Channel planning for distribution systems, and the n+9 and n+5 conundrums explained

When planning a distribution system you should, if possible, avoid having signals on channels five or nine channels apart. Thanks to modern TV receiver design the problems caused by these channel combinations are far less important than they used to be, but nevertheless it’s worth avoiding these channel spacings if you can, especially if the system will feed any older TV sets. The channel plan for systems that carry both analogue and digital signals should also avoid these channel relationships where possible, treating both analogue and digital signals as potential interferers. A useful chart for plotting your channel usage can be found here.

On a system using DSB (double sideband modulators) for in-house analogue channels a common practice is to use every second channel. This automatically avoids n+5 and n+9, except that DSB modulators produce a lower sideband, which is actually a mirror of the output channel and is centred on the vision carrier frequency. This sideband extends some way across the lower adjacent channel, so a system using every second channel for a DSB modulator effectively has something on every channel, making n+9 interference a possibility. However, the channel filters that should always be used to combine modulator outputs reduce these unwanted lower sidebands somewhat, so the problem is greatly reduced in practice. If the system also carries a standard broadcast group of analogue signals the sequence of n, n+2, n+4, n+6 etc will probably have n+5 or n+9 clashes with one or more of the broadcast signals. This means that a channel in the ‘every second channel’ sequence might have to be left unused.

Years ago it was out of the question to use adjacent channels, but with modern TV sets a system using VSB modulators (which suppress the unwanted lower sideband) for the in-house channels will probably use every channel (n, n+1, n+2, n+3, etc). Such a system will obviously have n+5 and n+9 channel clashes, but it can work perfectly well as long as the TV sets are modern and are free from vices, such as radiating their IF excessively or being susceptible to interference from signals nine channels above the one being viewed. It’s always interesting when adjacent channel working is imposed on an existing system that hasn’t had it before, because it sometimes happens that certain quite respectable looking tellys will be quite unable to cope, whereas some supermarket cheapies will be OK.  Generally though, the Panasonics and Sonys and similar sets will handle the adjacent channel input best.

The cause of the n+5 problem

The cause of the n+5 problem is that the local oscillator of a TV receiver runs at 39MHz above the channel it is tuned to. 39MHz roughly equals five channels, so if the local oscillator radiates it can cause interference to a nearby set that is tuned to a channel five above the one received by the first set.  

The cause of the n+9 problem

The n+9 problem is sometimes called ‘image’ interference. Ray Cooper wrote — in his inimitable style — a rather good explanation of the phenomenon, which I reproduce below.

The theory is simple enough: it's the practice that's a bit of a bummer.

Consider this. Suppose that you're receiving a signal on channel N. The local oscillator in the receiver will be running at a frequency on the high side of this, in order to give an IF frequency of 38.9MHz for the vision IF frequency. Since the local oscillator is above the wanted signal, you end up with a mirror-image of the original UHF one, and the sound IF therefore ends up at 32.9MHz, in round figures. Suppose, for the sake of argument, that N is channel 40, which is V = 623.25MHz,, S = 629.25MHz in round figures. The local oscillator will therefore have to run at 623.25 + 38.9 = 662.15MHz. Sound IF will then be 662.15 - 629.25 = 32.9MHz.

Okay so far. But suppose that there's also a strong signal on channel N+9 to be had, and suppose further that this signal gets as far as the mixer in your front-end. The mixer isn't proud - it will quite cheerfully convert this signal also into one in the IF passband, though this time, since the local oscillator is below the (unwanted) signal, no such mirror effect takes place. With our suppose-channel of 40, N+9 becomes Ch.49, V = 695.25MHz, S = 701.25. Our L.O. at 662.15 gets to work on this one too, and produces an IF of V = 695.25 - 662.15 = 33.1 MHz, S = 701.25 - 662.15 = 39.1MHz.

So you see that the IF produced is an unhealthy mix of a wanted signal and an unwanted, mirrored signal all occupying much the same spectrum. The detector in the receiver is quite unable to separate the two, and the end-results depend on how much rejection of the unwanted signal the tuner front-end can impose. If it has a decent amount of rejection, the unwanted signal ends up being small, and the end-result on the picture may merely be a nagging feeling in the back of one's mind that the picture isn't quite as tidy as it might be. If the rejection is poor, the picture can be ripped apart by patterning, and quite unusable. The suspicious character in the woodpile is probably the unwanted N+9 sound carrier, lying as it does only 200kHz from the wanted vision carrier, and heavily frequency-modulated to boot. There wouldn't be any way of filtering it out after the mixer that wouldn't completely wreck the vision signal. The wanted sound signal would also be severely compromised by having an unwanted vision signal, again only 200kHz away.

So what's the cure? Make sure that you don't have any N+9 signals in the four-channel combination from your local station, is a good start.  Unfortunately, the N+9 signal may be coming in from some other station.
If there are N+9 signals, the tuner front-end rejection of unwanted signals wants to be as good as can be.

It's a bind with domestic TVs, of course: but potentially it also affects transposers. However any decent transposer worth the name will have a channel-filter right at its input, which will only pass the wanted channel (plus bits of the two adjacent channels, unless you have a rather special filter). So the image-problem doesn't occur here. There is no reason why a filter which rejects an unwanted channel shouldn't be used with a domestic TV, so long as N+9 isn't a channel you want to watch. If it is, you'd have to make the filter switchable, which would be a bind of the first magnitude.

I invite others to enlarge upon these efforts (as the Railway Chairman said, having cut the first sod...)

There endeth the words of Ray Cooper. Thanks for that Ray.