Re: analog line clipping

At 04:42 PM 8/19/97 -0700, you wrote:
>
>Zonn said:
>
>"You set the clipping levels to clip just outside the screen. As a vector is
>drawn off the edge of the screen it hits the clip edge and moves no farther.
>Moving it to 0,0 at this point would cause a line to be drawn to the 0,0."
>
>My idea was to use the "watchdog" hardware on the input of the DACs, which
>return the DACs to mid-point if there is no vector activity, to drive any
>clipped line to the center of the screen if any of the four boundary
comparitors
>would fire while a line was being drawn. The fact that the line was clipped
>is latched until the "set initial DAC value" signal is asserted again,
signaling
>the start of the next line draw.
>
>The one thing I forgot to say was that the beam is blanked as soon as clipping
>occurs.
>
>The one thing I wonder, though, is if the deflection is fast enough to get from
>the center back to the screen edge when the line comes back into range.

It would certainly try! And that alone is going to stress the deflection
circuits. I think what you had in mind was something like letting the beam
*float* back to 0,0 which is not what would happen. Instead the deflection
board is going to try it's best to get to the center as fast as it can using
all the available current at it's disposale. This will push much more
current through the yokes (and deflection transistors) than a controlled
draw back to the center, where the speed of the draw is slower.

On the Cinematronics each vector is jumped to at the start of the draw, but
since most draws are "connected" to the previous draw the deflections
circuits have a pretty easy time of things.

Driving the beam to the center of the screen for every vector clipped would
be drawing two half screen vectors at the maximum slewrate, one to get to
the edge of the screen, and one to get back, for each vector clipped. Some
of these vectors were probably quite small (could simply be a vectored
font). There is a chance of distortion since when drawing a pattern on the
edge of the screen, the software expects the trace to already be there,

I'm not sure on the Cinematronics platform what the slew rate is for jumping
to the start of a new vector, so I'm not sure what that kind of jumping
around it's cabable of, but always jumping from the edge, to the center, to
the edge is much harder on the deflection circuits than just drawing around
the edge.

-Zonn
Received on Tue Aug 19 17:33:04 1997