Page 44 - Handbook of Biomechatronics

38 Reva E. Johnson and Jonathon W. Sensinger

no-load speed. Two common normalizations are often used. The first nor-
malization is with respect to mass (e.g., stall torque per kg of actuator, or
no-load speed per kg of actuator), as larger, heavier actuators can typically
produce better performance at the expense of larger, heavier designs. It is
important when reviewing these specifications to clarify whether the mass
includes all of the components (e.g., compressors for hydraulic actuators,
power sources, etc.) to ensure a fair comparison.

2.2.2 Torque and Speed Constant
The second normalization is with respect to electrical input. Electrical
motors produce more torque and spin faster if they are provided with more
voltage (which, for a given electrical resistance, in turn permits more cur-
rent). The torque constant (K t ) is a measure of how much torque per
amp a motor can produce. If SI units are used, it is nearly equivalent to
the reciprocal of the speed constant K v , which is a measure of how much
speed per volt the motor can produce.

2.2.3 Mechanical Power
These metrics of maximum speed and torque convey useful information
regarding the performance of an actuator. However, most actuators must
produce a range of torques across a range of speeds. Many designers accord-
ingly turn to the maximum mechanical power the motor can produce (typ-
ically in Watts). Maximum power for electric motors occurs at half of the
no-load speed and half of the no-load torque (Alciatore and Histand,
2003). Mechanical power can also be normalized by mass (e.g., W/kg). If
applications are in this vicinity of torque and speed it is a useful metric,
but if actuators operate away from that region, it can misrepresent the
comparative performance of different actuators. For example, many bio-
mechatronic actuators must either produce high torque (e.g., sit to stand),
or high speed (e.g., walking), but rarely both at the same time. For these
activities, the metric of maximum mechanical power is a poor proxy for task
performance (Sensinger, 2010a).

2.2.4 Envelope Visualizations
Envelope visualizations capture all four relevant parameters (torque, posi-
tion, velocity, and acceleration), but they do not provide a compact number
and cannot be easily normalized. The other metrics we have discussed (e.g.,
stall torque and mechanical power) provide compact, normalizable num-
bers, but they are only accurate for specific portions of the applicable space