Using the control algorithm developed, the apparatus can switch between ‘float’ and manual modes. In the ‘float’ mode a hoist will counterbalance the load so that it feels substantially weightless, allowing the operator to apply forces on the load itself to move it in the desired direction. It will also be highly responsive to operator force inputs due to the small dead-band, and so can be moved at high as well as low speeds. In manual mode a hoist will operate traditionally using a pendant control.

The need

The operational characteristics claimed for the apparatus enable applications in hoists and related equipment to be used for precise handling of large loads. In industry, specific applications may include the placing of an engine block on mountings within a vehicle compartment. This application requires both high and low speeds, plus precise multi-direction operation for engine location. The frequent alternative is for workers to resort to manual strength when a hoist with sufficient capacity to lift, as examples, an engine, car seat or battery has frustratingly clumsy operation for assembly duties.

The patent outlines the drawbacks of traditional and other previously developed hoist actuation and control devices. Traditional hoists have high capacity but have slow, non-variable speeds, with remote-control operation that is less than ideal for many manufacturing operations. Various counterbalancing devices have been developed to improve performance for these needs, with performance being assessed on the basis of:

– Being able manually to accommodate varying payloads easily when weight can change between manufacturing operations

– Automatically counterbalancing dynamically varying payloads as weight can change as assembly work proceeds

– Giving a small ‘dead-band’. Large dead-bands are detrimental if the dead-band is significant compared to the payload. They are usually the result of static friction in the counterbalancing device.

– A small physical size, especially to allow the maximum floor clearance of the payload

Spring balancers allow manual adjustment of the spring tension to accommodate payload changes but the possible counterbalancing force changes are only small, and it is difficult to dynamically change the counterbalancing force. Dead-bands are very small. The material properties of spring steel prevent these balancers prevent them from being scaled up for heavier payloads.

Pneumatic balancers have no sensor to measure force, so do not typically accommodate dynamically changing loads. Static friction is high due to airtight seals, giving a large dead-band. These balancers are therefore less than desirable for small loads, and can become cumbersome for large ranges of motions.

With a servomotor-controlled balancer, the spool torque, and hence the counterbalancing force, can be controlled accurately through controlling motor current. With no force sensor, these balancers do not typically accommodate dynamically changing loads. They are typically very inefficient at low speeds and high torques, as in hoists. To compensate, the hoist design can become cumbersome. A smaller motor can be efficient at high speeds and low torque, using reduction gearing to lower speeds. The latter may introduce significant friction – infinite in some cases and giving a very large dead-band.

Load-cell balancers allow hoists to counterbalance dynamically varying payloads. Hoist control is through a feedback controller where the actual load force is measured using a load cell. The motor is then servoed to correct for differences between the desired load and the actual load. Small dead-bands are achievable and designs can be physically compact. Load cell balancers are often suitable for both light and heavy payloads that vary dynamically.

Disadvantages are sensitivity to shock loads and high mechanical stiffness. Controller load gains need to be kept low for feedback control loop stability, resulting in sluggish response times. ‘Chatter’ is common in load cell systems when in contact with stiff environments.

A previous patent (US Pat. No. 5,650,704) associated with the current inventors disclosed ‘Series Elastic Actuation’ with an elastic element intentionally placed in series between a motor and load. This allows the introduction of high control gains, thus achieving low impedance and high force fidelity.

Despite improvements over load cell actuators, the actuator motion is bounded and typically small, as the elastic element has to move with the load. Rotary movement may be limited by twisting of sensor wires. Although this may be acceptable for limited motion applications such as robot arms, it is not for the large motion required in hoists and cranes.

How it works

A hoist used as the basis of the patented apparatus includes an armature sub-assembly with right and left armatures supported on the drive shaft and with a spool shaft connecting the armatures via bearings. In addition to the elements of a traditional hoist, the apparatus has a physically compliant measurement element (for example, compression spring, torsion spring, rubber element), deflecting in relation to the force on the load, constructed and arranged to provide at least partial support between the power transmission element and the base of the hoist, separate from the power source.

The hoist can also be designed with the power source arranged to partially support the load. An elastic element is coupled to a hoist baseplate and supports at least a portion of the power transmission element on the baseplate, without the elastic element exerting a substantial force on the power source.

The patent also covers a method of controlling a power hoist, which comprises measuring the deflection of the elastic element that provides support between the base and a payload without the support passing through the power source. Such a position transducer (potentiometer, strain gauge, optical encoder, etc) provides a signal to a controller both indicating the deflection measurement and giving a signal to actuate the power source in relation to the deflection of the elastic element.

Higher control gains allow friction and motor inertia to be masked substantially, resulting in further reduction of the dead-band. The patented invention offers advantages over conventional counterbalancing devices including:

– Option to manually change the counterbalancing force to accommodate varying payloads

– Automatic changing of the counterbalancing force dynamically to accommodate varying payloads

– A small dead-band with closed-loop feedback resulting from the high control gains provided by an elastic element for force sensing

– Physically compact design

– Inherent shock tolerance

In Fig 1 a pair of compression springs (46a and, not seen, 46b) are the complaint elements. In some designs one spring may be used or the two springs can have different properties. The drive train assembly includes a shaft encoder (20), a brake (22) and a servomotor (24), concentrically aligned and affixed. The motor output is coupled to reduction gearing (26). The drive shaft (32) is supported by a bearing (41). A drive gear (38) is mounted on the drive shaft. All are mounted on a baseplate or other suitable arrangement. A motor amplifier (34) and controller (36) are also mounted here. The armatures left (42a) and right (42b) carry bearings, which are connected by a spool shaft (50).

The compression springs are fixed to these armatures with the free ends on the baseplate. The position transducer (for example, potentiometer, 48) is connected between the left armature and baseplate. The spool assembly includes a spool gear (52) meshing with the drive gear and spinning freely on the spool shaft (50).

In an operational hoist set-up the operator controls the hoist with a pendant that includes push-button switches for selection of float or manual up/down modes. In float mode the payload is actively counterbalanced by the hoist with a closed-loop feedback control algorithm described below.

Thus, the operator can move the load up and down directly by hand. Alternatively, the operator can use manual up/down mode in which the equipment behaves as a traditional hoist. Velocity commands can be issued via the pendant control.

In lifting a load the servomotor powers the reduction gearing and thus rotating the drive shaft and gear. The drive gear, meshing with the spool gear, rotates the spool to wind the payload wire rope up or down depending on motor direction. The brake can be used to lock the servomotor in place, preventing movements of the payload except those small motions allowed by the compression springs. Brake operation is automatic in case of power failure.

Figs 2 and 3 show how the force on the payload can be measured. As the force on the payload rope increases, the springs deflect to counteract a proportion of the force. This is better understood, say the inventors, if the situation with the brake applied is considered. The spool gear is then unable to rotate freely about the spool shaft as the drive gear is held, and so the wire rope cannot unwind. However, the armature and spool sub-assemblies can still rotate with respect to the drive gear. Thus the payload produces a downward force on the armature sub-assembly causing it to rotate counter-clockwise (CCW) around the drive shaft. The springs provide a counterbalancing force to stop the CCW rotation of the armature assembly.


Referring to Fig 4, the counterbalancing force applied by the springs can be computed. It is assumed that the armature and spool subassemblies are in equilibrium so that the force on the spool sum to zero, as do the torques about the armature bearings. Forces from the load, springs, and drive gear are also assumed to be vertical.

The algorithm developed relates the payload (Fload) to the balancing force applied by the spring(s) and the radii of drive gear (Rgear), the load drum (Rcable), and distance between the spring(s) and the centre of the drive gear (Rspring). Thus:

Fload = Rspring – Rgear x Fspring / Rgear + Rcable

Algebraic manipulation has eliminated the element Fdrive gear. The force on the spring (Fspring) can be calculated using Hooke’s Law, knowing the spring constant of the spring.

The patent displays flow charts that detail the controller operation in the available modes.

Load force control

The control can control the force on the load by setting the current to the servomotor. If the actual force is greater than that required the servomotor causes the spool gear to rotate CCW to unwind wire rope and lower the load. This has the effect of dynamically decreasing the actual force exerted on the compression springs. If the actual force is less than desired, the servomotor is controlled to cause the spool gear to rotate clockwise, raising the payload and having the effect of dynamically increasing the actual force exerted on the compression springs. Other controller operation procedures to correct forces on the load are possible.

About the patent

This article is an edited version of US patent 7,090,200. The inventors are all attached to the Massachusetts Institute of Technology and three were founders of Yobotics, Inc.


This article