How do you think the new GigE standards will influence the machine vision industry?
Respond or ask your question now!
SHAKE, RATTLE AND ROLL
Although cameras for the industrial environment are typically designed to be shock and vibration tolerant, there is always some level beyond which they will fail. Sometime the failure can be catastrophic and sometimes it can be simple to solve. Perry Cornelius, advanced systems consultant at ABCO Automation Inc. (Brown Summit, N.C.), recalls a customer who called him about a system that was giving errors. On inspection, he discovered that the lens had been vibrated out of focus. “The lens was unscrewing,” he explains. “Eventually it would have just fallen off. We used a very small amount of nonpermanent Loctite on the lens threads to maintain position. It’s not enough that it’s going to damage the camera or lens but if it’s on a machine that’s vibrating, it will keep the lens from moving.”
Another solution is to mechanically isolate the vision system from the source of shock and vibration. Rather than mounting hardware to a vibrating conveyor, for example, mount it to the floor nearby so that the two are not coupled. Vibration-isolation tables seem to offer a more sophisticated solution but must be used with care. If the application involves finding a fiducial for inspection, for example, the system has to maintain a mechanical reference between the camera and the machine. A compliant support like a vibration-isolation table introduces an unknown offset, compromising accuracy. “Some applications will tolerate that and some will not,” says Perry West, President of Automated Vision Systems (San Jose, Calif.). “My preference is try to isolate the camera from the source of vibration and shock by mounting it separately.”
Launch conditions can impose multiple G’s of force in addition to a spectrum of vibration that changes over time and location on the spacecraft. Because these systems can rarely be repaired, reliability is paramount. Designers test camera prototypes and qualification units exhaustively but when it comes to the actual flight hardware, less is more. Modeling launch conditions can be difficult and as a result, the agencies may develop overly stringent numbers that can result in damage during testing. “There’s an enormous amount of attenuation and shaping in a shock spectrum that happens from the spacecraft,” says Michael Ravine, advanced projects manager and head of the hardware engineering department at Malin Space Science Systems (MSSS; San Diego, Calif.). “We managed to get the shock test waived for the camera on the Mars Global Surveyor. Afterward, they made measurements of the actual vibration environment and it turned out to be a factor of 50 lower than the test spec they had given us. That would have meant hitting the instrument 50 times harder [in tests] than it needed to be hit.”
Heat can pose significant problems in imaging, particularly dark current. Dark current consists of thermally generated electrons in the pixels in the absence of photons, creating a noise source. At room temperature, dark current is very small, but it rises with heat in a geometric progression, doubling approximately every 7˚ C. Typically, it leads to image degradation rather than catastrophic failure, but for some applications, that can be enough to render a system unable to perform satisfactorily.
If the ambient temperature is cool enough to draw off heat, the cheapest and easiest solution is conductive cooling through the camera housing. If not, a better approach is to put the camera into a cooled enclosure. Most cameras are rated for operation up to around 45˚ C. If the application takes place at temperatures only a few degrees above that, then simply blowing clean air into the enclosure can make a difference, provided that it has been filtered and dehumidified.