Video cable is kind of like Rodney Dangerfield. It doesn't get any respect, or very little when it does. Cable is always the low man on the totem pole 8/01/1998 8:00 AM Eastern

Video cable is kind of like Rodney Dangerfield. It doesn't get any respect,or very little when it does. Cable is always the low man on the totem polein an installation, behind those wonderful video projectors, touchscreencontrols, lighting systems and racks of audio gear. Yet, the improperselection of video cable-and even connectors-can generate problems withsignal distribution and reproduction.

Right now, you may be thinking that cable is cable, right? As long as itsays 72 V or 75 V, it should be all right. What's the big differencebetween RG59 and RG6? If I use 98% shielded cable, then I will have lesssignal loss than the guy using 80% shield. Better yet, I'll go todouble-shielded cable, and I'll make sure to use gold-plated connectorsthroughout the system; they'll offer less resistance than nickel-platedplugs and jacks.

Well, maybe. Perhaps it would be worth digging a little deeper under theinsulation to see just what makes up a video cable and examine how thingslike characteristic impedance, dielectric, loss and shielding all come intoplay.


Coaxial cable was actually developed around the time of World War II by theAmerican Phenolic Company, later called Amphenol and currently part ofAllied Signal. It was intended to replace open-wire balanced transmissionlines for radio stations, particularly in installations where the antennawires needed to be run behind bulkheads, around metal objects or in areaswhere people could come into contact with a hot wire carrying severalhundred volts of RF energy.

The principle was simple-a center conductor would transport the energywhile an outer conductor formed both a shield and ground return. Thisunbalanced transmission line needed to be fed by a signal source with thesame impedance, and the coax would have to be connected to a load (read:antenna) with the same impedance. Otherwise, the flow of RF energy would beimpeded, and the communications system would not function as planned.

Early coaxial cables had a characteristic impedance of 50 V or 52 V, stillthe standard for the communications industry. Open-wire transmission lines(and balanced antennas) have a characteristic impedance of around 200 V, sobalun (balanced-unbalanced) transformers were needed to match antennas tothis newfangled type of cable and prevent return currents from flowing backdown the shield. These baluns also were wound with 4:1 turns of wire,effecting the impedance match from 200 V to 50 V.

When the first television sets came along, the antenna systems were largelyconnected by 300 V, ribbon-style open wire. Early sets connected directlythrough a pair of terminals (remember antenna terminals?), but when cabletelevision came along, it was necessary to again transform a balanced,open-wire signal to an unbalanced, shielded coaxial connection. Once again,a 4:1 balun was used, but this time the resulting output impedance was 75V. This has become the standard impedance for coaxial transmission linescarrying baseband video and computer video signals.

Why not just use open wire transmission line? For one thing, it hasconsiderably less loss than coaxial cable, especially at higher frequencies(25 MHz to 200 MHz). It is also a lot cheaper than coax. On the other hand,the characteristic impedance of open-wire balanced line changes if it getswet, if ice forms on it or if it passes too close to metallic objects.That, in turn, introduces an impedance "bump", which degrades systemperformance. Coaxial cable is immune to the effects of water and ice ifproperly sealed and can be attached to metal without knowing it is there.

The inner sanctum

The combination of center conductor, dielectric insulation and outer shieldare what determines a coaxial cable's impedance, not its diameter. Thereare small 75 V cables, and there are very large 50 V types. There are evenexotic cables like RG-62 (93 V) to be had in all sizes. Thus, we have verysmall 50 V or 75 V cables for jumpers within a chassis (RG-174 and RG-179),larger diameters for short runs between equipment (RG-58, RG-59 and RG-6)and even larger diameters for long runs (RG-8 and RG-11).

Depending on the frequencies in use, we can also obtain solid,polyethylene, Teflon and even air dielectrics within our chosen cable.What's more, we can spend a few dollars and get 93%, 95% and even 97%shields around the dielectric, and those shields can be tinned copper, barecopper, silver wire or even a combination of wire and foil. We can also getdifferent color insulating jackets, but that's strictly a vanity item andhas no tangible effect on performance.

Does increasing the shield coverage decrease signal coupling and radiationfrom the cable? Absolutely. If you are wiring a boardroom on top of askyscraper in a large city, it is best to use dense shield coverage toreduce the chance of stray RF signals from entering back into your system.A given piece of coaxial cable can act like a relatively inefficientantenna if its length corresponds to a quarter-wave multiple of an unwantedRF signal. If the cable is not properly terminated at both ends, the resultwould be some unusual audio coming through your system.

Does increasing shield coverage decrease signal attenuation through thecable? Not at all. That's strictly a function of the diameter of the centerconductor or conductors and the absorption properties of the dielectricmaterial. At baseband video frequencies, solid plastic dielectric doesn'tabsorb much energy. Go up a 100 MHz or so, however, and bingo; it may startto heat up and dissipate energy like a big resistor.

Different hydrocarbon-based plastics behave better at higher frequencies,including polyethylene and even Teflon, which also has relatively highbreakdown voltage rating. But we eventually reach frequencies high enoughwhere any solid dielectric simply absorbs too much of the energy passingthrough it. At that point, we use a rigid, corrugated transmission linewith an air dielectric and small spacers or flexible air waveguide with nospacers at all. Such lines are often used at microwave relay stations, cellsites and satellite uplinks and downlinks.

In the commercial A-V business, we'll generally use one of the three cablesizes I mentioned earlier. These will also be available in bundles (Triax)and stand-alone cables for passing RF and high-frequency data or inplenum-wire configurations for running through service panels and alongsideHVAC, electrical and plumbing.

About wire

The center conductor of a cable may consist of many small wires or a singlelarge wire. If the cable is going to be flexed or bent, it's always betterto go with a stranded rather than solid-center conductor. Cable resistanceand impedance won't significantly change. Also be aware that there are manyoff-brand cables that may vary considerably in impedance. Although thisisn't a major issue, some cheap cables might be as high as 80 V or lowerthan 70 V.

Two other cable characteristics are worth mentioning. You may see aspecification that reads "contaminating jacket" or "non-contaminatingjacket". Don't worry, you won't get a communicable disease from thesecables. What this term defines is the cable's "groundworthiness", or itsability to be directly buried without using conduit or PVC piping. As arule, coaxial cables with "non-contaminating" labels can be buried in soilwithout further protection, although there's no absolute guarantee theywon't eventually absorb moisture, develop cracks or have small insectstunneling into them. Cables marked contaminating must be encased withinsome sort of water-resistant, bug/rot-resistant conduit before placing themin the ground.

The last characteristic of coaxial cable is its velocity factor, a termthat applies to how fast an AC signal travels through the cable comparedwith its transit in free space (that good old air dielectric). If thevelocity factor of air is one, then any transmission line is going to besomewhat slower, but not so slow that we humans would notice anythingdifferent.

As the velocity factor of a given cable changes, so does the effectiveresonant length of the cable at a given frequency, especially if itterminates into a non-resonant load or if it is used as part of an antennasystem. For most cases, you won't be concerned about velocity factor whenusing coaxial cable as part of a pure 75 V to 75 V single cable connection.However, you'd better be consistent with your cable types when passing Y/Cor YUV/RGBS video signals-don't mix and match cables here.

You'll also see a specification for capacitance per foot, usually inpicofarads (pF). That is because coaxial cable resembles a capacitorphysically and electrically, meaning that it can have reactance in anon-resonant circuit. Coaxial cable can also resemble an inductor at radiofrequencies, and some manufacturers use it to wind high-power baluns and RFchokes for antennae.

How about a plug?

For coaxial cable to be of any use, we've got to use some sort of plug andjack at each end. There are plenty of those to choose from too, butfortunately our industry has pretty much standardized on a handful. The badnews is only one is really a 75 V connector; the other plugs don't evencome close. Let's check them out:

BNC connectors are generally called "professional" video connectors in theindustry. Frankly, it ought to be the only video connector we ever use. TheBNC connector was first developed as a water-resistant, positive-locking,constant-impedance RF connector for the U.S. Navy after WWII. Its initialsstand for Bayonet Naval Connector, and it actually resembles a section oftransmission line with a locking sleeve on it.

The distance between the center pin (usually silver, often gold-plated) andthe outer shield is consistent, and the connector also has an airdielectric. It's a secure plug, which makes a secure connection, and itdoesn't even show up on sophisticated cable impedance-measuring deviceslike time-domain reflectometers. BNCs are generally good up to 2 GHz-wellabove the bandwidth we'll use in the foreseeable future.

RCA connectors first appeared in the early 1950s and were used to connectthe outputs of stand-alone record players to hi-fi systems. (Rememberhi-fi?) RCA plugs are the complete antithesis of BNC plugs-cheap,unreliable connectors with no specific impedance, no positive locking, andthey're inexpensive as well. (Well, one out of three isn't bad.) Cheap ornot, RCA plugs and jacks have been around so long that they've evenoutlived the company that invented them.

Today, we have numerous varieties of RCAs on the market withnickel-plating, gold-plating, stronger center pins, threaded barrels,better "grip" barrels and even cable strain reliefs. Well, you know whatthey say about changing a zebra's stripes. RCA connectors may be okay asterminating plugs and jacks for Hi-Z audio cables, but their use as videoconnectors simply reflects a desire on the part of the manufacturer to savea few dollars. Gold-plating them is even more ridiculous. Use BNCs whereyou can, you'll be happier for it.

Type "F" connectors may have actually wrestled away the title of "World'sCheapest Connector" from the lowly RCA, except that no one's gold-platingthem yet. These marvels were designed as a fast, cheap way for cablecompanies to install drops to customer's homes, and they can't even besoldered as they use the cable's center wire as the conducting pin. Ifanything, they may be more positive-locking than an RCA plug, simplybecause they thread on to the jack.

As long as you don't repeatedly flex an "F" connector, they're prettyhappy. But their characteristic impedance varies quite a bit around 75 V,mainly because part of the connector is deformed when crimped to the cableinside. "F" conectors are fine for bringing in external RF-modulatedsignals from antennas and CATV/MATV lines, but don't use them forprofessional video installations.

DIN connectors are another example of a consumer-level connector that hasmigrated into the professional world without any good rhyme or reason. DINjacks and plugs have four pins and no consistent spacing or dielectric,making them another non-constant-impedance connector. Many "S" cablesthemselves have high loss figures; the generic 6 foot (1.8 m) S Videojumpers that come with DVD players and desktop projectors are particularlyprone to loss.

For long cable runs, a pair of matched coaxial cables with BNC plugs wouldbe a much better choice, assuming your DA or switchers have BNC jacks thatcan be enabled for Y/C output. Otherwise, look for premium S-Video cableand hope for the best.

You may encounter some other connectors from time to time. One very commonplug for wireless microphone systems is the PL-259/SO-239 combination,otherwise known as the UHF plug. This is a total misnomer; PL-259s are noteven close to constant-impedance connectors, but they became widespread inthe communications industry in the 1950s and 1960s. They're still made(some even have silver-plated center conductors) and are okay for use below30 MHz.

You may even encounter the odd-looking TNC plug, which is nothing more thana threaded BNC and retains all of the BNC's electrical performance. Thelarger brother to the BNC is the type N plug, which will work with 1/4 inch(6.4 mm), 3/8 inch (9.5 mm) and 1/2 inch (12.7 mm) coaxial cables. Nconnectors are also constant-impedance plugs and can handle a lot morepower than BNCs or TNCs, plus they can be used up to 3 GHz. Somemanufacturers of UHF wireless mic systems are employing type N connectorsfor antenna inputs.

Terminally yours

Coaxial cable works best and is happiest when both the video source and thedelivery point (projector, monitor, signal processor, etc.) look like 75 V.At this point, just about all of the signal is being delivered where itshould be. If, however, the projector, display or signal processor/switcherdoesn't look like 75 V, some of that signal will be reflected back down thecable as standing waves.

This phenomenon can lead to signal phasing problems, creating color errorsand even sync problems. The picture may appear too bright or becomeunstable, which is all the more reason to make sure that the correcttermination is used at either end of the cable. By the way, you can use a50 V BNC on a 75 V system in a pinch. The mismatch amounts to about 1.5:1or 4% reflected power, hardly anything to get upset about.

Now that we've reviewed cables and connectors, a couple of myths should bedispelled. First, gold-plating a connector does very little to improve itsperformance in a given video signal distribution system. The resistance ofgold is indeed lower than nickel or tin-plating, but that's a differenceyou won't be able to measure without the most sensitive of volt-ohmmeters.

Where you may see a long-term difference is a video connection in a hostileenvironment, such as one with lots of dust, smoke, corrosives or moisture.The truth is, properly plated connectors shouldn't oxidize substantially innormal use. Silver does oxidize (needs protection) as does tin.Nickel-steel alloys hold up very well with time. Take a look at thoseRCA-to-BNC adapters you've been lugging around for years. See muchoxidation on them? I've got several Radio Shack, garden-variety RCA/BNCadapters that have at least ten years of staging mileage on them; nocorrosion is evident.

Of course, gold is a very stable metal with excellent conductivity. Itsbenefits are better realized in circuits where the resistance or impedanceis very low to begin with, such as on the tabs of a power transistor in anaudio amp or in speaker terminals. A change of only a few tenths of an ohmin these circuits can amount to a substantial percentage of the systemimpedance. On the other hand, a change of 0.1 V in resistance is wellwithin the tolerances of the coaxial cables used in a video signaldistribution system and isn't even worth worrying about.

The other myth seems to have its roots in the broadcast video industry andstates that component (YUV) video signal distribution is very difficultbecause the cables must be exactly the same length. Well, we've beenporting around 3-wire, 4-wire and 5-wire RGB computer video signals forsome time without serious consequences. How hard could it be?

Baseband video has a color burst frequency around 3.6 MHz and a bandwidthof 6 MHz. An electrical wavelength at that frequency is about 84 m long(83.79 to be exact). So an error of a couple inches at that frequencyamounts to .0006% of a wavelength, hardly enough to introduce a serioussync or color phasing error. Both of these are far more likely to occur inimproperly biased and phased distribution/buffer amps.

If you are running YUV signals in your system, make sure you usegood-quality cable with a known velocity factor. Go for the best connectorsyou can install, although even a lowly RCA plug won't introduce much in theway of phasing errors. Be sure your active distribution and switchingelectronics are properly phased and have sufficient bandwidth, and check tosee that every cable is looking into a 75 V load.

Lastly, don't scrimp on quality cable and connectors. The additional coastis generally minimal, but reliability will be increased considerably. Buildenough bandwidth into the system for future expansion; if you're cablingfor video and Internet graphics, make sure the system can handle XGA andHDTV as well. For a workstation/boardroom install, use cable that willhandle 150 MHz to 200 MHz of signal before showing appreciable loss.

Want to read more stories like this?
Get our Free Newsletter Here!
Past Issues
November 2017

October 2017

September 2017

August 2017

July 2017

June 2017

May 2017

April 2017