FERRANTI 20T6

This set is the last in the line of developments that started with theband-I-only 20T4 in march 1954. The 20T4 later became available with a factory-fitted add-on-style band I/III tuner. This was soon replaced by the creatively titled 20T5 with a pukka 2-band tuner and further circuit refinements. The 20T6 contains further, if somewhat minor, circuit refinements and, with larger C.R.T. sizes now becoming available, one of the last of the projection sets.

An unusual design feature of this set is a semi-push-pull video amplifier. Valve Va forms a fairly standard video amplifier driving the CRT's cathode. It's screen current varies in inverse to the anode current, albeit at a lower magnitude. Thus a smaller inverted video signal appears across the screen resistor Ra, which is amplified by valve Vb so that the inverted signal is a similar magnitude to Va's anode. This inverted signal then drives the C.R.T.'s grid, giving double the effective C.R.T. video drive.

Oddly though the drive signals are A.C. coupled to the C.R.T. so I can't see how a correct black level was produced ; in the earlier 20T5 and 20T4 Va was DC coupled. I wonder why the change was made as it seems a backwards step to me.

As with all projection sets, the ith 25KV E.H.T. is generated by a seperate oscillator. With this level of voltage on a 2½-inch C.R.T., failure of either timbase would risk burning a line on the phosphor and hence all but the earliest projection sets were fitted with protection circuits that in effect set the brightness to below minimum during failure. The interesting feature added to this set is a a double diode valve (EB91) configured as a relay (!). The heater power for EB91 is generated from the line output transformer (much the same way as a conventional T.V. powered it's E.H.T. rectifier) and the EB91's diodes were used in bias circuits - the E.H.T. generator for example - where they'd normally bahave as a low resistance. Failure of the line timebase would therefor extinguish the EB91 removing, the bias voltages and inhibiting the EHT generator. Neat ! and an idea I've not seen utilised anywhere else. In addition this provided a delay at switch on allowing the timebases to have stabilised before enabling the EHT generator.

This is the last model introduced by Ferranti before they became re-badged Ekco models.