Measuring composite-video signal performance requires understanding differential gain and phase, Part 2 of 2

Measurement conditions for DG/DP in datasheets
When it comes to DG/DP in datasheets, there are some additional considerations to be aware of, with respect to measurement technique and test conditions. Although DG/DP has nearly become a standard specification in datasheets, the way it is measured is not necessarily standardized. As a result, some DG/DP specifications may not be entirely representative of how the op amp will perform in the actual application. Some of the following test methods and conditions are worth keeping in mind.

The end-points approach is not good practice. Merely taking the gain delta between the lowest and highest DC levels is not sufficient unless the DG/DP errors were perfectly linear with respect to dc offset. However, DG/DP curves can sometimes be quadratic or even cubic relations. In such cases, just measuring the end points may give better data than if the entire sweep range was measured, which is misleading.

The DG/DP published in the datasheet must also be based on testing with the proper chrominance and luminance parameters. The chrominance or sinusoid test signal must have an amplitude of 286 mV_pp (40 IRE), which is the standardized test signal amplitude for NTSC (for PAL, this amplitude is 43 IRE). The frequency of this signal should also be either 3.58 MHz or 4.43 MHz. The 4.43 MHz PAL condition is usually the worst case. DG/DP should be tested and published for both of the frequencies, or at least for the worst case. Performing tests with reduced chrominance amplitudes and frequencies will give better results, but will not properly characterize the part according to NTSC/PAL specifications.

Therefore, a DG/DP specification based on a test with a 1 MHz, 100 mV_pp sine wave may be impressive, but of little use to a video designer. The luminance or dc sweep range is also very important. Even in the previous discussion, this range was typically considered to be 0 IRE to +100 IRE, according to the NTSC luminance specifications. Based on this convention, the NTSC signal has positive-going video above 0 V and negative-going synchronization pulses below 0 V. Assuming that the op amp will always encounter positive-going video, a dc-sweep range from 0 to +100 IRE would be fine.

However, video system designers have been known to invert the video signal during intermediate stages within their closed systems for various purposes such as gamma processing. These designers may need to know the DG/DP of the op amp for an inverted luminance range. In other words, -100 IRE to +100 IRE is a more complete dc-sweep range for DG/DP tests. This will truly test the op amp over the entire dc range in which it might be actually used. Figures 5a and 5b illustrate examples of maximum DG/DP errors in the positive and negative video ranges.

(Click to enlarge image)

Figure 5a and 5b: Maximum DG and DP (-100 IRE to +100 IRE)
(Click to enlarge image)

The loading conditions of the tested op amp are another major consideration. Although successive stages in video systems may be ac-coupled at times, video op amps should still be specified based on a dc load. This should be standard video termination, which is double-terminated 75 Ω or 150 Ω equivalent loading. (Note that in Figure 3, the 50 Ω attenuator and 100 Ω series output resistor form a 150 Ω dc load on the device under test, while maintaining the 50 Ω overall test environment.)

DG/DP numbers based on loads any higher than 150 Ω are misleading since differential gain and phase are affected by the amount of load on the amplifier. An op amp will have better DG/DP with a 1 kΩ load than with a 150 Ω load on its output. On the other hand, DG/DP numbers based on loads lower than 150 Ω can be quite useful since the op amp may need to drive multiple video lines.

The total equivalent load resistances for two, three, and four video loads are 75 Ω, 50 Ω, and 37.5 Ω, respectively. However, not all datasheets will provide the DG/DP errors for more than one video load. Figure 6 shows DG/DP versus the number of video loads in a National Semiconductor op amp datasheet.

Figure 6: DG/DP versus number of video loads
(Click to enlarge image)

An additional caveat regarding DG/DP test conditions is the gain configuration of the device under test. DG/DP numbers for some amplifiers may not be the best at a gain of two. Some might specify DG/DP at unity gain or some other gain where the amplifier has the best numbers. However, video amplifiers are commonly used at a gain of two, and it is very important to provide DG/DP at this setting.

Conclusion
One of the most important factors within an analog composite-video system is how well an amplifier reproduces the composite video signal. Two extremely critical parameters in gauging the quality of this reproduction are differential gain and differential phase. These parameters determine how much the richness and shade of color varies due to changes in its brightness. Ideally, there should be no dependence on the level of brightness.

There are a variety of practices and techniques to measure and quantify this change. However, using a high-precision network analyzer is the method most highly recommended for complete and accurate data. In addition to varying measurement methodologies, DG/DP results are also dependent on an assortment of test conditions. The difference in conditions could be the difference between excellent and mediocre DG/DP numbers. The subcarrier frequency and amplitude, luminance-level range, amplifier-output load, number of video loads, and amplifier gain must all be set in a way that simulates a real video application. DG/DP in a datasheet should be based on these correct conditions. Only then will it be useful for predicting the op amp's behavior when reproducing actual composite video signals.

About the author
Cliff Win is a Senior Applications Engineer for National Semiconductor's Amplifier Product Line. Previously, he has worked extensively with analog and mixed signal video display ICs in National's CRT and Digital Television group. Cliff holds a Bachelor of Science degree in Electrical Engineering from the University of California at Davis.