To obtain high accuracy readings from a belt scale, 0.5% or better, the scale must be designed to remove the horizontal forces caused by the belt moving across the scale idler. There are different ways to do this; one is to use a pivot system.
The weigh idler is mounted to a frame that is supported on one end by the load cell and the other end is supported on a pivot, horizontal forces pull against the pivot and vertical forces are transferred to the load cell. Although this method is effective, it creates a maintenance point in the pivot of the scale. It also requires a large frame for the pivot system that makes the scale very large and difficult to install.
I remember once I was working on a pivot type belt scale in a mine in the north western United States. I had done several material tests and could not get the scale to perform. At the time I was young engineer and tended to be a little impatient. After enduring the cold for several days working to get the scale operating, out of shear frustration, I leaned back and gave the scale a kick with my steel toed boots. After that, the scale performed beautiful. While I would never recommend kicking a scale for obvious reasons in this case there must have been something in the pivot that was binding the scale causing hysteresis that prevented the scale from reading accurately. In this case, giving the scale a little jolt freed the pivot and allowed the scale to operate properly.
The Siemens MSI belt scale uses parallelogram load cells. These load cells are designed so they can only deflect vertically, when a horizontal force is applied it does not affect the load cell output. This prevents the scale from being affected by horizontal forces caused by the belt moving across the scale idler. These horizontal forces are removed without the need for pivots or check rods reducing the maintenance required for the scale and allowing for easier installation.