Techie Stuff

Introduction

Previous articles describe analogue [link here] and digital [link here] terrestrial television from the Oxford viewer's perspective. The purpose of this article is to give technical details of our antenna system that might be helpful to early adopters of the new digital terrestrial (DTT) system who are experiencing reception problems. A list of channels has also been drawn up [link here].

The original system - analogue reception from four regions

Until the recent advent of DTT, the system was optimised for analogue reception of the following:
 

Transmitter	Programmes and channels	Distance	Antenna	Amplifier and filters	Comments
Crystal Palace	BBC-1 London (formerly South-East) 26 Carlton Television London Weekend Television 23	56 miles	Antiference Extragain XG21 (23-bay multidirector) 19 dBD gain, Group A 30 feet high, external	Group A bandpass filter at input; Research Communications GaAsFET (25 dB gain, 0·4 dB noise figure, channels 23-33); channel 33 bandstop filter at output	Extreme fringe; adverse local topology; CCIR Grade 1-3
Hannington	BBC-1 South 39 Meridian Television 42	32 miles	J-Beam Multibeam MBM46 (13-bay multidirector) 17 dBD gain, Group B 19 feet high, internal	Labgear CM7068 (28 dB gain, 4 dB noise figure, Group B); channels 39 and 42 bandpass filter at output	Outer service; good topology; CCIR Grade 4-5
Oxford	BBC-1 South (Oxford) (formerly South-East) 57 BBC-2 63 Carlton Central South 60 Channel 4 53 Channel 5 49 The Oxford Channel 47	2·8 miles	J-Beam Logbeam LBM2 (15-bay log periodic) 11·5 dBD gain, wideband 19 feet high, internal	None	Line-of-sight; CCIR Grade 5. A set-top antenna is almost sufficient.
Sandy Heath	Anglia Television 24	49 miles	Triax Unix 100 (24-bay multidirector) 19 dBD gain, Group A 17 feet high, internal	Channel 24 bandpass filter at input; Research Communications GaAsFET (25 dB gain, 0·5 dB noise figure, wideband)	Fringe; good topology; CCIR Grade 2-3

Key: CCIR Grades: subjective reception quality (5: perfect; 1: poor). dBD: forward gain relative to half-wave dipole.

The four signals are combined in a fairly Heath-Robinson fashion, as follows. The Crystal Palace and Sandy Heath outputs are combined with a low-loss 'Y'; this output is combined with the Hannington output using a Group A : E diplexer (21-37 : 35-68) followed by a Labgear set-back amplifier (10 dB gain, wideband); this output is combined with the Oxford output using a Group K : C/D diplexer (21-48 : 48-68); the final output passes to a Labgear multi-port distribution amplifier fed also with VHF/FM.

Most of the other channels from the listed transmitters can also be received, although sundry alterations to the system would obviously be required for optimum performance. We do not necessarily recommend that others follow the above prescription, which evolved during experiments including other transmitters and which uses a zoo of components accumulated over time.

The Crystal Palace signal is exceptionally poor because of a nearby hill on the line of sight; a large, high external array with a special head amplifier proved essential. A grouped input filter was necessary because of breakthrough from the local Beckley signal. The Sandy Heath arrangement is similar, although signal strength is higher and we have resisted external mounting at present. Although there is no appreciable adjacent-channel interference between Carlton/LWT (channel 23) and Anglia (channel 24), the channel 33 bandstop filter prevents occasional (n+9) interference to Anglia by Crystal Palace BBC-2 during unusual weather conditions.

The Hannington array picks up strong backscatter from Beckley Channel 4 and Channel 5; to avoid ghosts, we have used pass filters for the two channels required; the filters could be placed after the head amplifier (for optimal signal-to-noise ratio) without it being swamped. The set-back amplifier included in the spaghetti makes up losses in combining the signals; it is an inelegant solution for lack of gain elsewhere.

Migrating to DTT

With a DTT receiver (Philips DTX 6370), good reception of all six current local multiplexes (then channels 34, 48, 52, 56, 67, 68) plus Hannington Digital 3 & 4 (channel 43) and SDN (channel 40) was available. [Note added 15 September 2000: some of the multiplexes and radiated powers have recently been changed; the table below has been amended accordingly]. Additional Hannington multiplexes could be obtained by altering the current filter arrangements; 40 and 43 managed to pass through the existing 39/42 filter anyway.

When setting up, the receiver scans upward through channels 21 to 68, listing the contents of usable multiplexes in its menu according to a predefined series of slot numbers (where possible). Consequently, when more than one transmitter is available, there is a dilemma when a second multiplex of the same type (i.e. containing items destined for the same slots) is encountered during the scan. With the SDN multiplex, the second seems to be ignored (even if of higher signal strength), whereas items from the Digital 3 & 4 multiplexes are given additional entries in vacant menu slots (in our case, thereby providing a choice between Carlton Central South and Meridian Television). When 'adding channels' to an existing selection, entries are replaced as they are encountered in the upward scan (with the first SDN multiplex encountered being retained, as previously): for instance, if the Oxford selection had already been stored when Hannington was added, Meridian (43) would replace Carlton Central South (68) in slot 3 and Carlton Central South would appear in a higher, vacant slot (and ITV-2, Channel 4, Film Four etc would have duplicated entries too) whilst the Hannington SDN channels (40) would oust the stronger Oxford ones (56). To prevent the latter, we used a channel 40 bandstop filter whilst setting up.
 

[Note added 21 January 2000: more recently, the storage behaviour in the presence of a duplicate of Multiplex 1 (BBC), which previously resulted in channels from the multiplex on the lower UHF channel being stored and those from the higher being discarded, has since changed so that, as with Multiplex 2 (Digital 3 & 4), channels from the higher are stored in vacant slots. This has enabled the subsequent differentiation in programming between the various BBC regions. Storage for channels in the SDN and other multiplexes still gives priority to the multiplex of each type which has the lowest UHF channel number, and discards the duplicates.]


In fact, perfect reception of the Oxford multiplexes was normally possible using a set-top antenna (distance 2·8 miles) except during wet and windy conditions when minor disturbances occurred, probably caused by intervening foliage. Most existing installations will have Group C/D antennae; the BBC multiplex on channel 34 is outside the group, but the gain of a typical C/D Yagi would only be about 4-7 dB down at channel 34 and that multiplex has a 10 kW ERP rather than the 1-3 kW of the other (in-band) five. Consequently, a viewer migrating to DTT using an existing C/D antenna might expect the BBC multiplex to appear at a similar signal strength to the others.

With our four-antenna system, impulsive interference was at first a problem on the Hannington multiplexes and, very occasionally, on the local ones too. Ignition interference caused momentary freezing, blocking or loss of signal when certain vehicles passed through the high-gain Group A beams; also, mains-borne switching transients caused visible minor disturbances. The interference could be eliminated by switching off the Group A amplifiers (with the loss of the corresponding analogue channels, of course), by minor re-siting of the Oxford antenna (our Hannington beam does not cross any nearby road, fortunately) and by dressing the feeders well away from mains wiring. Mains filters to the amplifiers or to the receivers did not appear to make any difference. We conclude that, although overloading was not a problem (as verified by experiments with attenuators), the high-gain Group A systems were introducing disruptive interference despite channel filtering (of course, simultaneous disruptions occur to the weak Group A analogue signals too). We are considering re-siting the Hannington array outside for an increased head signal: we estimate that the additional 6 dB or so is likely to eliminate our self-inflicted problem.

The new system so far - digital reception from two regions

The recipe for best results is described in the following table.
 

Transmitter	Multiplexes and channels	Distance	Antenna	Amplifiers and filters	Comments
Hannington	BBC South (1) 50 Digital 3 & 4 (Meridian Television) (2) 43	32 miles	J-Beam Multibeam MBM46 (13-bay multidirector) 17 dBD gain, Group B 19 feet high, internal	Labgear CM7068 (28 dB gain, 4 dB noise figure, Group B); channel 43-50 bandpass filter at output	Fringe; good topology; CCIR Grade 5
Oxford	BBC South (Oxford) (formerly South-East) (1) 34 Digital 3 & 4 (Carlton Central South) (2) 68 ITV Digital 1 (B) 52 ITV Digital 2 (C) 48 ITV Digital 3 (D) 29 (formerly 67) SDN (A) 51 (formerly 56)	2·8 miles	J-Beam Logbeam LBM2 (15-bay log periodic) 11·5 dBD gain, wideband 19 feet high, internal	None	Line-of-sight; CCIR Grade 5. A set-top antenna is almost sufficient.

Key: CCIR Grades: subjective reception quality (5: perfect; 1: poor). dBD: forward gain relative to half-wave dipole.

The two signals are combined with a low-loss 'Y'; the output passes to a Labgear multi-port distribution amplifier fed also with VHF/FM. As explained previously, additional filtering would be required to retain dual-region analogue capability. When combining an 'open' antenna directly with another amplified source in this way, care must be taken to avoid feedback and, of course, re-radiation.

Other Hannington multiplexes can also be received, although they are currently blocked out to prevent the receivers from tuning to them in preference to the local signals when setting-up; a channel 50 bandpass will be added when BBC-1 opt-outs begin, in the hope of adding BBC-1 South to the selection [Note added 21 January 2000: this has now been done]. We are using a Group B antenna for Hannington rather than the correct Group W (wideband) simply because it was to hand and covers the channels presently required.

Although the two-antenna configuration described works well, our original four-antenna analogue configuration is being retained in order to continue receiving Anglia, Carlton/LWT and the BBC regional opt-outs until further experimentation has enabled us to receive DTT reliably from Crystal Palace and Sandy Heath (at present, the Crystal Palace signals just limp into the 'poor' zone occasionally during tropospheric lifts). As explained above, this does render the Hannington DTT signal vulnerable to interference picked up by other antennae.

General comments about high-performance installations

The usual precepts of good UHF installation practice apply: some points are listed here.

Antennae

Antennae should be generous, mounted securely, well clear from each other and from metalwork. Ideally, they should be mounted externally, although they will deteriorate far more quickly than within the roof space - we have opted for higher-gain models placed internally where possible (some are still in pristine condition after 20-30 years, although quaint). Mountings and cables should not encroach on the element train.

Cables and connexions

It is important to use good-quality 75 ohm co-axial cable throughout, including flyleads. Currently-available double-screened satellite cable (CT-100 and CT-125 and their variants) with its low attenuation and good screening properties is perfect for DTT - when purchasing, just check that it is not an earlier 50-ohm type which might cause a mismatch to the standard 75 ohm terminations. Cable runs and drops should be well supported, with the usual care to avoid sharp bends, stretching, crushing and punctures to the sheathing. Cables should be separated from each other (crosstalk occurs readily) and from other wiring and should be as short as practicable. As few connexions as possible should be made; F connectors are preferable to the ancient Belling design; solder pins should be soldered. There are different species of F-connector for CT-100 and for the larger-diameter low-loss CT-125 cables. Application of silicone grease to terminals and, particularly, the exposed internal cable sheath (which is often very hygroscopic) is recommended.

Amplifiers

Any masthead amplifier should be placed at least 1 metre away from the antenna. As few amplifiers should be used as absolutely necessary (having said this, there are no fewer than five in the example above!) and care must be taken over signal levels if overloading and noise are to be avoided (signals can fade up as well as down). Any input filter should be mounted close to the associated amplifier to minimise stray pick-up within the intervening lead.

Channel levelling

The MATV practice of adjusting the level of each channel at the head end provides a flexible solution to the problems of differing multiplex powers and the disparity between the analogue and DTT signal levels, although it is probably too expensive for the average domestic installation. If a spectrum analyser is available, it is useful to check that the COFDM signal is not being distorted by filters. In a domestic context, it might sometimes be easier to use separate installations for analogue and DTT reception.

DTT reception: some points to consider when trying to eliminate occasional pixellation and/or freezing

• Signal level out of range -
• too much: although there might be a really good antenna, try reducing the signal (and the interference too) with an attenuator - ideally a variable type. Note that the higher-powered analogue signals, especially if on adjacent channels, can cause problems;
• too little: ideally, more and/or higher antenna and/or, less ideally, a low-noise amplifier with good overload characteristics;
• Impulsive interference (prime suspects: traffic; domestic appliances) -
• picked up by the antenna (try re-siting - bear in mind that even an external antenna might pick up mains-borne spikes as there is usually quite a lot of wiring in the roof);
• picked up by the downlead (try double-screened coaxial cable (including flyleads), e.g. CT100, if you haven't already used this - and make sure that the antenna has an integral balun (the more basic 'contract' types usually don't, resulting in a mis-match at the cable termination which can reduce the cable's immunity to interference pick-up); and, of course, use good-quality plugs soldered where necessary);
• caused by 'saturation' in a masthead or other amplifier that does not have enough signal overhead to deal with interference cleanly;
• picked up directly from the mains input to the DTT receiver or devices connected to it - some types of mains filters designed as 'surge protectors' can still let through quite a bit of RF even if guaranteed not to kill the devices plugged into them;
• Multi-path interference (ghosting) -
• fixed: although DTT is relatively immune to this, it can be susceptible. Remedy: better antenna and/or very careful alignment (sometimes a reflected signal is better than the 'direct' one);
• time-varying, from traffic or trees in the wind: as above. It seems to be something of an Achilles' heel even with the COFDM system used in the UK.


Tricks such as siting the antenna so that the roof screens it from nearby traffic can often solve problems. Sometimes corresponding disturbances to the analogue reception, e.g. ghosting or interference, might give clues. If problems tend to occur at a certain time of day one might suspect, say, a neighbour's appliance (try listening to the background with a medium-wave AM radio at problem times). It is possible that signal propagation varies with the time of day and that at certain times the top or bottom of the correct range of signal strength is reached - in that case the weather would also influence it noticeably.

Wideband or Grouped?

"You'll need a wideband, Squire"
 
Unless there is sufficient signal strength, this popular cure-all suggestion can be a fallacy. Many DTT transmitters' multiplexes fall within a recognised channel group. The gain of a 'grouped' version of a given antenna design will be higher than that of an equivalent wideband version - this is particularly marked with Yagis at the lower end of the UHF band. If analogue reception in a different group is required, it might be preferable to use separate antennae. The downside of using a grouped antenna for DTT is, of course, the possibility that the channels might be reassigned or that additional out-of-group multiplexes might be added at a later date.

Jon Godwin
skull@wolfbane.com

Links

• [Link to analogue transmitter information here]
• [Link to article about analogue TV here]
• [Link to article about digital terrestrial TV here]
• [Link to digital transmitter information here]
• [Link to DTT channel list here]
• [Link to some suppliers' contacts here]