Yeah. agreed, the severe ringing in the video is the source of the problem, until you can nail down the source and fix it, you will never have a clean lock, like EM suggested, if you could somehow feed in a cap coupled composite video signal at point B, right after the video det, it would rule out the IF section as a source of the problem, but it looks like it needs to be about 6v P-P (ref W1) to do any good.
