First, Let's take a look at part 76 …..
"..... such as intermodulation products, second and third order distortions or discrete-frequency interfering signals not operating on proper offset assignments."
Restated, this means anything within the channel bandwidth (excepting noise) that shouldn't be there.
The important thing to keep in mind about Coherent Disturbance tests is that it includes more than CSO and CTB. Automated CSO/CTB measurements are fine for some things but don't quite meet proof of performance test requirements. The reason is that automated routines perform measurements only at the common CSO/CTB frequencies (that's at the visual carrier, +/- 750 kHz, and +/- 1.25 MHz for the standard channel plan in the US). Here's a fairly common example of a coherent disturbance that automated CSO/CTB measurements would miss:
The small bump slightly to the right of the second graticule line is CTB. The large "beat" (actually, a CW carrier) on the far right is 3.75 MHz above the visual carrier frequency. In this case, it was caused by the local oscillator of channel 97 (A-3) leaking into the system. The local oscillator frequency for channel 97 is at 103.25 + 45.75 = 149.00 MHz. Which is 3.75 MHz above channel 18's visual carrier. This particular coherent disturbance is particularly significant because it is near the color subcarrier frequency (3.58 MHz) and causes color beats in the picture. If only automated CSO/CTB test methods are available from your test gear be aware that you may be missing some very important coherent disturbances. Carefully monitoring subscriber comment along with looking for beats in the picture during installs and service calls may provide evidence of such disturbances.
Another reason for using interactive instead of automated test methods is that, because a narrow resolution bandwidth filter and heavy video filtering is used for the coherent disturbance test, it is necessarily slow. If you use automated techniques, the analyzer will check at each of the pre-programmed frequencies, whether there's anything there or not, resulting in the channel being out of service for longer than necessary - and you might still get the wrong numbers!Fortunately, Coherent Disturbance tests can be run properly, accurately, and quickly using semi-automated or manual methods.
Performing coherent disturbance tests properly, accurately, and quickly.
Don't use automated methods that only measure CSO and CTB. The key is to use interactive manual or semi-automated methods that allow the operator to select the disturbance(s), if any, to be measured. This allows a check across the entire channel and measurement of only the largest disturbance(s). If no disturbances are visible, the channel can be returned to service immediately.
For one popular analyzer, this is the difference between having the channel out of service for more than 90 seconds (and still not doing a thorough test), and having it off for 10 seconds and doing the test properly! When disturbances are found, the largest one can be measured first, and assuming it meets the requirements, the channel can be returned to service in less than 30 seconds.
Using a preselector filter.Consider alternatives to using tunable preselector filters. It's very easy to make mistakes with tunable filters, especially when looking for signals across the entire channel bandwidth
There are a couple of reasons for the filter. One is to minimize intermodulation distortion (such as CSO and CTB) produced in the spectrum analyzer. With many of the newer analyzers, you don't need the filter if you keep the analyzer's input level in the 5 dBmV to 10 dBmV range. Alternatively, use a fixed tuned filter that is several channels wide. For example, a 100 MHz wide filter effectively reduces the number of channels to 16, dramatically reducing the likelihood of significant beats being generated in the analyzer.
Another reason for a preselector filter is to reduce the effects of high Peak to Valley ratio across the system's spectrum. For example, here's a plot of the 24 hour tests for a (not particularly well maintained) system.
Let's say that the carrier level for channel 2 is 19 dBmV and for channel 78 its 10 dBmV. The thing to be aware of is that, when performing some tests, most cable TV analyzers follow a sequence something like this:
1. Measure total power across the entire operational spectrum of the analyzer (maybe 1 MHz to 1800 MHz). This information is used inside the analyzer to automatically adjust its input characteristics - such as input step attenuator or, in some cases, turning on a preamplifier.
2. Automatically adjust the input attenuation or gain to ensure the analyzer is not over driven (is not generating significant beats).
3. Set other internal parameters - resolution bandwidth, video filter, sweep speed, averaging, etc as needed to the test to be performed.
4. Perform the measurement - this may require some interaction with the user to position markers, etc.
For the input spectrum shown above, there is approximately 9 dB difference in amplitudes between channels 2 and 78. The analyzer settings don't change. Therefore, there will be approximately 9 dB difference in the measurement range for the two channels. So, if we can measure CSO and CTB down to, say 65 dB on channel 2 before we hit the noise floor of the analyzer, we may only be able to measure down 56 dB on channel 78. Of course, each of those test results would indicate compliance with the FCC requirements but it would be better to have more measurement dynamic range.
Now, if we were to use a bandpass filter (fixed or tuneable) in front of the analyzer the total power input to the analyzer would be greatly reduced, the analyzer would automatically use less input attenuation resulting in greater dynamic range for the measurement. We would expect better agreement between the test results and have a better understanding of actual system performance.
Non disruptive coherent disturbance tests (including in-service CTB).
If you run coherent disturbance tests at the tap instead of at the output of a converter (get your "good engineering practices" statement ready), consider running in-service coherent disturbance tests. If the spectrum analyzer only takes readings during quiet lines in the vertical blanking interval, all coherent disturbances except those, such as CTB, that fall near the visual carrier frequency can be measured without service disruption. Even CTB can be measured without disruption if a device is installed in the headend to turn off the visual carrier during one or two lines in the vertical blanking interval. See the information about the TVMS model 4200 In Service Test Processor in the Products section of this web site. Again, its important to test for all coherent disturbances - not just CSO and CTB regardless of the test method employed.
