Today I worked on the image horizontal centering. As you'll see in the earlier pictures, the image is shifted off to the right. That could be due to one of two things: either the video signal is shifted in time relative to the HSYNC pulse, or there is a DC current through the yoke which shifts everything to the right.
The first of those is easier to address. I built a circuit with a 74LS123 to delay the HSYNC pulse by a variable amount, up to maybe 20us. (Incidentally the Classic analog boards have a horizontal position control which works on a similar principle.) I'll make a schematic later, but here's the result with 4.3us of delay on the yoke HSYNC pulse:
Nice and centered! Just for kicks it can be delayed way too much to see what happens:
Here we see the video folded over during the retrace time. Not useful of course, but it also shows another problem. With the scan lines visible it shows that the picture is, in fact, systematically shifted to the right. That is, the scan covers more of the right side of the CRT than the left. That must be from DC current through the yoke in my simple driver circuit.
I may have to make a fake flyback primary (probably just a big inductor) to be able to build a circuit more similar to the original. C15, which I've removed, is really useful because it ensures that there's no DC current through the yoke. Unfortunately it can't be used in my simplified driver or the result is no signal at all. The other alternative is just to leave it as is and use the delay trick. My main concern is the extra heat the DC current may be generating in the yoke.
Finally, a strange effect. When I increase the power supply voltage on the yoke circuit beyond about 17V, all of a sudden my delay circuit goes unstable and produces an odd pattern where each scan line starts in a different place:
I'm pretty sure this is a crosstalk issue, probably noise on the ground line, which will be solvable with more compact wiring.
The first of those is easier to address. I built a circuit with a 74LS123 to delay the HSYNC pulse by a variable amount, up to maybe 20us. (Incidentally the Classic analog boards have a horizontal position control which works on a similar principle.) I'll make a schematic later, but here's the result with 4.3us of delay on the yoke HSYNC pulse:
Nice and centered! Just for kicks it can be delayed way too much to see what happens:
Here we see the video folded over during the retrace time. Not useful of course, but it also shows another problem. With the scan lines visible it shows that the picture is, in fact, systematically shifted to the right. That is, the scan covers more of the right side of the CRT than the left. That must be from DC current through the yoke in my simple driver circuit.
I may have to make a fake flyback primary (probably just a big inductor) to be able to build a circuit more similar to the original. C15, which I've removed, is really useful because it ensures that there's no DC current through the yoke. Unfortunately it can't be used in my simplified driver or the result is no signal at all. The other alternative is just to leave it as is and use the delay trick. My main concern is the extra heat the DC current may be generating in the yoke.
Finally, a strange effect. When I increase the power supply voltage on the yoke circuit beyond about 17V, all of a sudden my delay circuit goes unstable and produces an odd pattern where each scan line starts in a different place:
I'm pretty sure this is a crosstalk issue, probably noise on the ground line, which will be solvable with more compact wiring.