quote:

Just one understand this - and you get a cookie ! or something




That will hold true on any set that does not use a commutating power supply. STK392-1XX series ICs allow for (IIRC) commutation of the negative supplies separately. The preamp must be fed with full Vcc all the time of course, but the actual output supplies are separated for that purpose. '

From an engineering standpoint, only one supply has to be commutated to enhance efficiency during or near retrace. So although for example red and green outputs will have opposing spikes in the wavform, it is a simple matter to invert the phase to use the feature. It really is nothing more than switching the C yoke wires and inverting the input, or even just reconfiguring the STK to be inverting or non-inverting.

Some Toshibas do this via transistors. You would think they would be power transistors but they are not needed. They are either on or off, and they are on for a very short time. Even Vcesat isn't a big problem then. Some Hitachis on the other hand simply used a winding off the flyback to provide the boosted Vcc at or slightly before retrace. That's why there was a phase adjustment in the service menu. If it was off too much, the STKs would clip.

NAP does not use commutating power supplie(s), and therefore usually the best setting is zero DC offset. However when the set does use a commutating power supply, with these chips it will usually be on the negative leg. In some cases a certain amount of DC offset in the output is optimum, and it will usually be positive in polarity. Note that it doesn't matter if it moves the bulk of the red or blue raster(s) left or right, the designer can determine the phase of the drive, as well as whether the STK is configured to invert or not. This, with the extremely easy reversal of the actual C yoke feed and return, you always get the phase you want.

To determine what is happening, I look at a couple of different things. First of all, if it is converged after the ICs and whatever resistors are changed, it is usually OK. If the set is five years old, well, do you give a five year warranty ?

Our problem starts when it fails too prematurely. If the thing is seven years old and the STKs have never been changed, going above and beyond the call of duty is a waste of time. But if you weren't there first, give it some thought. When you take the old STKs out you can get an idea of how long they have been there. In some cases it is obvious that the failure was caused by the lack of heat sink compound, or the improper application thereof.

You do know the proper way is NOT to spread it. You apply a bead right at the center of the pressure of the mounting arraingment. Usually this is right between the holes. In other cases it is the contact point of the clip which holds it to the heatsink. You let it spread from there, which almost assures there will not be any voids. When you spread it first and then mount it, there will be voids. Air pockets. Think of hanging wallpaper.

What's more, the specs Stan provided convinced me even more to use the 120s. I can look at the print and see the supply voltages, and anyone knows that if it uses less voltage it's either not in need of as much correction or it pulls more current. One or the other.

Too late to make a long story short, but I say the 120s are the way to go. The lower thermal resistance is important. A couple of volts or mA either way does not mean as much.





Just one person in the whole world. That's all I ask.

T^T