The installation position of the sensor does not really matter.
The force (or torque) is introduced at one end of the sensor, at the other end of the sensor has an abutment, ie, the force or torque is taken there.
For small measurement ranges, it is useful to arrange the cable outlet to the non-moving side of the sensor. Otherwise, the movement of the cable could cause a force or a torque indicator.
For load cells, it may be advisable to mount the side with the cable outlet, for example, at the bottom of a container. Then, although the end is moved with the cable (by a few tenths of a millimeter), but you can make an elastomeric bearing foot on the floor and the load cell is often better protected against water and humidity.
For the correct supply voltage of the amplifier provides. With most measuring amplifiers it is 5V DC.
10V supply voltage can be too much at ultra-miniature sensors with 350 Ohm already, because the loss of heat can not be dissipated.
The errors caused by self-heating is greater than the gain in signal compared to the noise usually when the supply voltage is more than 5V.
Please the sensor supply voltage is not confused with the supply voltage of the measuring amplifier!
If "resolution" is the "smallest readable step", the question can be answered with the noise amplitude of the measuring amplifier: A good measuring amplifier should solve the measuring range from 0 ... + 2 mV / V in at least 10,000 readable steps.
This means that the noise amplitude is less than 2 mV/V / 10000 = 200 nV/V.
In the case of a sensor with a nominal force of 100 N and 2 mV/V output signal, you can read 0.01 N, only the third decimal point will fluctuate.
However, resolution does not mean accuracy: due to temperature-induced drift or by zero-point return error or creep error, the display may differ by more than 0.01 N after a few seconds or after a load cycle.
The resolution of the measuring amplifier is a function of the data frequency, and thus depends on the filtering of the measured data. The noise amplitude increases with a higher data frequency. Approximately 10 times the data frequency requires a 3-times noise amplitude. This is the same effect as in analog low-pass filters.
The noise amplitude is also influenced by the cable routing and the interference of signals in cables and leads.