The Basics of Verifying Communications Signals With an Oscilloscope

At one time, communications signals were found only in specific telecommunications applications. Now they are common occurrences in most consumer and business products. Almost every major piece of electronic equipment¾ faxes, telephones, and computers¾ is networked, and it seems as if everyone owns a wireless communications device like a digital Personal Communications Service (PCS) or cellular phone.

With the spread of communications circuitry into mainstream products, design and test engineers now must characterize and verify communications signals alongside the usual digital and analog signals. Since these products must be designed quickly and verifying communications signals is a relatively new task for many engineers, it is important to have a test solution, such as an oscilloscope, that is easy to learn and use.

Basically, the challenge of verifying communications signals with an oscilloscope can be divided into three tasks:

Connecting to the product.

Acquiring useful data.

Characterizing the design with specialized measurements.

Getting a Good Connection

Connecting to the product is the first challenge. Oscilloscope inputs typically are BNC connectors terminated with 50-W or 1-MW impedances. Some communications signals, such as DS3, use 75-W coaxial connectors. But others, such as the DS1 signal, use a bantam connector with a 100-W impedance.

Also, oscilloscope inputs are referenced to ground, while many telecommunications signals are differential voltages. To preserve the integrity of the acquired signals, use the appropriate adapter that converts balanced signals to single-ended—and match the oscilloscope’s impedance to the impedance of the application¾ 75 W , 100 W , 110 W , or 120 W .

High-speed electrical communications signals create the biggest connection challenge because they are differential. Probe bandwidth is very important, where data rates can exceed a gigabit per second, translating into 500-plus MHz signals with edge rates at much higher frequencies.

Applications in this area center on high-speed networking solutions based on the Fibre-Channel standards and Ethernet as it evolves in support of higher data rates. These advanced networks carry huge amounts of information, especially video data, in real time.

A differential probe is a must for these high-end applications and must deliver outstanding bandwidth and common-mode rejection performance. With an advanced differential probe, you not only will capture the true shape of the pulse, but also uncover the true noise entering the differential lines due to crosstalk.

Although electrical signals are being used in high-speed data transmission, many of the faster networking solutions are based on fiber optics governed by a set of optical standards. This approach provides high-speed transmission over longer distances than electrical solutions. But testing fiber-optic transmission designs opens new measurement issues.

Here the test setup becomes a set of instruments consisting of an oscilloscope, an optical-to-electrical converter, and if required, an optical reference receiver filter. The optical-to-electrical converter changes the received optical power into a voltage.

The reference receiver filter is an electrical low-pass filter, specified by optical communications standards such as SONET and Fibre Channel. This filter provides a test setup with a standardized frequency response for more consistent measurements because the reference receiver filter reduces the effects of overshoot and noise.

A variety of these interconnect devices is available. The more advanced devices not only connect to and condition the signal, but also communicate with the oscilloscope to automatically adjust the displayed scale factors and units. These powerful capabilities eliminate errors in interpreting displays and simplify documentation tasks.

Acquiring Useful Data

Once a reliable connection is made, now you must track down the pertinent information in a veritable ocean of data. Since telecommunications signals are highly unique and governed by detailed specifications, oscilloscopes designed for the communications industry include a triggering capability tailored to these standards. This advanced triggering makes it much easier to zero-in on subtle problems in communications design.

Some electrical standards even require the oscilloscope to trigger on a unique data pattern. For example, the ANSI standards for alternate mark inversion (AMI)-coded signals, such as for the 1.544 Mb/s DS1 signal, use a positive isolated one pulse for the pulse test. An isolated one pulse, a pulse preceded by and followed by two or more zero states, is used so that the pulse being evaluated is unaffected by intersymbol interference.

In the center of Figure 1 is a positive isolated one pulse from a DS1 signal captured using the AMI communications trigger. Because the oscilloscope has a two-threshold trigger circuit, it decodes the AMI-encoded pulses, and recognizes and triggers on the isolated one pulse.

Simplifying Measurements

Characterizing communications signals is the most difficult and important part of signal verification. You must examine all relevant parameters and determine if the communications signal is within the standard’s specifications. Measuring all these parameters one at a time is tedious and can easily result in errors.

To streamline verification, most standards specify the shape of the signal by defining a mask. Mask testing provides faster, more reliable characterization and test of communications signals. You simply overlay the mask on the required signals and immediately determine if the signal is in compliance. There are two kinds of masks: pulse templates for lower-speed electrical signals, and eye diagrams for high-speed electrical and optical signals.

Mask testing of electrical signals has been performed with oscilloscopes for many years, but new features in advanced oscilloscopes make this task much easier. These capabilities include automatic generation and display of the mask by the oscilloscope, calibrated variable time delay and voltage scales, communications signal triggering, autoset of the signal to the mask, and automatic mask-violation counting. The combined features make testing quicker and more accurate.

Pulse Parameter Masks

Several parameters must be measured on lower-speed electrical telecommunications signals. The parameters that define the shape of the signal include minimum voltage, maximum voltage, rise time, fall time, and pulse width. A pulse mask specifies the shape of the signal in amplitude and time so you can make all the relevant measurements at one time. The signal is compliant if it does not exceed any of the mask boundaries.

An example of an ANSI pulse mask for a DS1 signal is shown in Figure 2. The mask consists of two polygons or mask regions, labeled 1 and 2 in the figure. For the signal to be compliant to the mask, it must remain in the area between the two mask regions.

Once the mask is displayed and the oscilloscope is capturing pulses, you can use the oscilloscope’s mask-counting feature, if available, to automatically count the number of samples that fall outside of the standard mask limit. Figure 2 shows a DS1 standard waveform that has failed the mask test. It also shows the number of hits (violations) for each mask region (Mask 1 and Mask 2).

Eye-Diagram Masks

An eye diagram is a display of many overlaid waveform acquisitions that resembles the shape of an eye. Eye diagrams may be generated by displaying the data signal while triggering on the clock, or by triggering on the data itself and displaying the data signal after a few periods of delay.

When a pseudorandom data stream is displayed as an eye diagram, the overall quality and stability of the communications system can be observed. Data-stream problems, such as excessive noise or timing jitter, tend to close the eye. In addition, the rise time and fall times of the data signal and pulse parameters, such as duty cycle, overshoot, and undershoot, can be observed. So in one display, you see how well the system is operating.

Eye diagrams almost always are used when measuring high-speed communications signals, including optical signals. Mask testing enhances the usefulness of the eye diagram test.

When a standard eye-diagram mask is displayed, you can quickly determine if the signal is operating within the limits specified by the standard. If automatic measurement of a mask hits is turned on, the oscilloscope can monitor the signal and notify you of any problems.

Take, for example, the OC3/STM1 SONET/SDH signal, a popular 155-Mb/s standard used in local area networks and telecommunications transmission systems. To view this signal, you will need to connect the signal to the oscilloscope using the appropriate optical-to-electrical converter and optical reference receiver filter. Next, you should select the OC3/STM-1 eye diagram trigger and rely on the oscilloscope’s infinite persistence to hold and display all the acquired data values, as shown in Figure 3.

The OC3/STM1 mask shown in Figure 3 consists of three regions. A compliant eye diagram surrounds the hexagonal center-mask region and remains between the top and bottom regions. The display on the right indicates that the automatic mask test has found no mask violations.

Although automatic mask testing is convenient, particularly in the manufacturing environment, there are times when you will want to perform manual mask testing. This is especially true during the design and debug phase of the product.

In the manual mask testing mode, the oscilloscope displays all the acquired waveforms, perhaps over millions of cycles, with respect to the mask. Consequently, you can see any anomalies that might creep in across thousands of acquisitions. The resulting display gives a wealth of information that can be a real asset when characterizing or debugging your design.

Important Measurements

Along with the straightforward measurements made with a pulse or eye mask, you should consider several important complex measurements. These include jitter, extinction ratio, and average optical power.

Jitter is an important measurement, especially for high-speed signals. Cycle-to-cycle jitter can cause the eye opening to close and affect a receiver’s capability to decode a data stream. As the jitter increases, the data-transition points move closer and closer to the decision point of the receiver and eventually the bit error rate of the system will increase.

Cycle-to-cycle jitter can be broken into types: deterministic and random. Deterministic jitter is caused by the data bits preceding the current bit in the data stream. Random jitter is due to random noise.

If the deterministic or pattern-dependent jitter is negligible, random jitter can be characterized and measured by statistically analyzing the data using an oscilloscope’s histogram capability. The oscilloscope should display the rising edge, falling edge, or eye crossing where the jitter will be measured, and then create a histogram of the region where the jitter is occurring.

If the histogram of the placement of the signal edge is a normally distributed curve, the standard deviation is equal to the rms jitter of the waveform. The observed peak-to-peak jitter or other histogram measurements also can be turned on to characterize the jitter (Figure 4).

There are two unique measurements specified by the Bellcore and International Telecommunications Union—Telecommunications sector (ITU-T) standards for optical systems. These measurements are extinction ratio and average optical power.

The extinction ratio measurement is the ratio of the average power level for a logical one to the average power level of a logical zero. The higher the ratio, the more margin the system has to resist added noise.

The extinction ratio automatically is performed by some communications oscilloscopes, making the measurement easy to do. But even if it is automatically done, follow these recommendations for an accurate extinction ratio measurement:

First, an optical reference receiver filter must be used for the extinction ratio measurement. Since the data rate will be high and only average power levels are desired, the filter’s integrating effect will give an accurate representation of the average power levels even at high data rates.

Secondly, be sure that the dark level or zero-light level of the system is correctly measured and accounted for before any optical power readings are performed. Once the dark-level correction has been done, ensure that the signal being tested is a stable eye diagram. Then the automatic extinction ratio measurement can be run.

Average optical power is another measurement specified by the SONET/SDH standards. The measurement can be performed with an oscilloscope with an optical-to-electrical converter or with an optical power meter.

The accuracy and dynamic range of the oscilloscope must support this measurement. Primarily, this measurement must determine the strength of the communications signal. The more power the signal has, the further it can travel in the fiber-optic cable. This obviously is an important measurement when characterizing or debugging a communications system.

Examples of the extinction ratio (in dB) and average optical power (in dBm), along with the statistics (mean and standard deviation) of the measurements over time, are shown at the right in Figure 5.

Ready to Verify

As communications capabilities continue to be incorporated into mainstream electronic systems, you will need to verify communications signals. Fortunately, today’s advanced, yet affordable, digital oscilloscopes can dramatically shorten the learning curve and simplify the task of verifying and debugging communications signals. If you use a scope with acquisition and measurement capabilities tailored to communications standards, connecting, capturing, and characterizing these signals will be a straightforward, productive process.

About the Author

Scott Davidson is a product marketing manager for Tektronix’ oscilloscope business. He has held a number of positions during 14 years with the company, ranging from market analysis to electrical engineering. Mr. Davidson earned B.S.E.E. and M.S.E.E. degrees from Montana State University. Tektronix, P.O. Box 500, M/S 39-729, Beaverton, OR 97077, (503) 627-4763.

July 1997