LabVIEW and IMAQ Control Production of Precision Alloy Microspheres

Author(s):
Robert Hamburger - Bloomy Controls

Industry:
Semiconductor

Products:
Vision, LabVIEW

The Challenge:
Improving speed, efficiency, and control of manufacturing process of metal microspheres measuring 80 to 800 microns in diameter.

The Solution:
Redesigning image acquisition and controls portions of system using National Instruments products.

Abstract
Bloomy Controls, Inc. developed an image acquisition and control system for a manufacturer of precision alloy metal microspheres - measuring 80 to 800 microns in diameter - that are used in hi-tech manufacturing. The manufacturer employs a proprietary process to create the microspheres. As free-falling, liquid metal droplets cool and solidify into spheres, an image acquisition system identifies, counts, and measures the spheres. This data feeds into a PID loop that manipulates the flow of liquid metal thereby controlling the size of metal spheres formed.

Introduction
In this manufacturing process, molten metal is metered through a vibrating precision aperture. The liquid metal droplets free-fall through a pressure chamber as they cool and solidify into spheres. Previously, the manufacturer used a machine-vision system to measure and control the size of these droplets. In this prior system, a strobe backlit the spheres, which were imaged using a camera fitted with a telescopic zoom lens. A frame grabber board acquired and processed the images with customized C code. This code performed pattern recognition to identify and measure the size of the spheres. Simple PID processing compared the time-averaged sphere sizes with a desired setpoint and produced a process correction parameter. A signal generator board generated a signal of the required frequency. This waveform then was amplified and fed to a mechanical transducer driving the vibrating aperture. This achieved measurement and closed-loop control of sphere diameter.

Although the general strategy employed was sound, this system’s particular implementation had the following significant weaknesses that the manufacturer wished to address:

• Slow image acquisition and processing
• Low number of spheres identified and measured
• Little adjustment capability associated with the image acquisition and controls
• Poor closed-loop control of the desired sphere diameter
• Undocumented source code that was difficult to maintain or modify

Bloomy Controls built a solution using National Instruments products to increase the overall speed and effectiveness of the image acquisition and processing and to provide the controls necessary to adjust and optimize the system over a variety of operating conditions.

System Description
The system developed by Bloomy Controls incorporates National Instruments LabVIEW, IMAQ Advanced and PID toolkits, and image acquisition and analog output plug-in boards into a standard Pentium III PC running Windows NT. The PC provides access to all operator controls and displays the measured parameters. The system uses the existing strobe and camera equipment, upgrading only the image acquisition and signal generation hardware.

 A PCI-1408 monochrome acquisition card captures images from a high-resolution industrial black-and-white CCD camera. A PCI-6711 analog output card generates a variable-frequency waveform that, after suitable amplification, drives the mechanical actuator that produces the metal alloy droplets.

Bloomy Controls used several of the basic and advanced IMAQ functions in LabVIEW to perform image capture and manipulation operations. The use of a programmable Region of Interest (ROI) eliminates unnecessary acquisition and processing of image areas that do not contribute relevant data. A Grab Acquisition receives the most recent image in the fastest manner possible. Using IMAQ Threshold.vi, the image then is contrast-enhanced to convert it into a binary format for subsequent handling. The resultant binary image is examined by IMAQ Circles.vi, which identifies, counts, and measures the spheres. These values are returned to the main control program for display, closed-loop control, and saving to a data file.
Once images are acquired and data is produced, the application uses a selectable data rejection scheme to prevent widely out-of-tolerance spheres from disturbing the steady-state process. However, the diameters of all recognized spheres are saved in a tab-delimited file for later analysis. A running average of the acceptable sphere diameters is then created and input to a PID VI, along with a setpoint reflecting the desired target sphere size Both this running average and continuously updated Standard Deviation are displayed, notifying the operator immediately of any problems with the process. The correction signal produced from the PID algorithm is then fed into a Programmable Signal Generator VI, which updates the output drive frequency to the mechanical actuator producing metal spheres. Thus, closed-loop control of the process is obtained.
The application’s Main Screen, shown in Figure 2, displays all critical measured parameters and provides controls for adjustment of the process. Up to three different images may be selectively displayed, allowing the operator to fine-tune the various image-processing settings. Specific control palettes appear as "pop-up" panels through the use of VI Server. Each of these separate control clusters is linked dynamically to the main control loop, which continues to run uninterrupted in the background. In this manner, screen space is used efficiently while allowing complete adjustability of the process without disturbance.

Results
This new machine vision and control system for precision alloy microsphere manufacturing has produced numerous benefits over the old system. Because of the powerful image processing and analysis tools provided by LabVIEW and IMAQ, the system is capable of capturing up to 15X more microspheres in a given production run. This has, as a result, increased the in-tolerance process yield by more than 250 percent. All important measured parameters are automatically logged and stored to a file for future reference. Because statistical results are produced while the process is in operation, the operator may take immediate action to adjust or fine-tune various controlling parameters. Efficiency, accuracy, and repeatability all have been increased as a result of this LabVIEW/Bloomy Controls solution.

Browse All Case Studies »