Co-ax cable quality - how much does it matter?
In June last year, in the 'letters' column, I raised the subject of low-grade aerial and satellite download cables and gave examples of the sort of reception defects they can cause. Awareness of these problems made me wish I had a valid way of comparing the various cable types in common use.
The effects of inferior cable can be bad enough on a small domestic installation, but much worse - and more expensive to correct - on a distribution system that serves, say, a dozen flats. The TV system will be part of the electrical contractor's work, and will, in turn, will pass on the specialist part of the work to a local TV shop or aerial installer. The electrician will supply and install the TV cables from the head-end or repeater positions down to the outlets.The aerial contractor's job, on a small system like this (or indeed a much bigger one), is to come in near the end of the contract and install the aerial, dish, amplifiers, trunk cables, and everything else necessary to make the system work.
I always specify the cable type needed, and make it plain that if the electricians use poor quality stuff then I can not be held responsible for any problems that might arise. In the world of competitive tending the temptation to cut corners is strong however, so quite often my strictures about cheap co-ax are ignored. Of course, on many occasions I'm not even asked, because I'm not invited on the job until it's too late - after the walls have been plastered and decorated. Now, I'm never in any doubt that I'm going to make a fuss when this sort of thing happens - the only question is, how much fuss?It is one thing to have a vague feeling that signal levels at the farthest outlets from the head-end are likely to be 'a bit low' - but that sort of vagueness is not enough when you are threatening to withdraw all relevant warranties.
I realized some time ago that I needed a quantitative measure of the performance of different types of cable. Since the manufacturers of the 'budget' cables do not publish figures, I decided to perform a few simple tests to find out just how much the performance of the various cable types differs. Even if you never install anything more complicated than aerial feeding one TV set, I think the results will be of interest to you.
I have divided the commonly used cables into four groups: A, B, C, and D. All the cables are 75Ω 'downlead' types with an outside diameter of between 6.6 and 7mm.
From top to bottom, Type ‘A’: Semi-airspaced dielectric with copper tape and copper braid., Type ‘B’: Foam dielectric with copper tape and copper braid., Type ‘C’: dielectric with ‘silver paper’ wrap and copper braid.
Type ‘D’: Semi-airspaced dielectric with copper braid.
From top to bottom (fig 1) they are:
Type ‘A’: Semi-airspaced dielectric with copper tape and copper braid.
Commonly known as ‘copper on copper’, the best-known example of this class of cable is Raydex CT100. Other makes are Cavel QC100 and Hycomm HYC100. The designation ‘CT100’ is often taken in vain, no doubt to the great annoyance of Radex. Be aware that some cables sold as ‘CT100-type’ are nothing like the genuine article.
The expression ‘semi-airspaced’ refers to the construction of the white dielectric, and means that it has ‘cells’ (holes) running through it longitudinally. All semi-airspaced cables seem to have five cells (see fig 2). Copper prices move up and down all the time, but at the time of writing this cable should cost no more than £18 + VAT per 100m. Confusingly, this cable is sometimes incorrectly called ‘double screened’.
Type ‘B’: Foam dielectric with copper tape and copper braid.
These cables are identical to type ‘A’ products, except that they have a foam rather than semi-airspaced dielectric. Semi-airspaced cables can deform quite easily if mishandled, and for that reason foam cables, which are more robust when bent and crushed, are making a comeback. It is very difficult to fit semi-airspaced cable into a backbox without it kinking, so the re-introduction of foam cables looks like a good idea. The foam cables of yesteryear absorbed atmospheric and other moisture very readily, causing severe performance degradation, but the manufacturers assure us that the modern products are free of this defect. This type of cable is similar in price to the type ‘A’ ones, or perhaps slightly cheaper. The many different products available include Webro WF100 and Cavel QF100.
Type ‘C’: Semi-airspaced dielectric with ‘silver paper’ wrap and copper braid.
This cable is sold as ‘satellite downlead’, but Sky forbid its use on their installations. The screen consists of a transparent plastic wrap with a microscopically thin layer of silver coloured material bonded to it, and a very low-density copper braid. One peculiarity of these cables is that the dielectric will slide along very easily inside the screen. If the cable has been stretched slightly during installation, the inner core and dielectric can retract out of the ‘f’ plug shortly afterwards, causing severe installer confusion! These cables sell for around £8.50 + VAT per 100m.
Type ‘D’: Semi-airspaced dielectric with copper braid.
This cable has no foil wrap. It is commonly called ‘low loss’, a designation that originated in the early days of UHF transmissions, to distinguish it from the smaller diameter, solid dielectric cables used for VHF. It usually has a brown outer sheath, although the DIY sheds stock it in white. Over the years the braid density of ‘low loss’ has decreased mysteriously. In 1969 it was quite a job to unravel the braid when fitting a coax plug; now there’s hardly anything to unravel! Some manufacturers still produce this type of cable with braid coverage as high as 60%, but these products are rarely used. Much more common are the cheap versions with braid coverage of as little as 20%.
No manufacturer has ever pretended that this cable is suitable for satellite use, but this is persistently ignored by builders, and even by Bodgitt & Scarper Aerials and others of that like. What the motive is I don’t know, because these cables are generally no cheaper than the type ‘C’ ones. This cable is the site electricians’ favourite. Left to their own devices, this is what some of them will use for everything – UHF, satellite, surveillance cameras, dog leads, the lot. This cable is almost universally used for built-in downleads in new housing, where individual aerials will be fitted.
Top left: Foam dielectric (type ‘B’), Top right: Semi-airspaced (type ‘D’), Bottom left: Semi-airspaced (type ‘C’)
Bottom right: Semi-airspaced (type ‘A’)
I tested types A and B first, and was relieved to find that my results corresponded pretty closely with the various manufacturers’ figures. This suggested that my experimental method was valid. The method, in fact, was very simple. I laid out exactly 50m of each cable, making sure there were no kinks or sharp bends. I used a Vision modulator (from Satellite Solutions) as a signal source at UHF, and Sky digital transponders at satellite IF. I checked both of these sources at regular intervals during the tests to make sure that there was no variation. The measuring instruments were recently calibrated spectrum analysers. Fig 3 shows the results.
Signal attenuation per unit length: comparison of cable types
Cable types A and B performed almost identically, so I have shown both as one line on fig 3. Losses climb to 18dB per 100m at the top of the UHF band, and 33dB at the top of the satellite IF band. Cable type C is significantly more lossy at 24.5dB and 44.5dB for the same frequencies.
There are many different ‘type D’ cables, and frankly you only have to look at some of them to see that they are about as much use as wet string. For these tests I used one of the better products. Even so Type D comes in at 32.5dB for top UHF and a massive 66dB—double the figure for types A and B—for the top of the satellite IF band.
What do these figures mean in practice? If we leave aside questions of cable deterioration with age (of which more later) probably not all that much where cable runs are short. The problems will arise where cable runs are longer than average, and where signal levels or carrier to noise ratios are marginal to start with. Take as an example the following scenario. A wideband UHF aerial feeds a simple domestic distribution amplifier via 10m of cable and one or more of the downleads from the amplifier to the outlets is 25m long. The signals carried include analogue channel 21 transmitted at 500kW and a ‘must have’ digital multiplex on channel 67, transmitted at 10kW. Sounds familiar? If type A or B cable is used the overall loss on channel 67 will be 6.3dB. If type D cable is used the overall loss increases to 11.4dB, and signals on channel 67 will be attenuated 2.5dB more than those on channel 21. This will add to an already very unsatisfactory signal level imbalance, and could increase the chances of digital drop out. Where cable runs are 30m or longer, type ‘D’ cables are quite inadequate for UHF, and of course absolutely hopeless for satellite IF.
Although type C performs significantly better than type D, in my opinion it is so far behind types A and B that it should not be used for good quality UHF or satellite installations. Budget domestic installations, maybe. Cables of this type are sold as ‘satellite’ cable, and the unwary could quite reasonably suppose from this that their performance is good enough even for the more demanding installations.
Incidentally, distribution systems carrying satellite IF have each downlead going back to a polarity switch, and since the switches are normally located together in large groups (fig 4), the downleads are likely to be long. In the case of the block of twelve flats mentioned earlier, all the downleads will run to the one amplifier and switch unit, so some of the cables might easily be 40m in length. Where satellite cable runs exceed 30m, I prefer to use CT125. This is a larger diameter version of CT100.
Inadequately screened cables will both radiate and receive signal. This is a difficult thing to measure properly unless you have an electronics laboratory, but I was able to carry out a simple experiment that gave comparative, though not absolute, figures.
I laid out 50m of each of the cables under test, along with an additional length of cable type D. The latter, the ‘transmit’ cable, was connected to a high level signal source, and the other cables were tested on their ability to receive from it (or, I suppose I should say, their inability to not receive from it!). All five cables were bundled very loosely together with cable ties at 1m intervals, to simulate the sort of proximity that would be found if the cables had been installed in a wall cavity, across a loft, or whatever. The far ends of all the cables were terminated with 75Ω. The test was done on one frequency only: 727MHz.
The crosstalk figures below are simply the differences between the signal level entering the transmit cable and those leaving the receive cables.
Crosstalk from type D
Although the signal sources and measuring instruments were 5m apart, the results for cable types ‘A’ and ‘B’ were, I think, compromised slightly by direct transmission from source to instrument. This is likely to happen with a ratio of 80dB. Slight movement of the connectors caused a fluctuation of a few dB, so for this reason, the figures for cable types ‘A’ and ‘B’ are probably slightly pessimistic. The crosstalk from types ‘C’ and ‘D’ was much more ‘solid’.
I repeated the experiment, but this time with all the cables reduced to 20m. The results were virtually unaltered. I also attempted the experiment using type ‘A’ cable for transmission, but could get no meaningful result from cables ‘A’, ‘B’, or ‘C’.
All this strongly suggests that if all the cables in an installation were types A or B, crosstalk would be unmeasurably small. Type C’s performance is perfectly adequate, but look at type D, returning –29dB! Remember that analogue video needs a signal to noise ratio of at least 46dB.
This simple test confirms what a lot of installers have always suspected. Many and varied are the interference problems than can be cured by replacing cheap coax or flyleads with CT100. Satellite IF leaking into a UHF feeder, computer noise entering the flylead of an adjacent TV set, maintained lighting chargers putting white lines across all the TV screens in the building – the list is endless. Downleads inevitably pass alongside or at least near mains cables, and given terrestrial digital TV’s susceptibility to impulse interference, type ‘D’ cable is simply not suitable.
Coaxial cables deteriorate with age, mostly due to the gradual ingress of moisture. Visible evidence is a yellowing of the dielectric and a dark discolouration of the copper. Even when a cable doesn’t show these signs, its performance may fall off severely over a period of years. Type ‘D’ cables seem to suffer most, probably due to a more permeable outer sheath and the lack of a foil screen that serves as a moisture barrier. TV distribution systems often share ducts and voids with district heating schemes and other plumbing, and type ‘D’ cables in such a humid, damp environment will become astronomically lossy after a few years. The signal losses are much worse at higher frequencies, so if you are quoting for the conversion of a system from Group A analogue to wideband digital, beware!
In my opinion Cable type ‘D’ should not be installed behind the plaster in a new building. I suspect that it picks up moisture as the building dries out, because the deterioration seems to set in very rapidly. This can become a serious problem early in the life of the building. The way technology changes these days, it seems very shortsighted to use anything less than CT100. The cost difference, after all, is at most only a few pounds, and who knows what signals and frequencies we will expect these cables to carry during their lifetime?
Kinks and bends
The characteristic impedance of coaxial cable depends partly on the ratio between the diameter of the inner conductor and the diameter of the screen. If the cable is forced into a tight bend the ratio changes and an impedance ‘bump’ is created. This isn’t the place to go into cable impedance, standing waves and what have you, but take my word for it, impedance bumps are a Bad Thing! The minimum bend radius is usually taken as about ten times the cable diameter, whatever the cable type.
The performance of coaxial cable will suffer if it has been ill-treated during installation, by kinking, forcing into small bends, or crushing. I suppose I should have set up some sort of comparative test in which the different cables were (a) subjected to a pretty violent installation by disgruntled electricians on piece rates, and (b) installed by placid electricians keen on transcendental meditation. But I didn’t. Take it as read: if coax is squashed, kinked, twisted, scorched, or stretched, its performance will suffer.
What happens if a cable is bent repeatedly? This can happen during a difficult installation, or during normal use over a long period. Cable type ‘C’, with its transparent plastic wrap and thin layer of conductive material, suffers badly. Tiny radial cracks appear in the conductive coating, and since the braid is very skimpy, impedance ‘bumps’ are likely. I must stress, though, that repeated flexing of any coaxial cable will cause damage. The copper foil of cable types ‘A’ and ‘B’ can crack, and of course, the inner core of all cables will eventually snap. Special flexible coaxes are available, with seven-strand inners, solid dielectric, and a dense braid of fine wire, but these cables are expensive and ‘lossy’, so are really only suitable for short interconnecting leads.
Crushing and trapping
Cable clips should be the correct size, and cable ties should not be over- tightened. Cables can be accidentally trapped or squashed, especially on a building site. When planning cable routes try to anticipate the actions of ‘other trades’. Cables in lofts should not run where they will fall victim to the plumbers’ size 12 boots, for instance. Clip to the side of timber that will be walked on, not the top. If it seems likely that a cable run will be mistreated in this way, use type B, since the foam dielectric cables are physically tougher than any of the semi-airspaced ones.
Cables to avoid
The list is many and varied. One such cable has no braid, just a few strands of wire running longitudinally, and some sort of shiny (and allegedly conductive) coating attached to the inside of the outer sheath. Apart from the obvious screening deficiencies, it is difficult to make a convincing connection to this cable, and it is very susceptible to kinking and crushing.
Sometimes electricians will use oddments of coax left over from previous jobs, and this is where you are likely to encounter cable designed for baseband video use. These are easy to spot because they have a solid dielectric. They are very lossy at UHF, and utterly hopeless for satellite use. Remember, to the average electrician ‘coax’ is ‘coax’, and if he has half a roll left over from a surveillance camera installation he will use it for TV downleads. Similarly, 50Ω and 93Ω cables can appear. These are useless for TV. Any cable with an overall diameter of less than 6.5mm is highly suspect. Some cables have a type number printed at intervals along their length. Amongst the hundreds of different types of unsuitable coax available, the following are, in my experience, commonly found on building sites masquerading as TV downleads: URM43, URM70, URM76, RG58, RG59, and RG62.
If you encounter an unfamiliar cable, I suggest that you take a sample away and test it. Even a short length, say 10 metres, will show excessive signal loss if compared directly with the same length of CT100. Carry out the test on high UHF channels or satellite IF. Thin braid cover is a sure sign of inadequate screening.
For distribution systems or good quality domestic work use a ‘copper on copper’ cable, with either semi-airspaced or foam dielectric. If the cable will, unavoidably, be forced into tight bends, or might be crushed, use a foam type.
For long runs, especially satellite, consider the use of a larger diameter cable, such as Radex CT125 or Cavel QC125.
For budget domestic work, type ‘C’ cables are probably the best choice. Since these cables cost about the same as ‘low loss’ (type ‘D’) there seems to be no point in using the latter. But bear this in mind: the cost difference between the best cable and the worst is only about £1.50 on a standard domestic aerial job. I have to say that the only cables you will find in the back of my van are types A and B.
Of course, most installers know that good cable is essential for satellite use, and I hope that this article has clarified the differences between ‘good’, ‘not so good’ and ‘bad’ cable. The message hasn’t quite got through to many, though, that UHF also demands good cable. I think our trade should recognize that cable quality is an important issue, particularly for digital reception.
Distribution system head-end, ready for installation. The earth rails at the bottom are the connections for 20 downleads, some of which will be 30m long. The connections on the right are the feeds to the repeaters.