(From: Jeroen Stessen (Jeroen.Stessen@philips.com).)
Check datasheets of TDA4580, TDA4680, TDA4780, TDA8390:
These video control ICs have automatic black level stabilization. At the top of each field it generates 3 measurement lines (R,G,B). Through a BCI beam current info wire it measures the beam current. It adjusts the amplitude of the measurement pulses so that it gets a fixed current level back, usually around 10 uA. Actually it measures a delta-voltage, the current is determined by resistors on the picture tube panel. The result is used as an offset for the video signals so that black voltage is put at zero beam current. (3 capacitors near the IC store 3 voltages, a 4th capacitor is used to store an ultra-black reference to compensate for leakage current.)
The pulses are easily seen on the screen if you turn down the vertical amplitude (height), you'll see 3 dim lines at the top.
When the set first starts up, the measurement loop can not close because cold cathodes can not produce any beam current. That would cause the loop to increase the black level to maximum. To prevent that, first it is measured whether the cathodes are warm yet. Instead of the 3 small measurement pulses a much larger (peak-white level) warm-up pulse is generated. This looks like a big fat white line at the top (above the screen edge).
The video control IC waits for the beam current to increase (to approx. 5 mA for the 3 guns together) before it will release the black level control loop. This is measured over the same BCI line, a diode adds an extra load resistor to this line so that a large beam current is required to cross the detection level. After this level is reached (was it +8 or +9 V ?) the black level stabilization starts, and after almost a second the picture is unblanked. It should be correct at once.
(I have often argued that this long delay is not pleasing to the customer who would rather have a faint picture earlier.)
Reasons why the warm-up trip level may not be reached are:
If there is too little beam current in general then I would start with re-adjusting the VG2 screen voltage. Because of the automatic black level stabilization this will NOT have any obvious effect on the picture. The stabilization just follows the VG2.
If VG2 is too low then the peak drive limiter will kick in too early and you can't get a decent contrast. Also the picture will be less sharp because at a lower cutoff voltage a smaller cathode area is used plus there's your problem with the warm-up detector.
If VG2 is too high then the video signal can not achieve blanking level anymore and you will see retrace lines. Also the cathodes will wear out faster because at a higher cutoff voltage a smaller cathode area is used, you'll have a sharper picture but not for long.
The proper adjustment procedure for VG2 is in the service doc, you need to observe the measurement pulses on an oscilloscope. But a little twist won't hurt too much, especially if you have already written off the CRT.
(I have the advantage of having known the people who invented this system many years ago, they were once my teachers.)