the bandwidth of the analog front-end circuit , the sampling rate of the sampling circuit , the number of ADC bits , and the four most basic indicators of digital real-time oscilloscopes, such as triggering and memory depth .



In this issue, we will once again start from the basic bandwidth indicators to discuss the evolution of oscilloscope bandwidth-related concepts and the bandwidth and analog front-end performance of Keysight's latest mid-range oscilloscope MXR.



Take another look at the following mind map about oscilloscope analog bandwidth:


Figure 2: Real-time oscilloscope bandwidth mind map


For a digital signal measurement system based on an oscilloscope, there are generally three bandwidth definitions before the signal is digitized, i.e., before ADC sampling: the bandwidth of the oscilloscope itself, the bandwidth of the probe, and the bandwidth of the system. In this article, we mainly focus on the bandwidth of the oscilloscope itself.


In addition, the calculation method of bandwidth under different frequency response modes of oscilloscopes is slightly different. The most traditional frequency response mode of oscilloscopes is Gaussian frequency response. The frequency response mode of oscilloscopes with Gaussian frequency response is similar to that of low-pass Gaussian filters. The disadvantages of oscilloscopes with Gaussian frequency response are mainly reflected in the large in-band loss and insufficient out-of-band suppression ability.


The frequency response of early oscilloscopes and many current low- and medium-bandwidth oscilloscopes is Gaussian or quasi-Gaussian. Many modern high-bandwidth oscilloscopes use another frequency response, namely flat frequency response. The main reason for this differentiation is the cost of configuring high-sampling circuits on high-bandwidth oscilloscopes. The two frequency response curves are shown in the figure below.


Figure 3 Oscilloscope frequency response



Next, let’s take a look at the bandwidth we talked about today.


1. Oscilloscope Bandwidth


The definition of oscilloscope bandwidth is: input a sine wave signal, keep the amplitude unchanged, and increase the signal frequency. When the signal amplitude displayed on the oscilloscope is 70.7% of the actual signal amplitude (i.e. 3dB attenuation), the corresponding signal frequency is equal to the oscilloscope bandwidth.


The MXR series offers a wide range of bandwidth options, from 500MHz to 6GHz. The entire series supports on-site software upgrades from low-bandwidth models to higher-bandwidth models. At each channel input, there is a low-noise front-end custom chip, which is a 130ns BiCMOS chip with an internally integrated analog filter, so customers can choose the bandwidth of the filter (the oscilloscope also supports digital filters, which are optional, including brick-wall frequency response filters, 4th-order Bessel filters, and bandpass filters, which also enables on-site bandwidth software upgrades.)


It should be noted here that when using a brick-wall frequency response filter, its bandwidth can be set to the maximum bandwidth of the oscilloscope. When using a 4th-order Bessel filter, its bandwidth can be set to about 2/3 of the maximum bandwidth of the oscilloscope. The bandpass filter is designed for phase noise analysis applications and is not usually used.


Xiao K used an MXR608A engineering prototype with a bandwidth of 6GHz to perform a frequency sweep test. He input a sine signal with an amplitude of 500mV, starting from a frequency of 1GHz and increasing the signal frequency to 6GHz. The measurement results are shown in the table below.


Figure 4 MXR bandwidth measurement environment


For a Gaussian frequency response oscilloscope with a bandwidth of 6GHz, if a sine wave signal of about 1GHz is input and the amplitude is 500mV, the measured amplitude value should theoretically be 490 mV, and for a sine wave signal of 2GHz, the measured amplitude value should theoretically be about 470mV. However, the test results of the prototype of this project show that the amplitude value is 497.4mV at 2GHz. After removing the influence of the connecting cable and the connector, the oscilloscope measures a 2GHz sine wave signal with almost no attenuation. At the 3GHz frequency point, the measured signal amplitude is also close to 500mV, which is far superior to the Gaussian response product. For the measurement results of the sine wave signal at the 4GHz frequency point, if the attenuation of the probe and the connecting line is removed, the amplitude error is about 10%. All these show that MXR uses a flat response.


If we calculate based on the 3dB attenuation bandwidth of the oscilloscope, we can preliminarily calculate that the amplitude corresponding to 6GHz is 353.5mV. From the measurement results, the MXR608A oscilloscope with a bandwidth of 6GHz, even if the amplitude of the 6GHz sine wave signal is only attenuated by 20%, it does not decrease by 3dB. This also shows that the MXR mid-range oscilloscope is not a Gaussian frequency response oscilloscope, but a flat response oscilloscope, so it can accurately measure signals and components of all frequencies within its bandwidth. In fact, the bandwidth of the 6GHz model of the MXR series products can be set to 6.3GHz at most, with an additional 300MHz !


The above only provides amplitude results. The horizontal frequency and period measurement results are very accurate even at 6 GHz.


Conclusion: As an oscilloscope with a flat response, the MXR has excellent characterization capabilities for signals within its bandwidth, providing strong support for various precise measurements including consistency testing and digital receiver applications.



2. Empirical formula or rule for selecting oscilloscope bandwidth


1. Calculate the required oscilloscope bandwidth using the measured sine wave frequency


If the measured signal is a sine wave, the required oscilloscope bandwidth is 3 times its frequency, or the required oscilloscope's own rise time should be 3 times faster than the measured sine wave signal. Since the frequency response of early oscilloscopes, whether analog or digital, is Gaussian, the definition of the oscilloscope's own bandwidth is 0.35/Tr, where Tr is the oscilloscope's own rise time (10% to 90%). The oscilloscope's own rise time is 3 times faster than the object being measured, so people have used this formula for a long time, oscilloscope bandwidth = 0.35/Tr x 3, where Tr refers to the rise time of the measured signal (10% to 90%). This statement assumes the premise that the oscilloscope has a Gaussian frequency response and the measured signal is a sine wave. This statement does not apply to modern high-end real-time oscilloscopes, because their frequency response is not Gaussian, but mostly maximum flatness, or between Gaussian and maximum flatness.


Today, based on the MXR series oscilloscopes with flat response, the amplitude error of the 3GHz sine wave signal component is basically negligible. For the 5GHz sine wave, the amplitude error is about 10%, and for the 6GHz, the amplitude error is 20%. If the instrument performance is known, the MXR can be fully used for 6GHz sine wave signal measurement, which means that the 6GHz bandwidth performance of the MXR is quite impressive.


2. Calculate the required oscilloscope bandwidth using the square wave or clock signal frequency of the signal being measured


If the measured signal is a square wave or clock signal, people believe that the oscilloscope bandwidth should be able to observe at least the 5th harmonic of the measured signal. This statement assumes that the 5th harmonic energy of the measured square wave signal is still above the oscilloscope noise floor, while the 7th harmonic energy is submerged in the oscilloscope noise floor. The question is, how to determine this point, for example, the signal rate of PCI-E 3.0 is 8Gbps, the fundamental wave is 4GHz, and the fifth harmonic is 20GHz, so does the PCI-E 3.0 test necessarily require a 20GHz bandwidth?


The answer is that if the energy of the 5th harmonic is very small and falls below the background noise of the oscilloscope you are using, you still cannot observe the 5th harmonic with a 20GHz bandwidth oscilloscope. Therefore, as early as around 2008, the PCI-Sig Association recommended a bandwidth of 13GHz for the consistency test of PCIE3.0, because the 20GHz oscilloscope products in the industry at that time had a large background noise and could not characterize the 5th harmonic at all.


The MXR series mid-range oscilloscopes have a bandwidth of 500MHz-6GHz and provide consistency testing for all the current mainstream medium signal rate standards in the industry, including (but not limited to):



In addition to the above standard consistency tests, the MXR 6GHz model can also support signals with rates below 4Gbps, such as HDMI1.4b, PCIE1.x, SATA 1.5G/3.0Gbps, and Rapid IO.


It should be noted that for DDR3/LPDDR3, 1866 MT/s, the clock frequency is 933MHz (also the fastest rate of data bus conversion), so the MXR 6GHz (hidden bandwidth to 6.3GHz) can collect the 7th harmonic, so it can basically support this rate.


3. Use the fastest rise time of the signal being measured to determine the required oscilloscope bandwidth


This is an accurate formula, which is applicable to more occasions because it does not have any specific assumptions. This formula comes from Howard W Johnson's book "High Speed ​​Digital Design --- A handbook of Black Magic".


Assuming you can accept a 3% measurement error, for an oscilloscope with a flat response, BWscope = 0.5/Tr x 1.4 (the measured signal gives a rise time of 10%~90%), BWscope = 0.4/Tr x1.4 (if the given rise time is 20%~80%). For an oscilloscope with a Gaussian frequency response, BWscope = 0.5/Tr x 2 (the measured signal gives a rise time of 10%~90%), BWscope = 0.4/Tr x2 (if the given rise time is 20%~80%).


Figure 5 MXR series model rise time


Based on the flat response, MXR can measure the rising edge of the signal:

  • Trisetime (10-90%) = 0.5/BWscope X 1.4 = 0.5/6GHz * 1.4=116 ps

  • Trisetime (20-80%) = 0.4/BWscope X 1.4 = 0.4/6GHz * 1.4=93 ps


In actual measurement, the signals to be measured with a rise time greater than the above can be well characterized.



Above, we have made a brief introduction to the bandwidth and actual measurement of digital real-time oscilloscopes and the latest MXR.




Tips: The relationship between bandwidth and rise time





The rise time specification of the oscilloscope is also closely related to the bandwidth specification of the oscilloscope. For an oscilloscope with a Gaussian frequency response, the rise time is about 0.35 / BWscope according to the 10% to 90% standard. For an oscilloscope with a maximum flat response, the rise time specification range is usually around 0.4 / BWscope, depending on the sharpness of the frequency drop feature.

Keep in mind that the oscilloscope's rise time is not the fastest edge speed that the oscilloscope can accurately measure. Assuming the input signal has a theoretically infinitely fast rise time (0 ps), the oscilloscope's rise time is the fastest edge speed that the oscilloscope can produce. Although this theoretical specification is not measurable because pulse generators cannot actually generate infinitely fast edges, the oscilloscope's rise time can be measured by inputting a pulse signal with an edge speed 3 to 5 times faster than the oscilloscope's rise time specification.



3. Other indicators of vertical system that are often overlooked


In fact, mid-range oscilloscopes have become a fully competitive red ocean market, with multiple brands and nearly 10 product series often making engineers feel dazzled and at a loss. Therefore, a simple bandwidth comparison can no longer fully characterize the differences between different products. A more comprehensive examination of the overall performance of the oscilloscope's vertical system is required, which of course involves more indicators. For example, the oscilloscope's background noise and channel input amplitude etc.


The background noise of the oscilloscope mainly comes from the analog front end of the oscilloscope. The industry generally defines it as Gaussian white noise. The wider the bandwidth, the greater the noise. Therefore, in actual work, in addition to adopting bandwidth restrictions for different signal tests, it is particularly important to choose a low-background noise instrument. For the significance of low background noise for eye diagram testing, please refer to the previous article "A Brief Discussion on the Significance of Low Background Noise of Oscilloscopes for High-Speed ​​Eye Diagram Testing" .


The latest MXR series inherits the excellent front end of the S series oscilloscope and provides the lowest background noise in the industry. At 6 GHz bandwidth 50mV/Div, the background noise is only 971uV , which is 80% of other similar products in the industry, providing the highest measurement accuracy for consistency and eye diagram testing. For power ripple testing, the N7020A 2GHz bandwidth power ripple probe is configured with 1:1 attenuation. When set to 2mV/Div, it provides 10mV level power ripple and noise measurement, and the background noise is only 91uV.


Low noise floor is also critical for many other small signal and RF signal tests, so I will not give too many examples.


The above is about the background noise of the oscilloscope, which involves measurement accuracy. Another indicator that is often overlooked is the maximum allowable input amplitude of the oscilloscope analog channel , which involves instrument safety and is crucial to the safe use of the instrument.


Since many consistency tests and digital receiver applications today often use cables to connect directly to the oscilloscope , the oscilloscope is basically "naked" when facing abnormal signal jumps, static electricity, and external EMI/EMC interference. Calm down and think about it. Have you ever encountered an oscilloscope's analog channel being damaged unknowingly, and then wasted time and effort to send it for repair?


Keysight's high-end InP V/Z/UXR series oscilloscopes are the only high-end products in the industry that support ±4V input in all gears.


The latest MXR mid-range oscilloscope has a maximum input voltage range of ±5V at 50Ω:


Figure 6: Maximum allowable input amplitude of MXR mid-range oscilloscope


This indicator itself should not be a technical bottleneck for mid-range products, but in fact there are differences between different products. One of the reasons is that due to bandwidth or noise reduction, the attenuator with sufficient performance is not used, thus sacrificing the input amplitude allowed by the analog channel. In contrast, all real-time oscilloscopes of Keysight provide the best and balanced indicators.


There are some other indicators related to the vertical system, such as vertical scale sensitivity, offset setting, etc. At present, there is not much difference between the products of various manufacturers, so they will not be explained in detail here.


Figure 7 MXR608A front view





The above is all the content about MXR in this issue. If you have any different insights or stories about the bandwidth or vertical system of the oscilloscope, please share them in the comment area.

Finally, let me make a hard advertisement and come to participate in the 2020 MXR mid-range oscilloscope "Cloud Marathon" - the second stop !





Related reading



references


1. "Principles and Applications of Digital Oscilloscopes", Sun Dengliang, Shanghai Jiaotong University Press, 2012.

2. Advanced Applications of Modern Oscilloscopes: Testing and Usage Skills, Li Kai, Tsinghua University Press, 2017.

3. "A Brief Discussion on the Significance of Low Oscilloscope Noise Floor for High-Speed ​​Eye Diagram Testing", Keysight Technologies (China) Co., Ltd., Huang Teng

4.《High Speed Digital Design --- A handbook of Black Magic》,Howard W Johnson, Prentice Hall, 2003.



About Keysight Technologies

We are committed to helping enterprises, service providers and government customers accelerate innovation and create a secure and connected world. Since the founding of HP in 1939, Keysight Technologies has been operating independently as a new electronic test and measurement company on November 1, 2014. We continue to uphold the same entrepreneurial spirit and passion to start a new journey, inspire global innovators and help them achieve goals beyond imagination. Our solutions are designed to help customers innovate in 5G, automotive, IoT, network security and other fields.

Learn about our history of relentless pursuit of industry innovation:

www.keysight.com/cn...


Long press to identify the QR code above

Follow Keysight Technologies official service account


Latest articles about

 
EEWorld WeChat Subscription

 
EEWorld WeChat Service Number

 
AutoDevelopers

About Us Customer Service Contact Information Datasheet Sitemap LatestNews

Room 1530, Zhongguancun MOOC Times Building,Block B, 18 Zhongguancun Street, Haidian District,Beijing, China Tel:(010)82350740 Postcode:100190

Copyright © 2005-2024 EEWORLD.com.cn, Inc. All rights reserved 京ICP证060456号 京ICP备10001474号-1 电信业务审批[2006]字第258号函 京公网安备 11010802033920号