I would not call the T4 computer manual nor would I confirm it being more accurate than the radar.
The T4 had a gyro stabilised sighting glass and the computer was fed airspeed and altitude etc. If the computer was U/S the sighting head could be used in manual mode.
In the Varsity, at 4,000 feet and 150 kts 17 feet was possible.
For any accuracy the Vulcan bombsight had to be calibrated. We would go to El Adem range to drop 16 single 100lb practice bombs from around 45k. Forward vision was around 15 miles. With a forward throw of 7 miles the plotter had just one minute to identify the target and align the aircraft. To assist the nav rad, who also controlled the bomb selection, would direct the aircraft toward the target. Once sighted the plotter could steer the aircraft though a rotary knob and the auto pilot. The bank sensitivity could be adjusted.
Apart from the poor visual range the weak point was the physical linkage to the gyro stabilised sighting glass hence the need for calibration. On our first run everything was fine until the plotter started to direct aircraft. I saw the bombing director indicator drift off and aborted the run. I did the same for the next two runs until the skipper said to let the plotter drop one. It landed 1,600 yards off. The next at 1,400. Thereafter we dropped the rest on radar.
As it happened the aircraft had only been calibrated the week before and we should not have been allocated that one. Back at Cottesmore I challenged the IX Sqn plotter. He agreed that the first bomb was way off so he simply aimed off. Prat. That would be no use to the next crew on an operational first run attack.