It's easier than you might think
You've just been told you need to start using the reverse path in your system for digital services. From what you've heard and read, it's going to be a major headache. You're expecting the worst. You're filled with fear, uncertainty and doubt . . . but you shouldn't be.
Making the reverse path work correctly has the potential of being a big project in a small system, or a huge project in a big system. But, with proper planning and a few tools, you can minimize the difficulty, and reduce time and cost of implementation. It is planning and execution that make the difference. Figure 1 shows an ideal reverse path.
Three key factors affect reverse path operation: ingress, impulse noise and common path intermodulation distortion (CPID). In this article, we'll take an in-depth look at ingress and determine a strategy for dealing with it. We'll also cover briefly what causes impulse noise and CPID, and how to identify these factors.Sources of ingress
Let's start with ingress. You'll probably find that with a hybrid fiber/ coax (HFC) plant, better than 90 percent of all ingress problems will come from the customer premises. Even with older plants with little or no fiber, the majority of ingress enters the plant at the drop. If you have adequately maintained your plant for minimum leakage (CLI), ingress should also be minimal.
Organization and planning are the key. Remember, set-up and alignment of the reverse path is similar in many respects to forward path set-up, but is also very different in others. If you have been plagued with CLI problems in the system, you are likely to find ingress a serious problem. However, even a tight system may still yield some surprises with unexpected ingress.
The reason is some sources of ingress can have very high field strength. Examples many of us have dealt with are pager interference to channel 19 or ham radio repeater interference to channel 18. In the return frequencies between 5 and 47 MHz, there are amateur radio frequencies in use at 7.0 to 7.3 MHz, 10.1 to 10.15 MHz, 14.0 to 14.35 MHz, 18.068 to 18.168 MHz, 21.0 to 21.45 MHz, 24.89 to 24.99 MHz, and 28.0 to 29.7 MHz. Amateurs may operate with up to 1500 watts peak envelope power (PEP) on all these frequencies except 10.1 – 10.15 MHz, where power is limited to 200 watts PEP.
Like television viewing and Internet surfing, ham radio is a usually a leisure-time activity. Thus, ingress may be worse at times when the most stress is put on the cable system. The 27 MHz Citizens Band is limited to much lower power, but many illegal high-power amplifiers are in use. While interest in CB has waned over the last few years, in some areas, CB activity is still at a level where use of return frequencies in Channel T-10 may not be wise. Figure 2 shows interfering carriers.
Finding an ingress point where activity is as intermittent as with ham radio may be difficult. If no employee of your system is a ham, it is a good idea to get to know several of the amateurs living in or near your system. Once you have isolated an area where ham ingress is occurring, it is often possible to have a ham come into the area with a mobile transmitter. He can then transmit while your technicians check connections, amplifiers and drops nearby. It is often possible to eliminate or significantly attenuate ingress in this manner.
In systems where little fiber is in place for cascade reduction, interference from distant short wave transmitters may become more of a problem over the next few years, as sun spot activity increases, and skip signals become stronger and higher in frequency.Solving ingress problems
Now that you have some idea as to the source of interference, let's look at how to minimize problems. Loose connectors or line connectors that have been improperly installed cause most of the difficulties you will encounter in the outside plant. Similarly, at the drop, you can expect problems related to F-connectors that are loose and/or corroded. Single-shielded drop cable can be another source of ingress. In the case of some older plant, there may be enough degradation in the drop cable to cause not only problems with the reverse path, but also for downstream digital services.
First, let's consider the mechanical health of your system. If you have an older plant, you may need to spend time looking for cracked, corroded or loose fittings. New plant will only be as good as the contractor who installed it. A quality installation monitored by system personnel will be nearly free of ingress and CLI. Obviously, fiber will be free of ingress problems, but does not have the "headroom" of coaxial plant. Ingress in the coax portion of the plant can result in impulses that overdrive the laser and cause clipping.
The effects of ingress on Internet service will manifest themselves as loss of speed in data transfer because of increased resends. When severe, ingress will cause communication to end abruptly. With telephony, the results may cause the subscriber to experience the same type of drop outs and noise typical of cellular telephone communications. Ingress can also cause loss of communication between the server and the set-top box.
One of the best tools for troubleshooting ingress is a signal level meter, as shown in Figure 3. Signal level meters are available today with ingress monitoring capability. This monitoring capability will allow a number of frequency windows to be set up, each with its own level threshold. Anytime one of these thresholds is violated, data that is time and date stamped is stored, with the option of saving the spectral display as well. Using such a tool, it becomes easy to determine the frequency of the ingress or if wide band interference is often coming from an electric appliance.
While outside plant ingress can be severe, it is usually limited to a few points and a few discrete frequencies. The nasty problems more often come from the subscriber premise. In bringing up a reverse path system, many operators have chosen to place a high-pass filter at the drop. This is expensive and time-consuming, but it does allow you to handle problems on a case-by-case basis as new services are deployed. Remember, the least shielded device in your plant is apt to be the subscriber's TV. If the set is directly connected to the system, not through a set-top box, you may have to place a high-pass filter at the input to the TV as you try to bring new services into the home.
When in-home wiring was the responsibility of the cable operator, some subscribers added illegal outlets to their homes. With the change in rules and more information on how to properly add outlets, some homeowners are following accepted practice, but some are not. One approach is to assume the worst, and rewire. In some cases this is faster and less costly than troubleshooting. In any case, an installer should examine the visible wiring, and measure the loss between the ground block and the set-top box or modem. Cable loss can be easily predicted, even if the length of wiring requires an educated guess. Any result more than 2 dB from the estimate should be suspect. At this point, it is time to consider which is the most cost-effective — to reinstall the wiring or troubleshoot. More often than not, wiring is the less expensive, and often more reliable, choice.
Don't forget to look at what is coming out of the house, also. Be certain that an appliance, illegal intercom system or some other device in the home isn't degrading your system performance. The list of devices that generate noise or RF is long.
This is a good time to look at the entire system. As mentioned earlier, there are similarities between forward and reverse path maintenance. It's fairly easy to isolate problems when parts of the forward path are in trouble. Subscriber phone calls quickly localize the problem to a broken "limb, branch or twig." The reverse path is different. A broken "twig" can affect the entire tree, right back to the root (the headend). Troubleshooting, therefore, takes place in reverse, often starting with a problem that is identified in the headend. The problem may be ingress on a discrete frequency. It may be wide band interference that is so broad in frequency the only indication is that the entire noise floor is raised. Or it could come in the form of laser clipping.
Start at the combined test point, making a snapshot of the spectral display of the interference. You may want to compare it with other measurements as you move out into the system. This is especially true if you've determined there are multiple levels of interference not caused by the same problem. Then look at the individual feeds to the headend and follow the fiber or truck to the source of interference. You'll want to follow the problem trunk until the ingress has been isolated to a trap distribution leg. Next, work your way out into the system until you have found the distribution leg, and finally, the tap where the noise or ingress is entering the system.
As you gain more experience in troubleshooting, you will learn the signatures of the various problems. This will lead to faster detective work. You may find some of your first forays into the field a bit frustrating, but they will always be interesting and challenging. Early in the process of troubleshooting the reverse path, I used a communications receiver and a spectrum analyzer so I could both see and hear the problem. Today, that can be done on a single handheld instrument. With this tool, you are quickly able to determine the type of interference and may be able to hear a station ID or location.
Because most reverse path problems can be traced to the customer drop ports at the distribution taps, let's review a few steps you can take to isolate the interference. First, check the output side of the tap. If the ingress is still present, move further downstream. If the ingress is gone, move to the input side. If the ingress is present on the input side, disconnect the customer drops one-by-one until the ingress goes away and you have isolated the problem to the customer premise. Now the sleuthing begins. Go to the ground block; check for ingress there. And so on, back into splitters and equipment the subscriber has connected to the system.
You may not have to go any further than the ground block. Often, ground blocks are a major source of ingress. In one situation, there was a corroded block with a high level of RF present because of a frequency resonant length of drop cable (which placed a high RF level at the offending point). In this case, every 1.47 MHz, a very high level of RF was present, up to 14.7 MHz, the 10th harmonic of the broadcast station! Using the communication receiver, the station was quickly identified, long before anyone did the math to determine what the fundamental frequency might be. The drop was located more than a mile from the radio station, but because the RF level on the drop was high enough, the operator decided to change the length to get it out of resonance. In this case, the only choice was to lengthen the drop by looping it back on itself.
Perhaps the most unusual interference I've encountered came from an unlikely source — a tube type audio amplifier. For some unknown reasons, the output tube, a single-ended 6V6, was emitting a dirty signal about every 100 kHz up to 65 MHz. The tube was replaced, and the signal went away. Sometimes, you have to expect the unexpected.Impulse noise and CPID
Electrical/mechanical interference is sometimes dubbed the "hair dryer" syndrome. Like the signals from ham radio operators and CBers, this signal is usually intermittent, but fortunately, it is usually of shorter duration. This is interference you will likely never track down. Typically, this type of noise comes from motors with brushes. Induction type motors are usually very quiet in nature. Brush type motors are the most common on hair dryers, vacuum cleaners and electric mixers. But before you pass this off as a unisex problem, I should add that electric drills and other small AC power tools generate impulse noise as well. Some larger motors are capacitor start, induction run motors which may generate a fairly severe noise spike when starting. Again, expect the unexpected. Some older door-bell transformers have thermal protection that may fail and arc continuously. Usually, subscribers will find and fix this problem because it interferes with their AM radio and TV. But if they listen to FM and their set is on your cable system, they may not be aware of the problem. Figure 4 shows impulse noise.
In these cases, isolate with a high-pass filter until the subscriber fixes the problem. And, no, you're probably never going to find every noisy hair dryer or mixer. But what you do need to do is monitor at the headend continually, looking for a rise in the noise floor. If a problem suddenly appears, it most likely came with a new batch of installs, which is a place to start looking. The answer, once the source of the interference is found, will be to "tighten" the wiring at the subscriber premises and the drop. In some cases, it may be necessary to work with subscribers to have them fix or replace the offending appliance and perhaps, in the most severe case, disconnect service. Most subscribers will be reasonable if the problem is explained and a solution can be mutually agreed upon.
Common path intermodulation distortion, like the others discussed, can be avoided with good maintenance practices, and, while sometimes difficult to isolate, it can be easily fixed. CPID has a unique signature when viewed on a spectrum analyzer. CPID will display third order harmonics, evenly spaced throughout the return path spectrum. Like each of the other distortions, it can be traced back to a set of causes. These usually include poorly installed connectors, corrosion, dissimilar metals used in connectors, or kinks and cracks in rigid coax. Figure 5 shows CPID.
Now that the most common problems have been identified, let's go back to the basics — the three most important things to remember in maintaining a cable TV system — levels, levels and levels. If levels are properly set and maintained, most everything else falls into place. In fact, levels are more important in the reverse path than in the forward path because, as stated earlier, there is less headroom and less tolerance for noise in the reverse path. Levels that are too high will result in laser clipping and intermittent service, and levels that are too low do likewise.
Digital signals are robust, but they do not degrade gracefully. Digital signals are subject to the cliff effect, which simply means they work, or they don't — black or white with no shades of gray.
The impulse noise that interferes with an analog video signal may be annoying, but you can still watch it. A digital signal may fail to work at all if it receives too much noise or interference.
With the advent of interactive digital services which require reverse spectrum to operate comes a new set of challenges. As these services grow in popularity, plants will have to be upgraded and maintained to a higher level than in the past. Fortunately, a new generation of handheld tools will help cable TV engineers and technicians meet these requirements, and help solve this new set of problems.
e-mail: firstname.lastname@example.org 
|About the author|
|Linc Reed-Nickerson is product development manager, TV/Communications Test Business Unit, Tektronix Inc.|
|For more information about reverse path testing, see Tektronix Application Note 21W-11671-0, "Troubleshooting Reverse Path Interference Using the RFM151."|