issue: May 2009 APPLIANCE Magazine Manufacturing Technology Expanded Online: A Quick Look	Share  Printable format  Email this Article  Search
Video imaging puts high-speed production line/automation faultfinding into tiny camera heads.

From an engineering standpoint, there is “no worse word associated with faultfinding than ‘intermittent’,” says Andrew Bridges, director of sales and marketing, and Nils Lassiter, application engineer, for Photron Inc. (San Diego, CA, U.S.; www.photron.com). “The Fastcam MC2 has several trigger methods that make it a straight-forward matter to have the camera record when the problem occurs, be it minutes, hours, or even days after the system is armed.”

The cameras can also be used in destructive testing, in which a part is subjected to extreme stresses, and packaging drop tests.

The light-sensitive, CMOS (Bayer system color), 10-µm pixel, single-sensor camera is available in three models. All three feature 512 × 512-pixel resolution; Model 500 features frame rates up to 500 frames per second (fps), Model 2K offers frame rates up to 2000 fps, and Model 10K operates at frame rates from 60 to 2000 fps at full resolution and up to 10,000 fps at reduced resolution.

The appliance manufacturing engineer’s biggest benefit may come in the ability to record synchronized high-speed video, with a small camera head. “When it comes to camera size, the smaller the better,” Bridges and Lassiter say. They note that the tiny camera heads share many components with those engineered for use in vehicle crash work with the company’s Fastcam MH4 camera system. “The small size enables engineers to mount the high-G cameras on top of, under, or even inside of just about any component they wish to study under the worst real-life conditions they will ever encounter.”

The camera can work with two camera heads, connected for opposing views, for true 3-D motion analysis.

“As with any high-speed video camera, the sensor is the critical element,” Bridges and Lassiter tell APPLIANCE magazine. “The Fastcam MC2 uses a proprietary CMOS sensor designed specifically to Photron’s requirements. Once photons have been converted to electrons through the pixel’s light-sensitive area, this voltage is transported from the sensor through multiple channels, or taps, to adjacent buffers. This analog voltage is digitized by on-chip analog-to-digital convertors (ADC) prior to communication from the camera heads for storage in the processor’s digital memory. The sensor can be globally ‘gated’ or shuttered, independent of the frame rate, with exposures ranging from 1/60th of a second to six microseconds (6 µs). This ensures that every pixel is exposed at the same time to provide crisp imagery that is free of motion blur.”

The precision digital camera system has clock accuracies (oscillator frequency stability at ± 50 ppm) in excess of several parts per million. “The CMOS sensor features an array of 512 horizontal pixels by 512 vertical pixels, each of them 10 µm square,” Bridges and Lassiter explain. “This combination of precision timing, and attention to the sensors’ physical build, means we can achieve extreme accuracies both spatially and temporally. Modern motion-analysis software can track a point’s motion to roughly one-tenth of a pixel’s physical size, enabling engineers to measure exactly where a point is at any precise point in time, at what speed it is moving, how fast it is accelerating or decelerating, the frequency it is vibrating, and whether it came into contact with any other parts throughout its motion.”

What does this mean for the end-user? “Accuracy, and the ability to see exactly what happened.”

Appliance Production Applications

“Photron high-speed systems have also seen plenty of use in production faultfinding situations, e.g., as a line that stamps out parts from sheet steel, [in which] there is an unidentified roller, upstream from the stamp that occasionally causes a tear in the edge sheet metal roll,” Bridges and Lassiter tell APPLIANCE. “When this tear reaches the stamping machine it causes a web break and the entire line must be shut down, rethreaded, and restarted. The MC2 event-capture system could be pointed at the suspicious rolls and operated in the endless record mode where it would continue to record high-speed video until it received a stop trigger from a tension sensor, indicating a web break.

The last few seconds of video from the camera system would then be viewed from any computer in the plant and the production engineer could evaluate it frame by frame and identify the offending roller.”

In the same way the video system could also help the production engineer evaluate welding or high-speed pick-and-place operations using the same mode of operation.
The unit is made to be quite easy to add on the manufacturing line. The power supply that comes with the camera requires 110 V ac power, or the camera can be connected directly to an available 18 to 36-V dc power supply. Installation also requires a place to bolt the camera and, if required, an illumination. This easy installation means the camera can be moved from one part of the line to another as needed. The camera can be mounted on any axis as the software allows the user to easily rotate the image.

A key issue for an installation is field of view, which dictates what lens may be required.

A good zoom lens may solve some installation concerns.

Another key factor is lighting, as high-speed cameras require additional lighting due to their short exposure times. Bridges notes that there are many third-party lighting solutions on the market, from simple dc-powered LEDs to complex and more-expensive flicker-free illuminators. Flicker can be an issue, as many domestic lights turn on and off with the frequency of the ac supply. This appears as fluctuating light intensities when recordings are made at speeds not synchronized with the 50-Hz supply.

The camera maker also offers its own optional LED illumination source, specifically made for use with the camera system.

“This illumination can be operated continuously or synchronized with the high-speed video capture,” Bridges and Lassiter explain.

 A Quick Look—read more online: