Distortion101

Updated July 17, 2011

Click here to go to our main page on microwave measurements

Click here to go to our page on spectrum analyzer measurements

Click here to go to our main page on microwave receivers

This page was contributed by Matrix Test Equipment, Inc..

Matrix Test Equipment offers some excellent technical literature that will further your understanding on the topic of distortion, that you can download at: http://www.matrixtest.com/techlit.htm.

MTN-107 Test Setup for Measuring X-Mod, CSB & CSO
Describes a method that uses a mean square circuit as a detector and a fast Fourier analyzer as a readout.

MTN-108 Notes on Composite Second & Third Order Distortion
Presents formulas for determining the number of second and third order distortion products that can be expected in a group of carriers. Graphs are shown depicting the distortion components versus frequency.

MTB-109 The Relationship of Intercept Points & Composite Distortions
Presents a discussion of the relative strength of the various second and third order distortion products.

MTN-110 Extending the Limits of Composite Intermodulation Distortion Measurements
Methods for measuring distortion that occurs near the noise level.

MTN-111 PM to AM Conversion of Distortion Products by Slope Detection in TV Receivers

Here's a clickable index to the material presented on the page that you are currently looking at. It is presented in a "frequently asked questions" format:

1. What causes distortion?
2. Can passive devices cause distortion?
3. What types of distortions are important in a cable system?
4. What are composite triple beat and composite second order distortions?
5. What RF bandwidth filter should I use for distortion measurement?
6. How do I measure CTB and CSO?
7. How can the effects of coherent carrier frequencies on the consistency of distortion measurement be minimized?
8. What is the relationship between carrier level and second and third order distortions?
9. What are the important considerations in setting signal source levels for my DUT?
10. How can I tell if distortion is coming from test equipment (e.g. spectrum analyzer) or the DUT?
11. What is carrier-to-noise ratio (C/N) and how is it measured?
12. When measuring C/N how do I compensate/determine test equipment noise contribution?
13. How do I measure C/N in the presence of digitally modulated carriers?
14. What is crossmodulation?
15. How is crossmodulation measured?
16. Can I make a two-tone intermodulation distortion measurement and extrapolate to get the CTB, CSO and crossmodulation?
17. We usually measure the amplitude intermodulation products, are phase intermodulation products present and are they a problem?
18. Why does the carrier level drop when the carrier is modulated?
19. Why is square modulation used as a test signal when the TV signal SYNC is clearly a much narrower pulse?

Frequently asked questions about distortion

1. What causes distortion?
Distortion is the result of a nonlinear transfer function. If the output of the device under test (DUT) is not an exact replica of the input, distortion products are present. Although usually caused by active devices, passive components such as couplers and connectors can cause distortions.

2. Can passive devices cause distortion?
Yes, absolutely! Distortion in passive devices, also known as passive intermodulation, is a very real problem. At some level all components generate distortion. Significant distortions can be caused by RF connectors because they may have dissimilar metals or dirt and oxidation at their contacts, ferrite devices such as splitters and couplers, relays, attenuators, PIN diodes, carbon resistors, and even coaxial cable with an impure dielectric. The main problem is usually due to the center conductor, not the ground. You can minimize passive intermodulation by using the same metal for all conductors that make mechanical contact to each other.

3. What types of distortions are important in a cable system?
The important distortions are primarily second and third order. Higher order distortions do occur but at normal operating level they usually are small enough to be ignored. Second order distortions are the result of a second order non-linearity. They are formed by one mixing operation and result in products such as A+B, A-B, 2A, 2B. Third order distortions are the result of third order non-linearity and are formed by two mixing operations such as (A)+(A)+(A) or 3A, (A)+(A)+(B), 2A+B, A+B+C. Crossmodulation is a third order distortion. The products do not all have the same amplitude, for example the 2A product in 6 dB lower than the A+B product. For more information on this subject, see MTN-108.

4. What are composite triple beat and composite second order distortions?
Composite triple beat distortion (CTB), also called composite third-order beat, is the ratio of carrier to composite third order distortion products. Composite second-order distortion (CSO) is the ratio of carrier to composite second-order distortion products. Both of these distortions result from multiple carriers experiencing second and third order non-linearity, which generate multiple distortion products. These products are measured as a group; thus the term composite distortion as opposed to discrete distortion.

Triple beat (CTB) products fall close to the carriers because of the choice of equally spaced carrier frequencies. Second order (CSO) distortion products fall +/- 1.25 MHz from each carrier. A small quantity of second order products fall 0.75 MHz from some carriers. They are a result of the 10 MHz instead of the usual 6 MHz spacing between Channel 4 and Channel 5. The definition of these distortions refers to spectrum analyzer measurements. The spectrum analyzer does not read the power in CTB or CSO with 100% accuracy, but because it is used as a standard other methods for making the measurement are adjusted to agree with spectrum analyzer measurements. See MTN-110 for more information.

5. What RF bandwidth filter should I use for distortion measurement?
If CSO is being measured then the BW should be at least 2.5 MHz in order to pass the second order products that may appear 1.25 MHz above and/or below the video carrier or expected center frequency. For non-NTSC plans or FCC frequency offset carriers, this bandwidth may vary.

6. How do I measure CTB and CSO?
The basic way to measure these products is with a multiple frequency signal generator and a spectrum analyzer. For almost all applications a band pass filter is required at the input of the spectrum analyzer because the spectrum analyzer, when driven with multiple carriers will generate its own distortion products. For CTB we seek the ratio of the carrier to the cluster of distortion products that fall in a narrow band centered at the carrier frequencies. For CSO the measurements are the same except the distortion products fall +/- 1.25 MHz from each carrier. For a description of the CTB and CSO products and how they vary across the band, see MTN-108. For these measurements the analyzer is set as follows.

1. Frequency carrier frequency
2. Resolution Bandwidth 30 kHz
3. Video Bandwidth 30 Hz or lower, Video averaging is also useful
4. Span 200 kHz, Nominal

Then: 1. Supply the required number of carriers to the input of the DUT. 2. Measure the output carrier level at the channel being measured. 3. Turn that carrier off. 4. Measure the peak of the distortion. The ratio of carrier to the peak of the distortion is defined as the CTB. A similar procedure is used for CSO except that the products are +/- 1.25 MHz from the carrier. For these CSO tests the carrier should be off.

7. How can the effects of coherent carrier frequencies on the consistency of distortion measurement be minimized?
If the carriers in the multiple frequency generator are set for example, to within 1 kHz of the correct frequency, all the composite beats will fall within a 6 kHz band. This stresses the measurement because the high crest factor of the distortion. One way to minimize is to randomly offset the signal source frequencies by up to 5 kHz. The Matrix ASX-16B and later models have the capability to remotely adjust the carrier frequency trim. Matrix CATV Equipment control program has the capability to automatically randomize the carrier frequencies. The use of a square law detector eliminates the crest factor problem and extends the lower limit of the measurement. See MTN-110.

8. What is the relationship between carrier level and second and third order distortions?
When measured on a dB versus dB scale, the signal has a slope of 1, second order distortions have a slope of 2 and third order distortions have a slope of 3. This means for every 1 dB change in carrier level the second order distortions change by 2 dB and the third order distortions change by 3 dB. Therefore, carrier to second order distortion (CSO) changes 1 dB for 1 dB change in carrier level and carrier to third order distortion (CTB) changes 2 dB for 1 dB change in carrier level.

9. What are the important considerations in setting signal source levels for my DUT?
It is very important to correctly adjust the signal level to the DUT, because every dB of discrepancy from the expected level may cause a 2 dB of error in third order distortion measurements. When leveling an amplifier for example, the target reading at the measuring device should account for any cable or system losses, or attenuator pads from the output of your device. In addition, if dBmV units are being used to specify the target levels, an adjustment must be made if the output of your device is 75 ohms and the input of the measuring device is 50 ohms.

10. How can I tell if distortion is coming from test equipment (e.g. spectrum analyzer) or the DUT?
To determine if distortion is coming from the spectrum analyzer, increase the attenuator value at the front end of the analyzer and observe if the distortion decreases. If it does, then the distortion is being created at the spectrum analyzer. Make sure there is adequate bandpass filtering between the device under test and the spectrum analyzer. Another powerful test is to jump out the DUT and measure the distortion. Here the distortion should be well below the required distortion level. Typical distortion levels for Matrix equipment is well below –100 dB.

11. What is carrier-to-noise ratio (C/N) and how is it measured?
Carrier to noise ratio is the ratio of the carrier signal power to the noise power in some specified channel, usually expressed in decibels (dB). For the analog channels the noise is assumed flat and the result of thermal and amplifier noises. Measurements are made with a spectrum analyzer. The carrier is used as a reference and the noise is usually measured using the “noise marker” of the spectrum analyzer set to a quiet spot in the band. The noise marker reading is now normalized to 4 MHz and the measured C/N is the ratio of the carrier to the normalized noise level, measured in dB. The 4 MHz normalization frequency is for the American standard, other standards may require a different bandwidth.

12. When measuring C/N how do I compensate/determine test equipment noise contribution?
One way is to disconnect the device under test from the test equipment and observe if the noise floor drops. If it does not, then it means that the noise of the device is below the noise of the test equipment and you cannot determine the real noise of the device. An amplifier with sufficient gain placed after the filter can bring up the noise of the device so that it rises above the noise floor of the test equipment. If the noise drop is at least 2 dB then a correction may be applied to determine the correct noise floor of the device.

13. How do I measure C/N in the presence of digitally modulated carriers?
The distortion from a digital source manifests itself as noise. The total of the test device noise, thermal noise and distortion from the digital carriers is called carrier to composite noise (CCN). If the noise level is measured without the digital carriers, the result is the carrier to noise. The difference in the noise levels is a measure of the intermodulation products from the digital carriers.

14. What is crossmodulation?
Crossmodulation or XMOD is third order distortion product which transfers the modulation of one carrier to another carrier. Crossmodulation is defined as the ratio of the residual modulation on a carrier compared to the modulation of the carrier at 100% modulation. On a TV picture crossmodulation may appear as a faint image from another channel. This is not to be confused with an image from a distant station at the same frequency as the desired channel.

15. How is crossmodulation measured?
Crossmodulation is measured by supplying one or many modulated carriers to the Device under test along with one carrier operating CW. Crossmodulation is the ratio of the residual modulation to 100% modulation. For this measurement, the direct measurement of the 15.750 kHz sidebands turns out to be unsatisfactory. There are many reasons for this, one is that for the measurement the carrier must be present and the carrier noise and CTB components limit the level of the measurement. Another problem is that many amplifiers generate phase modulation distortions, which are generated by the nonlinear reactances in the semiconductors. Phase modulation sidebands appear on the spectrum analyzer spectrum display but are not important for our measurement. Actually what we are interested in is AM crossmodulation primarily because the TV video modulation is AM. So how do you measure crossmodulation? The answer is to first demodulate the carrier in question. AM demodulation is immune to the PM modulation so the spectrum analyzer can operate as an excellent demodulator with the following setup.

1. Frequency: set to carrier frequency
2. Span: 0.0 Hz
3. Detector mode: linear
4. Resolution BW:1 MHz
5. Video BW: 100 kHz

Extract the vertical output signal, usually available on the back panel of the spectrum analyzer, and connect it to a low frequency spectrum analyzer tuned to 15.750 kHz (note that many RF spectrum analyzers can tune to an input frequency as low as 15 kHz and it may be tempting to use an RF spectrum analyzer for this application but they are not satisfactory because their performance is very poor at low frequencies.) The magnitude of the level of the 15.750 kHz is measured first with the test carrier modulated and again with the test carrier operating CW. The ratio of these voltages is the crossmodulation. There is one other important point, this method measures crossmodulation with reference to 100% modulation. This is equivalent to measuring the ratio of the amplitude modulation sidebands with and without modulation. Sometimes another definition for crossmodulation is used which defines it as the ratio of the first sideband to the carrier.

16. Can I make a two-tone intermodulation distortion measurement and extrapolate to get CTB, CSO and crossmodulation?
You can certainly obtain data this way, but the question is how well do these measurements correlate to "real" data. For perfect conditions the measurement of two-tone second order and third order distortions should allow the calculations of all other conditions. See MTN-109 for simple calculations. Experience has shown that extrapolation to multiple frequency measurements results in unacceptable variations.

17. We usually measure just amplitude intermodulation products, are phase intermodulation products present and are they a problem?
In many amplifiers phase intermodulation exists because the semiconductor junction capacity is a function of the applied RF voltage. The resultant products are as real as the amplitude intermodulation products and at exactly the same frequencies. The PM and AM intermodulation products are usually not in phase but do contribute to the overall distortion. The analog signals are almost immune to closely spaced pairs of phase intermodulation products (see MTN-111) but QAM signals can be affected. Widely spaced PM products behave as AM products and because of phase offsets may add or subtract amplitude.

18. Why does the carrier level drop when the carrier is modulated?
For normal AM modulation the carrier should stay constant but for the TV industry only downward modulation is used. For 100% square wave downward modulation the carrier decreases by 6 dB from the CW level and the first sidebands are 3.9 dB below the new carrier.

19. Why is square modulation used as a test signal when the TV signal SYNC is clearly a much narrower pulse?
Square modulation is easy to generate and has become the standard. Also, all modulations are synchronous for the same reason.