My logic as a non-engineer and a novice ATM:
Usually the mirror rests, either on the floating support points or silicone. Its weight keeps it in even contact with the pads / supports during scope movement, holding collimation. If the supports are springs, each of them presses up on the mirror with its own force, and the mirror would bounce around during scope movement -- probably would not hold collimation. The springs might also somehow cause astigmatism -- putting uneven forces on the bottom of the mirror, or pressing the mirror against the retaining clips.
Your original cell looks like a great (tried and true) solution.