Colour Faults: Visual Guide

Some of the faults that are peculiar to colour sets. The following examples were taken from a standard British PAL receiver fitted with a shadow mask picture tube.


The neck of a shadow-mask tube contains three inde­pendent electron guns whose beams must each pass through the shadow-mask to reach only the red, blue or green phosphor dots respectively, so that each gun scans a raster of one primary colour with no impurity or tainting with another colour. If a purity error is present, objects will change colour as they move across the screen and a monochrome picture will contain tinted areas. Purity errors are most visible on the red raster and there­fore purity, is checked by switching off the green and blue guns so that only a red raster is seen.

Most sets are equipped with gun cutoff switches for this purpose; on a few (e.g. Decca CTV25) the guns can only be cut off by backing off the A1 (background colour) presets. Photo 1 shows a red raster with typical purity errors showing.



Loosen the wingnuts on the scan coil assembly and slide the coils towards the tube flare. Adjust the purity rings (these are two ring magnets on the tube neck, rearward of the convergence assembly. They can be adjusted to aid or oppose each other and produced magnetic field across the tube neck at any angle) for pure red at the centre of the screen. A magnifying glass can be used to check that no green or blue phosphor dots are illuminated here. Now bring the scan coils rearward until the screen is uniformly red as in Photo 2.

If necessary retrim the purity rings to remove any remaining purity errors in the corners, Now switch on the blue and green guns singly to check that there are no purity errors in these colours either. If pure rasters cannot be obtained the shadow-mask has probably become magnetised and requires degaussing; check whether the set’s automatic degaussing circuit has failed. With all three guns switched on, a colour set should provide a, good quality picture on a monochrome trans­mission. The grey tones in the picture should have neutral colour. If there is any overall colour tint the grey scale adjustment procedure recommended by the manu­facturer should be carried out.



There should be virtually no coloured fringes visible on the edges of objects. Presence of these means the three primary-colour rasters are not exactly converged (regist­ered) together. There are two sets of convergence controls,” tor static and dynamic convergence. Coloured fringes over the entire picture as shown in Photo 3 are cured by adjusting the static convergence controls on the tube neck. These are three circular magnets which provide radial shift of the red, blue and green rasters respectively, plus a fourth blue.lateral control for sideways movement of the blue raster. Set these static controls for correct registration at the centre of the screen. The cross at the centre of Test Card F is a useful aid for this. Recheck purity after static convergence adjustments.


Even with good static convergence, fringes may be visible in parts of the picture away from the centre as in Photo 4. This is cured by careful setting up of the dynamic convergence controls. A crosshatch pattern generator should be used for these critical adjustments. Failing this it is just possible to use a test card but it is humanly impossible to set up dynamic convergence on an ordinary picture. The skill of achieving a good result is only learnt by practice, a useful hint is to turn each control slightly less than appears to be called for at any stage of the process.


Complete absence of one of the three rasters leaves a picture of the complementary colour, e.g. no red raster leaves a cyan picture. Photo 5, no green raster leaves a magenta picture. Photo 6, and no blue raster leaves a yellow picture (not shown). The cause of the fault might be:-

  • A beam cutoff switch at ‘off’, or
  • Extreme mis-setting of the grey scale controls, or
  • Incorrect A1 voltage           }
  • Incorrect cathode voltage  } — at the faulty gun
  • Incorrect grid voltage         }
  • Internal failure of the tube.

On a receiver with primary-colour drive to the tube cathodes the fault is very likely to lie in the relevant colour output stage.

col5 col6


A block diagram of the decoder used in PAL receivers is shown. A crystal-controlled reference oscillator (4-43 MHz approx.) feeds two synchronous demodulators which detect the two colour-difference signals R-Y and B-Y present in the composite colour signal. Subsequently G-Y is obtained by electrical summing of these two.


The reference oscillator must run in phase with the modulating oscillator at the colour transmitter. To ensure this a ‘burst’ of about 10 cycles of 4-43 MHz is included for reference immediately after the line sync pulse on colour trans­missions. However, the special feature of the bursts rn the PAL system is that on successive lines they are alternately 45° phase-advanced and 45° phase-retarded. Since the oscillator control voltage is smoothed the phase-lock loop does not attempt to follow these alternations and»the oscillator runs at the average phase of the bursts. The burst phase alternations do cause the phase discriminator output to contain an a.c. component of half line fre­quency i.e. 7-8 kHz. This is amplified by a 7-8 kHz tuned amplifier and is a vital signal known as the ‘ident’; it serves two purposes:-

(a). It establishes the correct operating phase for the PAL bistable. This is a multivibrator which is clocked (reversed) by line pulses to run at half line frequency and operates a phase reversing switch in the reference oscillat­ion supply to the R-Y demodulator. This arrangement is necessary because the R-Y phase is reversed at the transmitter on successive lines; without ‘ident’ the PAL bistable cannot ‘know’ in which of its two phases to start.

(b). It disables the ‘colour killer’. Since ident is derived from burst, absence of ident normally means the transmission is monochrome whereupon the colour killer removes forward bias from a transistor in the chroma amplifier so that no unpleasant coloured ‘noise’ reaches the tube.


Many decoder faults upset the ident so that the colour killer operates. Therefore if a set shows a monochrome picture when the programme is known to be in colour (and the set is correctly tuned) the first step should be to artificially disable the colour killer so that the state of the decoder can be diagnosed from the screen. Ways of disabling the colour killer vary; usually a 10 kfi resistor fitted with crocodile clips can be connected from the transistor supply rail to the base of the killer-controlled transistor to provide operating bias.

Failure of the reference oscillator to lock to the burst frequency appears as horizontal bands of colour across the screen, Photo 7. The more bands there are, the further off frequency is the oscillator.


Carefully adjust the oscillator frequency to see if it can be brought into lock to give correct colours. If the oscillator can be brought close to the correct frequency but does not lock, suspect”a fault in the burst amplifier or phase discriminator. If the burst is completely lost the colour is likely to ‘run through’ the picture so fast it can hardly be seen. An oscilloscope is needed to check such points as the burst gate pulse timing and the burst amplifier tuning.


If the PAL bistable runs in the wrong phase, R-Y is incorrectly demodulated and the colours are wrong—in particular, faces are bright green I See Photo 8.


Interrupt the signal several times to see if the bistable phasing is sometimes correct or permanently wrong. If sometimes correct, the bistable is not receiving ident phasing and therefore has a 50/50 chance of starting correctly. Check the discriminator balance, ident amplifier and bistable phasing diode. If the bistable phase is permanently incorrect the ident amplifier is likely to be off tune.



This is a much briefer (some 0’6 microseconds) delay which is fitted in the luminance channel to compensate for, the time taken by chroma to pass through the narrow-band­width chroma circuits and thereby ensure exact hori­zontal registration of colour and luminance. There is no adjustment on the luminance delay and any chroma/ luminance registration error is most likely to be caused by mistuning of the chroma amplifier. Luminance delay lines often, suffer from dry joints causing a characteristic double-edging effect on the luminance signal (mono­chrome picture), the colour being unaffected, Photo 9.



The purpose of alternating the R-Y signal phase on alternate lines is to cancel out phase errors in the signal path. In all British-made PAL sets a one-line (64 micro­second) delay is used to separate the R-Y and B-Y (not alternated) signals before demodulation. The delayed and undelayed chroma signals are summed to obtain B-Y and differenced to obtain R-Y. Accurate separation is only possible if both the amplitude-balance and the phase trim presets associated with the delay are set correctly; if there is any error there will be a difference in colour between adjacent lines—the so-called ‘Hanover blind’ effect, see Photo 10. In coloured areas of the picture, pairs of lines appear to crawl upward.



is to trim the chroma delay presets for minimum blinds, preferably viewing a colour bar pattern, to obtain the satisfactory result in Photo 11. A fault which can be mistaken for very severe Hanover blinds occurs if the PAL bistable sticks in one state; then there is virtually no colour on alternate lines. Often the trouble lies in the line pulse feed to the bistable.



The sound carrier is spaced 6 MHz from the vision carrier while the chroma subcarrier is spaced 4-43 MHz from it. Therefore there is a possible beat frequency inside the video range (1-5 MHz approx.) which could be generated by interference between sound and chroma. The i.f. strip response is carefully shaped to minimise this possibility. However severe misalignment or simple mistuning of the receiver can cause 1 -5 MHz patterning on the coloured parts of the picture—see Photo 12.





Leave a Reply

Important Notice: To leave a comment use the box below. You must tick "I'm not a robot" and then complete the puzzle, otherwise the automated comment system will reject your post and treat it as spam.

Your email address will not be published. Required fields are marked *