Springs are typically delivered in bulk and feeding them into an assembly line is often a considerable challenge for factories engaged in assembly and installation activities. The reasons behind this difficulty are numerous since springs as a shape have some distinct features. For comprehension, some of the main features of the spring have been listed below:
Especially the entanglement issue FN-1 appears to be well-known in the industry and there are some companies dedicated to solving this problem. However, most of their spring-separators rely on a pneumatic system and therefore require a source of air pressure (about 5.5 bar). The concept is based on the so-called Venturi-effect (Mills, David: Pneumatic Conveying Design Guide. Oxford: Butterworth-Heinemann, 2015. p. 70.):
Figure 1: System-Operational-Diagram of the Becmatic 550
Air flowing through the rear chamber feeds a random quantity of springs forward to the spinning chamber where a tumbling action separates them. Air pressure also feeds single springs into the outlet nozzle to be delivered down with a tube. A pulsed airflow clears the nozzle when tangled springs obscure the outlet by being pushed back to the rear chamber. Through this approach, springs can be fed one by one through the outlet in a controlled manner.
However, the usage of such a system has its drawbacks. Besides the costly aspect of these devices, providing a source of air pressure comes usually with additional complexity concerning connections consisting of airtight pipes leading to a substantial increase in maintenance workload. This would additionally restrain the flexibility of repositioning assets and it should also be considered that air pressure usually requires compressors which could emit potentially discomforting noise.
Further research revealed some alternative and less costly spring separating methods that could operate without the need for air pressure. These non-pneumatic spring separators are usually made up of a cylindrical tank (1) which stores the springs in bulk resting on a rotatable disk (2) with two protruding spikes mounted on a motor (3). The lower cylinder (4) houses additional electronics, a plug and control which consists of a single switch that can turn the motor on and off. When springs are needed the motor starts rotating the disk, causing random collisions between the spikes and the surrounding springs resulting in them getting thrown in all directions and occasionally they slip through a gap and leave the separator. This gap can be adjusted and when set up correctly according to the diameter of the springs it can feed the springs in its separated form.
Figure 2: Sectional view of the Spring Separator (Modified)
However, experiments illustrated that this method is incapable of feeding the springs in a controllable manner meaning the velocity, output flow, and orientation of the separated springs are random. The only way the user can control the device is with the gap size, the number of springs inside the container, and the duration of the switch manually turned on. For instance, when the switch was turned on for one second the number of separated springs exciting the device was between 0-5 springs. If the tank had too many springs inside, it choked the motor and if it had too few springs the output was zero until the gap size was increased (Table 1).
Table 2: Results of experiments when the spring separator was turned on for one second
Trials with different cycle durations did also not yield an improvement for achieving a reliable output of single springs. Consequently, significant modifications were needed to grant it feasibility for industrial applications. The modifications in this work aim to resolve the current shortcomings of control regarding the velocity, output flow, and orientation of the leaving springs.
Controlling spring orientation
To keep the springs separated and to avoid a new entanglement FN-1 the feature FN-5 was used and allowed the springs to be stacked and stored vertically on top of each other. For this purpose, a guidance and collecting unit was necessary to maneuver the randomly exiting springs in a vertical direction. For this matter, a funnel was considered and crafted by a sheet of A4-sized paper. After some trials, it turned out that the length and steepness of the cone-shaped funnel must meet some requirements set by the features FN-4 and FN-7. Since it had to handle springs with various velocities and orientations the angle of the funnel had to be within a range that was neither flat nor steep (Figure 3).
Figure 3: Development of the funnel based on the features of springs
The correct angle was determined by several trials and validations. Afterward, the measurements of the time-proven funnel were taken and 3D-printed with PLA. However, the printed part could not offer the same success rate as the paper funnel. A close examination revealed that the rough surface structure of the 3D-printed part was the underlying cause. As a solution, the initial funnel made from the paper was seamlessly inserted into the 3D-printed funnel and fixed the problem.
Controlling spring output flow and velocity
The stacked springs arriving from the funnel had the right orientation, but the output flow and velocity remained random. To address this lack of control a valve-like apparatus was developed that could retain and release the springs one by one on demand. For this matter, two linear electrical actuators and an inductive sensor were used. Once the sensor detected a spring in the chamber, the first linear actuator retained any potential spring above the spring and the second linear actuator opened the exit (Figure 4).
Figure 4: Valve-like apparatus in operation
Source: Own drawing
This valve-like apparatus can handle a variable output flow of up to five springs at a time while maintaining an output of a single spring. It should be noted that trials so far showed that there was no record of more than three springs entering the funnel, thus the current magazine capacity of five should be sufficient.
The control system is made with an ARM-based microcontroller NUCLEO-F446RE. The Controller can output a voltage of 5 and 3.3 VDC, however, the inductive proximity sensor requires a voltage of 12 - 36 VDC to operate. For this reason, an EL817 optocoupler was used to allow the controller to receive the sensing data of the inductive proximity sensor. The 5 VDC output is used to drive the relay for triggering the 230 VAC spring separator and the 3.3 VDC output is used by the two linear servos acting as retain valve and exit valve. The schematic of this can be seen in the following figure.
Figure 5: Schematic of optocoupler and inductive proximity sensor
The code of the spring separating unit is written in C++ which could run inside any IPC of an assembly line. From there it can grant communication via a serial connection which then could trigger the spring separator depending on given circumstances. In the case that the spring separator will run out of springs it will cycle a given number of times and then send out an error message.
In summary, this modification process (Figure 6) allowed the utilization of a low-cost spring separating device for industrial applications. The microcontroller allows not only to interact but also monitor the spring separator. These additional capabilities could be further exploited for applications such as predictive maintenance and be accessed as IoT. However, the utilized microcontroller isn’t designed for industrial standards, thus it should be considered to substitute it with more robust PLC equipment.
Figure 6: Stages of the modification process