What classification system do you have? Cyclones (vortex/apex dimensions)? Screw/spiral/rake classifiers? Type of mill ( dimensions?) Grinding media size? What is the current plant tonnage? Current %solids/density coming out of the grinding circuit?

If I understand it, you want to increase the cut size in (I'm assuming) the cyclones.  Assuming this is a conventional ball mill in closed circuit with cyclones, then in general, this can be achieved by installing a smaller apex, a larger vortex finder, decreasing the pressure in the cyclone header, and/or reducing cyclone feed density (i.e. less water to the cyclone feed pumpbox) - or some combination of all of the above.  However, if this is a longer term target, then you should look at decreasing the power input to the grinding circuit.  This can be achieved by dropping the ball charge in the ball mill, i.e. simply charge less balls per day (but don't stop charging balls), to drop the power draw.  Other options involve slowing the mill down, either by installing a VFD (somewhat expensive depending on the size of the mill motor) or changing the pinion (more expensive and significant down time).  As always, make one change at a time, and quantify the before and after to make sure you are heading in the right direction.  I hope this helps - Andrew

Do you want to increase the D50 while maintaining the D80 relatively constant?

Shahid Rind answer the questions we need to help you and try increasing your classifier density.

Read https://www.911metallurgist.com/blog/spiral-classifier-for-mineral-processing and https://www.911metallurgist.com/blog/screw-classifiers and https://www.911metallurgist.com/blog/mechanical-classifiers