Hi,

now, I can see no more pumps here, and no air.

To measure the weight being lifted at constant speed by a mechanism, using
the input power to the mechanism, you have to consider the fact that the
efficiency of the mechanism+motor system is not constant as the weight
changes. This is more than the torque, it relates to the particular kind
of mechanism, the way friction appears in bearings, and between gear faces
and more. With a small load the efficiency is likely going to be poor on
account of the motor and friction losses, then it will be good for medium
load, and then it will be poor with high load as elastic deformation
begins to increase friction and bearing rolling losses.

The only way out is to calibrate the whole system using a prototype. This
is best a final product taken off the assembly line because there are SO
many things that can change the curves. From experience I can say that the
effects I have quoted above will affect your theoretical read-out (based
on the motor curves) by between 5 and 30% (for low loads).

My experience is based on overload sensing circuitry in equipment that is
probably similar to what you are doing.

hope this helps,

        Peter