QAM measurements on the OMA software.
In addition to the numerical measurements provided on individual plots, the Measurements tab provides a summary of all numeric measurements, including statistics.
The Measurements tab- a summary of all measurements in one place
Make measurements faster
The Tektronix OM1106 software is designed to collect data from the oscilloscope and move it into the MATLAB workspace with extreme speed to provide the maximum data refresh rate. The data is then processed in MATLAB to extract and display the resulting measurements.Take control with tight MATLAB integration
Since 100% of the data processing occurs in MATLAB, test engineers can easily probe into the processing to understand each step along the way. R&D labs can also take advantage of the tight MATLAB integration by writing their own MATLAB algorithms for new techniques under development.Use the optimum algorithm
Don’t worry about which algorithm to use. When you select a signal type in the Tektronix OM1106 software (for example, PM-QPSK), the application applies the optimal algorithm for that signal type to the acquired data. Each signal type has a specially designed signal processing approach optimized for that signal. This means that you get results right away.Don’t get stymied by laser phase noise
Signal processing algorithms designed for electrical wireless signals don’t always work well with the much noisier sources used for complex optical modulation signals. Our robust signal processing methods tolerate enough phase noise to make it possible to test signals that would traditionally be measured by differential or direct detection such as DQPSK.Find the right BER
Q-plots are a great way to get a handle on your data signal quality. Numerous BER measurements versus decision threshold are made on the signal after each data acquisition. Plotting BER versus decision threshold shows the noise properties of the signal. Gaussian noise will produce a straight line on the Q-plot. The optimum decision threshold and extrapolated BER are also calculated. This gives you two BER values: the actual counted errors divided by the number of bits counted, as well as the extrapolated BER for use when the BER is too low to measure quickly.
Q-plot.
Constellation diagrams
Once the laser phase and frequency fluctuations are removed, the resulting electric field can be plotted in the complex plane. When only the values at the symbol centers are plotted, this is called a Constellation Diagram. When continuous traces are also shown in the complex plane, this is often called a Phase Diagram. Since the continuous traces can be turned on or off, we refer to both as the Constellation Diagram.The scatter of the symbol points indicates how close the modulation is to ideal. The symbol points spread out due to additive noise, transmitter eye closure, or fiber impairments. The scatter can be measured by symbol standard deviation, error vector magnitude, or mask violations.
Constellation diagram.
Constellation measurements
Measurements made on constellation diagrams are available on the “fly-out” panel associated with each graphic window. The measurements available for constellations are described below.| Measurement | Description |
|---|---|
| Elongation | The ratio of the Q modulation amplitude to the I modulation amplitude is a measure of how well balanced the modulation is for the I and Q branches of a particular polarization’s signal |
| Real Bias | Expressed as a percent, this says how much the constellation is shifted left or right. Real (In-phase) bias other than zero is usually a sign that the In-phase Tributary of the transmitter modulator is not being driven symmetrically at eye center |
| Imag Bias | Expressed as a percent, this says how much the constellation is shifted up or down. Imaginary (Quadrature) bias other than zero is usually a sign that the Quadrature Tributary of the transmitter modulator is not being driven symmetrically at eye center |
| Magnitude | The mean value of the magnitude of all symbols with units given on the plot. This can be used to find the relative sizes of the two Polarization Signals |
| Phase Angle | The transmitter I-Q phase bias. It should normally be 90 |
| StdDev by Quadrant | The standard deviation of symbol point distance from the mean symbol in units given on the plot. This is displayed for BPSK and QPSK |
| EVM (%) | The RMS distance of each symbol point from the ideal symbol point divided by the magnitude of the ideal symbol expressed as a percent |
| EVM Tab | The separate EVM tab shown in the right figure provides the EVM% by constellation group. The numbers are arranged to correspond to the symbol arrangement. This is ideal for setting Transmitter modulator bias. For example, if the left side groups have higher EVM than the right side, adjust the In-phase Transmitter modulator bias to drive the negative rail harder |
| Mask Tab | The separate Mask tab shown in the right figure provides the number of mask violations by constellation group. The numbers are arranged to correspond to the symbol arrangement. The mask threshold is set in the Engine window and can be used for pass/fail transmitter testing |
| Quadrature Error | The deviation of the transmitter IQ phase from 90 degrees. |
| IQ Offset | The ratio between the carrier leakage power and the signal power in dB. This metric is impacted by Quadrature Error, Real and Imaginary bias. |
| IQ Imbalance | The ratio of the real and imaginary constellation size in dB. It is related to the linear measure, Elongation. |
Color features
The Color Grade feature provides an infinite persistence plot where the frequency of occurrence of a point on the plot is indicated by its color. This mode helps reveal patterns not readily apparent in monochrome. Note that the lower constellation groups of the example below have higher EVM than the top groups. In most cases this indicates that the quadrature modulator bias was too far toward the positive rail. This is not evident from the crossing points which are approximately correct. In this case an improperly biased modulator is concealing an improperly biased driver amp.
Color Grade Constellation.
Color Grade with fine traces.
Color Key Constellation Points is a special feature that works when not in Color Grade. In this case the symbol color is determined by the value of the previous symbol. If the prior symbol was in Quadrant 1 (upper right) then the current symbol is colored Yellow. If the prior symbol was in Quadrant 2 (upper left) then the current symbol is colored Magenta. If the prior symbol was in Quadrant 3 (lower left) then the current symbol is colored Light Blue (Cyan). If the prior symbol was in Quadrant 4 (lower right) then the current symbol is colored Solid Blue.
This helps reveal pattern dependence. The following figure shows that pattern dependence is to blame for the poor EVM on the other groups. In QPSK modulation, the modulator nonlinearity would normally mask this type of pattern dependence due to RF cable loss, but here the improper modulator bias allows that to be transferred to the optical signal.
Field eye diagram.
Field eye measurements
| Measurement | Description |
|---|---|
| Q (dB) | Computed from 20 × Log10 of the linear decision threshold Q-factor of the eye |
| Eye Height | The distance from the mean 1-level to the mean 0-level (units of plot) |
| Rail0 Std Dev | The standard deviation of the 0-level as determined from the decision threshold Q-factor measurement |
| Rail1 Std Dev | The standard deviation of the 1-level as determined from the decision threshold Q-factor measurement |
In the case of multilevel signals, the above measurements are listed in the order of the corresponding eye openings in the plot. The top row values correspond to the top-most eye opening.
The above functions involving Q-factor use the decision threshold method described in the paper by Bergano1. When the number of bit errors in the measurement interval is small, as is often the case, the Q-factor derived from the bit error rate may not be an accurate measure of the signal quality. However, the decision threshold Q-factor is accurate because it is based on all the signal values, not just those that cross a defined boundary.
1N.S. Bergano, F.W. Kerfoot, C.R. Davidson, “Margin measurements in optical amplifier systems,” IEEE Phot. Tech. Lett., 5, no. 3, pp. 304-306 (1993).
Additional measurements available for nonoffset formats
| Measurement | Description |
|---|---|
| Overshoot | The fractional overshoot of the signal. One value is reported for the tributary, and for a multilevel (QAM) signal it is the average of all the overshoots |
| Undershoot | The fractional undershoot of the signal (overshoot of the negative-going transition) |
| Risetime | The 10-90% rise time of the signal. One value is reported for the tributary, and for a multilevel (QAM) signal it is the average of all the rise times |
| Falltime | The 90-10% fall time of the signal |
| Skew | The time relative to the center of the power eye of the midpoint between the crossing points for a particular tributary |
| Crossing Point | The fractional vertical position at the crossing of the rising and falling edges |
Measurements versus Time
In addition to the eye diagram, it is often important to view signals versus time. For example, it is instructive to see what the field values were doing in the vicinity of a bit error. All of the plots that display symbol-center values will indicate if that symbol errors by coloring the point red (assuming that the data is synchronized to the indicated pattern). The Measurement versus Time plot is particularly useful to distinguish errors due to noise, pattern dependence, or pattern errors.For more information: Tektronix