Steve from JH said:
The picture below comes from my playing around with a Stephenson 6-bar linkage program. The pivot positions are not perfect but pretty close. The red line between the left hand red dot and the point of the black angle above it would be the Felt axle path approximately.
In real life the path should be slightly more rearward. But this diagram shows how straight it is--straighter than any other design I think.
The thing that sets the Equilink apart from other linkages is that you can't calculate the anti-squat from simply knowing the axle path. The red link acts to short circuit the transmission of forces from axle to frame so that the bike has much more accelerational anti-squat than the axle path angle would indicate.
The design will pretty much eliminate squat from acceleration but does very little to reduce squat from rider bounce. Some riders will like this; some will not.
I agree with your assessment. There is a mechanical bias between low and high leverage torque rates when combined in the same mechanism. A higher torque leverage will override the influence of low leverage rate. The Equilink design's cross-link has higher leverage torque during acceleration tension but a much lower leverage torque relation during bump activation.
During pedal and acceleration there is more horizontal tension directed across the straighter aligned lower stay and lower frame link and so the cross-link would be a higher leverage factor binding the activity of the upper frame link against vertical movement. But vertical input from bumps or rider bob would direct leverage torque along the upper stay and upper frame link, and then the cross-link is in a low torque relation mainly keeping the system from collapsing and not binding the suspension so as when there is horizontal tension.
Looking at Francois's picture at the top with the shock and cross-link disconnect reveals the answer to the question I earlier had about spring rate. I had earlier thought there might be a falling rate spring tension from rather stiff flex stays. (falling rate springs or shock damping produces a pedaling platform effect). But it appears that the rear stays flex fairly easily so there would be very little spring rate influence within the travel range.
So there is something special about the Equilink. I look forward to demo riding one hopefully at Sea Otter this April. I'm curious if low pedal bob comes at a cost of mush deadening pedal performance like platform shocks do for monopivots compared to the lively and snappier performance of a more purely path based high anti-squat suspension such as VPP and the exceptionally well balanced dw-Link. I'm expecting the Equilink pedaling is also very snappy when the shock has minimal platform damping.