This case history shows how decibel knowledge, a handheld oscilloscope, and intuitive thinking can be used to solve a troublesome problem.
Now that we have a somewhat complete understanding of the decibel and how it applies to various types of signals, we can apply this knowledge, along with the many modern test instruments available today, in troubleshooting any electronic equipment operation problems that may come our way.
As an example of how this may work, let's review a case history involving closed circuit television, coaxial cable, and shoddy workmanship.
Background site information
We were called in by a client who complained about a very poor picture generated by a security camera on a closed circuit television (CCTV) video monitor on a simple point-to-point link. The picture had degraded from good to unusable over a short period of time and really got bad in the few days following a recent lightning and rainstorm. The general opinion was that lightning had somehow caused the problem.
Upon inspection, we found that the CCTV system consisted of National Television System Committee (NTSC) standard video generated and displayed in black-and-white. The main monitoring point of the CCTV system was a loading dock and a weather-proofed housed camera located on a 10-ft high post mounted atop the building's roof was used to monitor that location. The camera was connected to a video monitor in a security shack some 200 ft away via a 75 ohm coaxial cable, which was routed down into the shack by means of a vent pipe type of entry. Power to the camera was provided via a low-voltage DC link on another coaxial cable, which was installed right alongside the coaxial cable used to transport the video signal. Both the monitor and camera low-voltage DC supplies were simply plugged into a wall outlet convenient to the operator at the guard's shack.
Symptom
The picture on the video monitor looked as though someone had turned the contrast control all the way in one direction so that there was no contrast at all; the picture looked washed-out and was barely visible on the screen, which was a nearly uniform light gray.
Testing procedures carried out
Our first task was to go up to the camera at the roof and see what the video signal looked like as it exited the camera. This test was much aided by the fact that our handheld, 50 MHz bandwidth, solid state, digitizing oscilloscope with LCD display had an internal battery pack and did not require any AC power for operation.
First test. First, the coaxial cable was disconnected at the camera and a BNC style "TEE" fitting was installed. This fitting was equipped with a 75 ohm terminator resistor on one leg. Then, we connected our handheld oscilloscope into the remaining open end of the TEE. The result, as shown in Fig. 1, was a healthy NTSC composite video signal. Conclusion: the camera was clearly putting out a good signal, which was about 1.8V peak-to-peak across the 75 ohm termination. (There is also a DC component with the AC video signal.)
We then placed our handheld oscilloscope into its meter-mode and the above signal at the camera into the 75 ohm load was taken as a zero dB reference and stored into memory. This is shown in Fig. 2, where + 000.1 dBV DC is taken as being close enough to zero to do the job. Now, the "good" signal right out of the camera was available to be used again and again as a comparison with signals we would measure at different locations. We then would be able to see how much loss of signal occurred along the path, all of which was supposed to be a consistent 75 ohm.
The TEE was removed and the 75 ohm coaxial cable was reconnected.
Second test. The next test was made at the video monitor end of the cable and right at the point where the cable was connected to the monitor. Again, the TEE was used, but this time no 75 ohm termination resistor was used with it since the TEE was attached to both the monitor and the cable. Thus, there was a fairly good 75 ohm load on everything. The result of this test was that almost no video signal could be seen on our handheld oscilloscope's screen.
We then changed the oscilloscope's vertical scale from 500 mV/cm to 100 mV/cm and another measurement was taken, which is shown in Fig. 3. As you can see, the video signal is simply attenuated but does not appear to be distorted in any way that is easy to see. Conclusion: the signal loss was occurring along the 75 ohm cable path, or was it?
Video monitors have been seen to "load down" a signal due to an internal failure on its input circuit; as such, we didn't want to rule this possibility out. A quick test with the TEE and the 75 ohm termination resistor in place of the video monitor quickly ruled out this possibility; the signal was essentially unchanged from that shown in Fig. 3. Now, we really could conclude that the signal's loss was occurring along the 75 ohm cable.
Third test. We next placed our handheld oscilloscope into its meter-mode, while maintaining the connection to the TEE at the junction of the cable and video monitor. This allowed us to take a relative dB measurement reading, as shown in Fig. 4, using the original zero-level as the reference. (Remember, we did this at the camera end to establish a comparison reference.) Here we see that a - 13.7 dBV DC loss exists. This loss represents a voltage loss ratio of 4.84:1, or a signal loss of nearly 5V for every volt put into the cable!
How much signal attenuation should you expect on a 200-ft long, 75 ohm coaxial cable? A quick look at the coaxial cable manufacturer's Master Catalog gave us the approximate answer: around 2 dB of loss at 10 MHz for 200 ft of RG-59/U type cable as used in CATV applications. The whole attenuation chart is shown in the accompanying table below.
What we were seeing in this path was more than 11 dB loss over and above that stated in the manufacturer's literature. Also, the baseband video we were looking at shouldn't have a lot of really high-frequency in it; thus, the cable probably shouldn't attenuate as much as 2 dB (per manufacturer's literature) for 200 ft in any case.
Oh yes, since the manufacturer's information was provided only in dB form, what would we have done if we didn't understand dB and weren't working in terms of dB on our handheld oscilloscope? You guessed it. We would have had no idea what was "normal" and what was not on a coaxial cable run of the type being investigated. All we would have had was some guesswork, which is not a very good way to go in most cases.
Further analysis
What was happening on the cable? The BNC connector at the video monitor end was inspected and it looked OK, except that it seemed to be a little wet after it was handled and the cable was flexed.
Following this the same examination of the BNC at the camera end also failed to show any problems. We also made sure that the connections were well protected from the environment by the camera's enclosure.
Was the moisture a clue? Was it significant, or not? Past experience with coaxial cables with water inside of them showed that this condition caused severe signal attenuation.
Back to the rooftop we went to make a closer examination of the 75 ohm cable and its route back to the video monitor. First, we looked at the vent pipe, the rooftop penetration through which the cable was passed. We found that it was not equipped with a weatherhead and that the cable was simply stuck down into it from its open top. Sealing was done with some kind of putty or caulking material and it looked as though it was really dried-out. Thus, water (from the storm, remember?) could follow the cable's sheath down into the building around the bad seal.
But how did this condition let the water into the cable? We pulled the sealing material out of the vent pipe and then hauled the coaxial cable up out of it. About 10 feet down, we found a connection made up of two BNC fittings and a male-male adapter. The end of the cable going into the bottom BNC fitting from the building was mostly pulled out of the connector and the braid/sheath was fully exposed to any water flowing down the cable from above. In fact, the arrangement was a pretty good funnel for the water to flow into the cable between the outer sheath/shield and the inner dielectric material. Corrosion was also rampant in the damaged connector set since it had not been sealed from moisture in any way. Obviously, this was not good for reliable signal transport.
Where did this splice come from? After a little discussion with the personnel, we learned that the camera came from the factory with about 10 ft of cable. Rather than throw this cable away, it was simply kept in place and used by connecting it to the end of the cable being routed from the video monitor. There was no explanation as to why such a poor rooftop penetration was made; nobody would own up to it while we were there.
Solution
The whole existing 75 ohm coaxial cable run was replaced with a continuous length one. Where this cable came from and what its quality was, we didn't know and couldn't find out; it might have been surplus stock from somewhere (World War II?). After installation of the "new" cable, the signal at the video monitor end was again checked with the test TEE and the monitor in place. This signal is shown in Fig. 5. Here we see that there is still some attenuation, but nowhere near as much as before.
Again, using our handheld oscilloscope in its meter-mode, we made a relative dB measurement reading using the original zero-level as the reference. The "new" cable's signal loss, as shown in Fig. 6, is about -4.5 dB. Compared to the previous dB measurement readings [ILLUSTRATION FOR FIGURES 2 AND 4 OMITTED], this amount of signal loss is acceptable in this application, as was evidenced by the good picture on the video monitor.
ARTIFAX
EC&M article:
"An Introduction To The Decibel," July 1995. Cast: Article cost $9.95. Order No. 2258. Orders are taken via facsimile machines only. To order by fax dial 800-234-5709. (Have a credit card and your fax number ready when you dial by fax.)
Warren H. Lewis is President of Lewis Consulting Services, Inc., San Juan Capistrano, Calif. and Honorary Chairman of EC&M's Harmonics and Power Quality Steering Committee.
