|
|
|
|
|
|
by adrian_b
29 days ago
|
|
|
That was definitely a defect of the monitor, not a serious mistake of the programmer. Any damaged monitor should have been paid by whoever had approved such a monitor design, even in the case of a monitor integrated in an all-in-one computer, which was not supposed to be disconnected from the computer. I have also seen some monitors designed so badly, like the IBM MDA monitor mentioned at the link provided by userbinator, but all the TV sets that I have ever seen could only be synchronized to the desired scanning frequencies by the "sync" pulses, but they would oscillate freely in the absence of sync pulses. The cost savings by not having relaxation oscillators or PLLs that would drive the raster scan in the absence of sync pulses, and which would limit the range of acceptable scan frequencies to safe values, have always been negligible, even by the time of the vacuum tube TV sets, and much more so by the time of computer monitors. The elimination of physical safety devices is not an acceptable way to reduce costs, especially when the savings are so small as in this case. |
|
|
The vertical deflection drive appears to have some feedback, so maybe that could constitute an oscillator, but I can't see any feedback at all in the horizontal section, it looks entirely feed-forward. I get the impression they were just integrating the square wave into a ramp. The diagram shows 15us low out of only 50us total line time, which seems like quite unlike the typical 15.75 kHz sync of the time. I recall getting hold of the 6845 datasheet and trying in earnest to understand how to program that chip, and was baffled that the reference set of register values wouldn't produce a display. The fact I was missing at the time was that one had to start from 20kHz for the horizontal refresh.