Glenn, The geometry devils attack again! The bottom line with your setup is that you want an angular change in one arm to result in an equal but opposite angular change in the other arm. Right? The problem is that with a fixed distance between the two connection points, an angular change in the position of one point results in an unequal angular change in the position of the other point. This diagram might help.
You'll see that the distance between two points on the circumference of two equal circles varies as those points travel in opposite rotary directions.
Can't say I have a specific suggestion for you right now, but I cam confident that the final solution will involve some method of anchoring the motion to a known centerline, such as a fixed pivot point or a linear slide.