General idea
A mask is a set of polygons on the screen. The signal must stay outside the forbidden zones. The oscilloscope compares each acquired waveform with the mask and counts the violations. This automatically and repeatably evaluates what a technician would do by looking at the screen.
Where it is used
- Eye diagrams of serial links, where communication standards publish masks used to verify signal quality.
- Production testing: verifying that a signal stays within an acceptable envelope.
- Stability monitoring: detecting occasional events such as stray pulses or glitches.
How it is done
- Obtain the mask: load the one defined by a standard or create it from a good reference waveform, widening it by a horizontal and vertical tolerance.
- Set up the acquisition: stable triggering, suitable vertical and horizontal scales, and alignment of the signal with the mask.
- Define the termination criterion: a number of waveforms, a time or a number of violations.
- Run the test and observe the violation count and failure rate.
- Review the violations, which the instrument can highlight or save for analysis.
What influences the result
- The bandwidth of the oscilloscope and probe, which changes the visible edges and noise.
- The probe loading on the circuit, which modifies the signal.
- The stability of the trigger or clock recovery, which shifts the signal relative to the mask.
- The instrument’s own noise, which can cause violations that do not exist in the signal.
When it does not apply
A mask test judges the shape of the signal, not its cause. A failure indicates that jitter, noise, reflections or amplitude must be investigated with other tools. And a result with no violations only guarantees what was observed in the acquired waveforms: a very rare event may not have been captured, so the number of waveforms must be consistent with the error rate you are trying to detect.