General Best Practices
Rj/Dj-specific Best Practices
M1 OT-specific Best Practices

General

1. To achieve maximum accuracy, it is important to scale the signal to use as much of the vertical resolution of the scope's A/D converters as possible. Note that the scope's auto-scaling button will in general not scale the signal for best accuracy; rather it is scaled for best human visibility onscreen. M1™ will automatically scale the signal properly for best accuracy as part of AutoMeasure™ during startup.
2. If you wish to vertically scale the signal yourself rather than letting AutoMeasure do it, remember to allow some margin for a noisy signal. i.e. it is better to fill only 90% of the vertical resolution of the scope and capture spurious voltage spikes than to try to fill 99% and have a noisy signal exceed the resolution of the scope. Turning on Infinite Persistence mode on the scope will help you to see if you may be clipping any spikes. Note that this is more difficult to do on LeCroy and Tektronix scopes, as their A/D converters go a full division off both the top and bottom of the screen.
3. Deskewing your scope properly is critical to making accurate Differential and Delay measurements. See ASA App Note 99-13 for details on how to deskew your system.
4. If you have a deep-memory scope (i.e. a scope with greater than 1-2 Mpts maximum acquisition memory on a single channel) and wish to take deep-memory acquisitions, you will see much better performance if your PC has sufficient RAM. As a general rule of thumb, 1 GB RAM is sufficient for acquisition of 12-16 Mpts of memory without ever requiring the use of disk cache. Remember to multiply the number of points per channel by the number of channels being acquired to calculate the total acquisition memory being used.

5. If your signal has a shelf on the transition due e.g. to reflection, this can adversely affect the calculation of threshold crossing times. This is especially true when the shelf is close to the specified threshold level. If the shelf cannot be eliminated but accurate measurements are still required, you should change the threshold at which crossing times are calculated to be either well above or well below the shelf region. See ASA App Note 97-03 for additional information.

6. A channel to channel delay measurement, while useful for discovering trends or repetitive behaviors, will generally (and often significantly) underpredict the worst-case separation that is possible. This is because the true worst-case separation requires the simultaneous occurrence of worst-case events on both signals, which is a very rare condition. M1's PLL Toolkit calculates and displays the true worst-case separation values as well as the measured values to give you a better idea of the true outlier events that you need to be aware of.

Rj/Dj

7. While M1 allows you to choose from a wide variety of MJSQ-approved Rj/Dj extraction algorithms, as well as to specify various parameters of those algorithms, it is strongly recommended that you use the default SEEj™ method. SEEj has been calibrated and validated across a vast jitter space, and ASA considers it to be the most accurate Rj/Dj extraction algorithm available.
8. Each scope manufacturer has their own Rj/Dj extraction algorithm, and each of these algorithms can produce a different (and sometimes very different) result when measuring the same signal. This is caused not only by significant differences in the algorithms, but also by small differences in seemingly identical algorithms. Only M1 allows you to use the exact same algorithm on every real-time scope in your lab, ensuring correlation.

9. M1 provides several ways to track the convergence of your Rj/Dj extraction calculations, including red/yellow/green indicators on the RjDjView™'s tab and a plot of Rj/Dj/Tj vs acquisition. Not acquiring enough data to ensure convergence is a very real pitfall when making Rj/Dj measurements, and ASA is the only vendor who recognizes this fact and helps the user to understand convergence behavior.
10. While a large quantity of data is required to ensure convergence, it is actually preferable to acquire this data in a series of shorter acquisitions rather than one large acquisition. For example, acquiring ten 1Mpt acquisitions will provide the same amount of data as a single 10Mpt acquisition, but the 10-acquisition sequence will allow you to observe the convergence behavior and satisfy yourself that the results have converged while the single long acquisition will not.
11. The other side of Rj/Dj convergence is Rj/Dj divergence. This is when an Rj/Dj result that appears to be converged suddenly jumps in value, usually as a result of non-stationarity in the signal. Note that some Rj/Dj methods, particularly those that are not based on real-time capture, are effectively blind to non-stationarity. M1's Rj/Dj convergence indicators will also alert you if the results begin to diverge.
12. To achieve true correlation in Rj/Dj measurements between different scopes from different vendors, you need to be able to subtract out the effects of the scope's front-end noise on the measurement. From the Preferences / Measurement Preferences dialog, select the Noise tab. You can either enter your own estimated noise values, or let M1 guide you through a procedure to set those values for you.

M1 Oscilloscope Tools

13. ASA's website contains a variety of short videos to help you learn about various aspects of M1 OT™. It is strongly recommended that all new M1 OT users should view the Basic videos as well as any additional videos that are relevant to the work they will be doing.
14. If you don't want AutoMeasure to automatically bring up the default view set when starting up, you can turn that option off in Preferences / Measurement Preferences. You can also tell M1 not to take an initial acquisition, not to optimize the vertical scaling on your signals, or even turn AutoMeasure off entirely if you wish.
15. When measuring a differential signal, using M1 OT's differential mode (with a pair of single-ended probes) and a properly deskewed scope has significant advantages in accuracy and utility over using single-ended mode (with a differential probe).
16. For Repetitive Data signals, it is recommended to enter the exact pattern to match in the Signal Settings dialog rather than simply entering the pattern length in bits. When working with just the pattern length, M1™ will begin with the first edge in the acquisition and count out the desired number of bits. It is possible that this bit position will be in the middle of a pulse, and M1 will then shift to the next available edge to begin the next repetition of the pattern. This will cause the displayed data to be incorrect. Note that it is relatively simple to tell when this occurs, as the values shown will be well outside those expected; e.g. a Peak-Peak that is greater than the Unit Interval.


17. When analyzing deep-memory acquisitions, M1 will require additional time to complete all computations due to the volume of data being processed. There are status indicators next to the Basic View Controls at the bottom of the main viewing area that will help in tracking M1's progress through this process.

18. Turning on Sync Axes ties all of the 'vs Time' views together to provide a time-correlated view of what is happening to all measured parameters at the same time. ScopeTrack™ additionally allows you to view waveform behaviors close up in the ScopeView™ while retaining correlation to your position in the larger record.
19. M1's ScopeView animation runs at a very low priority, so it will not use processor time while M1 is performing other computations.
20. When running M1 inside of a Windows-based oscilloscope, be aware that M1 must share the processor and RAM with the scope application. In some cases, the scope application may be set to demand a large fraction of the processor cycles when it is actively acquiring. This can significantly slow down M1, as M1 is forced to use only the small remaining percentage of processor cycles available. If you think this may be occurring, try setting the scope to Single acquisition mode and seeing if M1 speeds up significantly as a result.

21. When creating a TestScript™, it is helpful to provide information to the user regarding which scope channel(s) the script expects to find each signal on. You should also notify the user when probing and/or signal settings need to be changed.
22. Collaborating with coworkers, customers, and vendors - regardless of where it the world they are - is as easy as clicking on the Send Waveform button in M1's Control Panel. Without leaving M1, you can email or FTP the entire waveform to anyone you choose. At the far end, all they need to do is to open the email attachment and the file will automatically be loaded into their copy of M1 Reader™.
23. When you want to make a major change in direction in M1, there are several options to help you get there quickly. You can close all current measurements, clear all data from memory, or even rerun AutoMeasure and redetect your scope as if you had exited and restarted M1. All three options are available under More Buttons, between the Exit and Help buttons.