“This application note discusses the most common things to test the quality of each design (especially embedded systems). It includes common sources of errors and other factors that can affect or degrade signal quality. It also introduces the techniques and procedures to be followed to achieve good data measurement. More importantly, it can be used as a guide for users to use RIGOL’s high-efficiency mixed-signal oscilloscope collection to deal with embedded design challenges.
“
This application note discusses the most common things to test the quality of each design (especially embedded systems). It includes common sources of errors and other factors that can affect or degrade signal quality. It also introduces the techniques and procedures to be followed to achieve good data measurement. More importantly, it can be used as a guide for users to use RIGOL’s high-efficiency mixed-signal oscilloscope collection to deal with embedded design challenges.
Embedded design, especially design work using low-speed serial signals, is one of the fastest growing areas of digital electronic design. A large number of modules in consumer and industrial electronic products, the demand for communication between FPGAs and processors is increasing at an alarming rate. The use of customized communication protocols and buses is critical to design efficiency and time to market, but there is a risk that it is sometimes difficult to analyze and debug. When using low-speed serial data in embedded applications, the most common sources and types of problems include timing, noise, signal quality, and data. We will recommend debugging techniques and functions available in modern oscilloscopes, which will make debugging these complex systems faster and easier.
Wrong time type
Timing is critical to any serial data system, but it can be difficult to find system timing constraints related to components, transmission length, processing time, and other variables. Let’s start with a simple 16-bit DAC circuit. First, make sure you understand the data and timing specifications of the protocol used. Does it sample data on the clock edge? How far can the clock and data differ when we still expect good data? In other words: have we defined a clock synchronization error budget? Once you understand these timing requirements, you can verify the Tx and Rx hardware subsystems through experiments. Now we can analyze the overall accuracy of system-level timing delays and conversions because we can directly measure logic and analog channels in a time-correlated manner.
Shown above is a simple example that measures a bit on channel 2 (blue) that drives the DAC output to produce a sine wave on channel 1 (yellow). Using parallel bus decoding (Figure 1), we can quickly understand the transition of this single line. But this does not provide us with all the information we need, because the DAC is using many data lines to set its output level.
Obtaining more complete data requires different methods. Let’s move all the DAC lines (Figure 2) to the digital input of the MSO. Now we can see how the digital line is truly coordinated with the DAC output.
For further research, we can simplify the decoding to display the hexadecimal value (Figure 3) and zoom in so that we can view the decoded data. In addition, if we want to see the changes in the bus graphically, we can use a function called plot in the Logic Analyzer bus menu (Figure 4).
In this way, the bit pattern can be displayed graphically for intuitive analysis. This is the perfect choice for use with DAC and A2D, because if something goes wrong with the encoding or decoding scheme, you can get immediate feedback.
noise
One of the most common problems in correct serial data measurement is the handling of system noise. The noise in these measurements can come from a variety of sources, including poor grounding, bandwidth issues, crosstalk, and electromagnetic interference (EMI) issues. Sometimes the problem lies in the equipment, but improved detection and measurement techniques can also significantly improve the results without the need to replace the equipment under test. A good first step is always to ensure that we are using the best measurement method.
We solved these problems in order from easy to difficult. First, we can take a look at our probe connection. Usually, we will use the alligator clip ground strap connected to the probe for grounding. Assuming we did it right, but there are still problems, we may need to switch to ground springs. The location of the ground spring connection is closer to the probe tip, and the connection loop area is significantly reduced. This can significantly improve noise and signal quality (Figure 5), especially for high-speed signals or signals that are sensitive to capacitance or coupling voltage. For these types of measurements, all Rigol probes are equipped with a standard ground strap and ground spring.
If the ground noise problem still exists, try to isolate the device from the ground. The oscilloscope is best grounded to the AC power ground through the plug. If the device under test or the rest of the system can be isolated from ground, ground loops can be eliminated. If there is still a ground noise problem, you can consider using a differential probe such as RP1100D (Figure 6), which can measure without reference to the oscilloscope ground.
Differential measurement may be the only way to clearly view certain low-speed serial data (such as the LVDS bus). Such a bus will deliberately move the reference line to maximize the bandwidth and increase the communication distance, but it may require true differential detection or the use of multiple channels of an oscilloscope to correctly view the signal. Rigol has several different probe types available for these measurements, including the RP1000D series of differential probes (usually used for high-voltage floating applications and RP7150 1.5 GHz differential probes (Figure 7) or high-speed data applications).
Signal quality
Monitoring and improving the quality of low-speed serial signals is a key part of the debugging process. Even if there is no noise, impedance mismatch, bandwidth and load errors will affect the signal quality. Now that we are carefully studying the exact nature of these signals, it is important to verify the way we perform these tests with an oscilloscope. For signal quality testing, we will use analog channels because they can best see what is actually happening in the signal, but we will still perform some decoding. This requires some additional consideration. In order to clearly see the data conversion, we should definitely use the highest possible sampling rate. Since we need to visualize the high-frequency components, sampling at 5 times the bit rate of the digital bus should be taken as a minimum. Sampling at 10 times the bit rate should allow us to see any problems. However, when we decode the signal, the oscilloscope may use a subset of the entire memory data to process the decoding analysis.
This is important because you don’t have to complete the decoding at too high a rate. When using a more nominal receiver to decode the data, this may obscure the problems you will find. On a Rigol oscilloscope, decoding is done on 1 Mpts of memory scattered throughout the acquisition. By setting the storage depth and the time of each division, users can determine whether they want to decode directly from the simulation point or from a subset. Sampling at 10 times the bit rate should allow us to see any problems. However, when we decode the signal, the oscilloscope may use a subset of the entire memory data to process the decoding analysis.
This is important because you don’t have to complete the decoding at too high a rate. When using a more nominal receiver to decode the data, this may obscure the problems you will find. On a Rigol oscilloscope, decoding is done on 1 Mpts of memory scattered throughout the acquisition. By setting the storage depth and the time of each division, users can determine whether they want to decode directly from the simulation point or from a subset. Sampling at 10 times the bit rate should allow us to see any problems. However, when we decode the signal, the oscilloscope may use a subset of the entire memory data to process the decoding analysis. This is important because you don’t have to complete the decoding at too high a rate.
When using a more nominal receiver to decode the data, this may obscure the problems you will find. On a Rigol oscilloscope, decoding is done on 1 Mpts of memory scattered throughout the acquisition. By setting the storage depth and the time of each division, users can determine whether they want to decode directly from the simulation point or from a subset. When using a more nominal receiver to decode the data, this may obscure the problems you will find. On a Rigol oscilloscope, decoding is done on 1 Mpts of memory scattered throughout the acquisition. By setting the storage depth and the time of each division, users can determine whether they want to decode directly from the simulation point or from a subset. When using a more nominal receiver to decode the data, this may obscure the problems you will find. On a Rigol oscilloscope, decoding is done on 1 Mpts of memory scattered throughout the acquisition. By setting the storage depth and the time of each division, users can determine whether they want to decode directly from the simulation point or from a subset.
data
The key to any low-speed serial application is to be able to quickly and easily view the data being transferred. This means that the ability to perform embedded decoding on the oscilloscope has been added. Decoding will affect the trigger and display of the oscilloscope at the same time. It adds the decoded bus display to the Screen of the instrument. You can decode the value as ASCII or hexadecimal, octal or binary data, depending on what you want to view. Now you can also trigger on these values to ensure that you are viewing the packets of most interest.
Look for
The key to proper oversampling and bandwidth
As mentioned above, correct sampling is essential to make correct measurements and fully debug low-speed serial interfaces. A good rule of thumb for analog signals is 5 times the bandwidth of the signal you want to measure. This will limit your rise time error to about 2%. To see the best details about high-frequency signal components, set up your oscilloscope so that the sampling rate is also 5-10 times. When using digital signals, this means sampling 5 times with a bit width. When sampling on digital lines or for decoding, the importance of oversampling is not so important, but please set up the measurement equipment so that it is similar to the final LSS receiver. This gives you the greatest opportunity to focus on substantial errors that may cause problems.
Ground, noise and differential signal
Proper detection and understanding of the use of differential reference signals and ground reference signals is important for debugging. If your data line is not grounded, make sure to understand the impact of ground loops and ground coupled noise on the measurement. Use appropriate probe technology and advanced noise reduction features on the oscilloscope to limit noise sources.If necessary, add differential probes to your measurement system to improve
Measure quality.
How to best view low-speed serial signals
There are many ways to analyze, view and evaluate LSS bus activity on a modern oscilloscope. The best method varies depending on the individual bit transitions that want to look at noise, speed, or synchronization. Do you want to view the complete data packet; or you want to compare the data packet and data packet timing in a longer period of time. Make sure that your benchmarking tool allows you to see everything you need, and is familiar with features such as zoom, recording mode, event table, depth memory, and automatic measurement, and how they interact and how they work when considering test plans Make the best conversion. Ideally, the oscilloscope enables you to view all the results you need and quickly switch modes to get more information.
Concluding remarks
Embedded design and debugging of digital data is an ever-increasing test requirement in a wide range of consumer and industrial applications. Having the right mixed signal oscilloscope can make viewing, analyzing and solving problems including timing, noise, signal quality and data easier and faster. This improves engineering efficiency and time to market. Rigol’s UltraVision-supported oscilloscope series includes mixed signal options from 70 to 500 MHz, as well as the standard or optional functions of the methods and measurements discussed here. It is a powerful benchtop instrument that provides uncompromising value with unprecedented value Performance.
This application note discusses the most common things to test the quality of each design (especially embedded systems). It includes common sources of errors and other factors that can affect or degrade signal quality. It also introduces the techniques and procedures to be followed to achieve good data measurement. More importantly, it can be used as a guide for users to use RIGOL’s high-efficiency mixed-signal oscilloscope collection to deal with embedded design challenges.
Embedded design, especially design work using low-speed serial signals, is one of the fastest growing areas of digital electronic design. A large number of modules in consumer and industrial electronic products, the demand for communication between FPGAs and processors is increasing at an alarming rate. The use of customized communication protocols and buses is critical to design efficiency and time to market, but there is a risk that it is sometimes difficult to analyze and debug. When using low-speed serial data in embedded applications, the most common sources and types of problems include timing, noise, signal quality, and data. We will recommend debugging techniques and functions available in modern oscilloscopes, which will make debugging these complex systems faster and easier.
Wrong time type
Timing is critical to any serial data system, but it can be difficult to find system timing constraints related to components, transfer length, processing time, and other variables. Let’s start with a simple 16-bit DAC circuit. First, make sure you understand the data and timing specifications of the protocol used. Does it sample data on the clock edge? How far can the clock and data differ when we still expect good data? In other words: have we defined a clock synchronization error budget? Once you understand these timing requirements, you can verify the Tx and Rx hardware subsystems through experiments. Now we can analyze the overall accuracy of system-level timing delays and conversions because we can directly measure logic and analog channels in a time-correlated manner.
Shown above is a simple example that measures a bit on channel 2 (blue) that drives the DAC output to produce a sine wave on channel 1 (yellow). Using parallel bus decoding (Figure 1), we can quickly understand the transition of this single line. But this does not provide us with all the information we need, because the DAC is using many data lines to set its output level.
Different methods are needed to obtain more complete data. Let’s move all the DAC lines (Figure 2) to the digital input of the MSO. Now we can see how the digital line is truly coordinated with the DAC output.
For further research, we can simplify the decoding to display the hexadecimal value (Figure 3) and zoom in so that we can view the decoded data. In addition, if we want to see the changes in the bus graphically, we can use a function called plot in the Logic Analyzer bus menu (Figure 4).
In this way, the bit pattern can be displayed graphically for intuitive analysis. This is the perfect choice for use with DAC and A2D, because if something goes wrong with the encoding or decoding scheme, you can get immediate feedback.
noise
One of the most common problems in correct serial data measurement is the handling of system noise. The noise in these measurements can come from a variety of sources, including poor grounding, bandwidth issues, crosstalk, and electromagnetic interference (EMI) issues. Sometimes the problem lies in the equipment, but improved detection and measurement techniques can also significantly improve the results without the need to replace the device under test. A good first step is always to ensure that we are using the best measurement method.
We solved these problems in order from easy to difficult. First, we can take a look at our probe connection. Usually, we will use the alligator clip ground strap connected to the probe for grounding. Assuming we did it right, but there are still problems, we may need to switch to ground springs. The location of the ground spring connection is closer to the probe tip, and the connection loop area is significantly reduced. This can significantly improve noise and signal quality (Figure 5), especially for high-speed signals or signals that are sensitive to capacitance or coupling voltage. For these types of measurements, all Rigol probes are equipped with a standard ground strap and ground spring.
If the ground noise problem still exists, try to isolate the device from the ground. The oscilloscope is best grounded to the AC power ground through the plug. If the device under test or the rest of the system can be isolated from ground, ground loops can be eliminated. If there is still a ground noise problem, you can consider using a differential probe such as RP1100D (Figure 6), which can measure without referring to the ground of the oscilloscope.
Differential measurement may be the only way to clearly view certain low-speed serial data (such as the LVDS bus). Such a bus will deliberately move the reference line to maximize the bandwidth and increase the communication distance, but it may require true differential detection or the use of multiple channels of an oscilloscope to correctly view the signal. Rigol has several different probe types available for these measurements, including the RP1000D series of differential probes (usually used for high-voltage floating applications and RP7150 1.5 GHz differential probes (Figure 7) or high-speed data applications).
Signal quality
Monitoring and improving the quality of low-speed serial signals is a key part of the debugging process. Even if there is no noise, impedance mismatch, bandwidth and load errors will affect the signal quality. Now that we are carefully studying the exact nature of these signals, it is important to verify the way we perform these tests with an oscilloscope. For signal quality testing, we will use analog channels because they can best see what is actually happening in the signal, but we will still perform some decoding. This requires some additional considerations. In order to clearly see the data conversion, we should definitely use the highest possible sampling rate. Since we need to visualize the high-frequency components, sampling at 5 times the bit rate of the digital bus should be taken as a minimum. Sampling at 10 times the bit rate should allow us to see any problems. However, when we decode the signal, the oscilloscope may use a subset of the entire memory data to process the decoding analysis.
This is important because you don’t have to complete the decoding at too high a rate. When using a more nominal receiver to decode the data, this may obscure the problems you will find. On a Rigol oscilloscope, decoding is done on 1 Mpts of memory scattered throughout the acquisition. By setting the storage depth and the time of each division, users can determine whether they want to decode directly from the simulation point or from a subset. Sampling at 10 times the bit rate should allow us to see any problems. However, when we decode the signal, the oscilloscope may use a subset of the entire memory data to process the decoding analysis.
This is important because you don’t have to complete the decoding at too high a rate. When using a more nominal receiver to decode the data, this may obscure the problems you will find. On a Rigol oscilloscope, decoding is done on 1 Mpts of memory scattered throughout the acquisition. By setting the storage depth and the time of each division, users can determine whether they want to decode directly from the simulation point or from a subset. Sampling at 10 times the bit rate should allow us to see any problems. However, when we decode the signal, the oscilloscope may use a subset of the entire memory data to process the decoding analysis. This is important because you don’t have to complete the decoding at too high a rate.
When using a more nominal receiver to decode the data, this may obscure the problems you will find. On a Rigol oscilloscope, decoding is done on 1 Mpts of memory scattered throughout the acquisition. By setting the storage depth and the time of each division, users can determine whether they want to decode directly from the simulation point or from a subset. When using a more nominal receiver to decode the data, this may obscure the problems you will find. On a Rigol oscilloscope, decoding is done on 1 Mpts of memory scattered throughout the acquisition. By setting the storage depth and the time of each division, users can determine whether they want to decode directly from the simulation point or from a subset. When using a more nominal receiver to decode the data, this may obscure the problems you will find. On a Rigol oscilloscope, decoding is done on 1 Mpts of memory scattered throughout the acquisition. By setting the storage depth and the time of each division, users can determine whether they want to decode directly from the simulation point or from a subset.
data
The key to any low-speed serial application is to be able to quickly and easily view the data being transferred. This means that the ability to perform embedded decoding on the oscilloscope has been added. Decoding will affect the trigger and display of the oscilloscope at the same time. It adds the decoded bus display to the screen of the instrument. You can decode the value as ASCII or hexadecimal, octal or binary data, depending on what you want to view. Now you can also trigger on these values to ensure that you are viewing the packets of most interest.
Look for
The key to proper oversampling and bandwidth
As mentioned above, correct sampling is essential to make correct measurements and fully debug low-speed serial interfaces. A good rule of thumb for analog signals is 5 times the bandwidth of the signal you want to measure. This will limit your rise time error to about 2%. To see the best details about high-frequency signal components, set up your oscilloscope so that the sampling rate is also 5-10 times. When using digital signals, this means sampling 5 times with a bit width. When sampling on digital lines or for decoding, the importance of oversampling is not so important, but please set up the measurement equipment so that it is similar to the final LSS receiver. This gives you the greatest opportunity to focus on substantive errors that may cause problems.
Ground, noise and differential signal
Proper detection and understanding of the use of differential reference signals and ground reference signals is important for debugging. If your data line is not grounded, make sure to understand the impact of ground loops and ground coupled noise on the measurement. Use appropriate probe technology and advanced noise reduction features on the oscilloscope to limit noise sources.If necessary, add differential probes to your measurement system to improve
Measure quality.
How to best view low-speed serial signals
There are many ways to analyze, view and evaluate LSS bus activity on a modern oscilloscope. The best method varies depending on the individual bit transitions that want to look at noise, speed, or synchronization. Do you want to view the complete data packet; or you want to compare the data packet and data packet timing in a longer period of time. Make sure that your benchmarking tool allows you to see everything you need, and is familiar with features such as zoom, recording mode, event table, depth memory, and automatic measurement, and how they interact and how they work when considering test plans Make the best conversion. Ideally, the oscilloscope enables you to view all the results you need and quickly switch modes to get more information.
Concluding remarks
Embedded design and debugging of digital data is an ever-increasing test requirement in a wide range of consumer and industrial applications. Having the right mixed signal oscilloscope can make viewing, analyzing and solving problems including timing, noise, signal quality and data easier and faster. This improves engineering efficiency and time to market. Rigol’s UltraVision-supported oscilloscope series includes mixed signal options from 70 to 500 MHz, as well as the standard or optional functions of the methods and measurements discussed here. It is a powerful benchtop instrument that provides uncompromising value with unprecedented value Performance.
The Links: LQ181E1LW31 LM64P844