标签: failure

I saw something I never expected to see on the way home a few evenings ago -- a set of traffic lights that had failed with all of them showing green.   By the time I arrived at the junction in question, there were police cars with their lights flashing all over the place, policemen standing in all four branches of the intersection, and a number of cars (or pieces thereof) piled up at the sides of the road.   That must be a horrible experience -- entering a junction brimming with confidence that you have a green light, and meeting cars coming from your right and left whose drivers also think they have right of way.   I remember them waffling on about this at university -- how to design systems so that they always fail into a known state. As I recall, they used traffic lights as an example, with the design being such that they would always fail to all red; never to all green.   But that was 35+ years ago now and -- like so many other things -- it's faded from my mind. Let's assume the simplest case of two roads crossing at 90 degrees (North-South and East-West).     Do you remember how to create this sort of design such that it will only fail with both directions red (or off)? And can you conceive how the lights I saw yesterday evening had managed to fail with both directions green?   Max Maxfield

In business school, they use case studies to see what clicks and what doesn't. In law school, they study old court arguments. But it's not often that an engineering school uses failure as a teaching tool.   I had this point brought home to me when speaking with Dave Nadler, a consulting engineer whose presentation, How NOT To Do Embedded Development! Practical Lessons From Real Projects That Almost Went Off A Cliff is scheduled for May 7 at ESC Boston. "In any of a number of professions," Dave told me, "you study failures to learn how to recognize impending disaster and avoid it. But in engineering we don't. That's a peculiar thing to our field." Yet in his experience most engineering projects that are trying to do something original don't end up well, for very unoriginal reasons.   Reflecting back on my own engineering schooling I can see his point. I have had dozens of courses on engineering theory, a few involving hands-on design, and only one that even touched on the topic of learning from past failures. It was the one-credit "Introduction to Engineering" course my freshman year at Virginia Tech, and in one of the lectures the instructor showed us a video of " Galloping Gertie ," the suspension bridge over the Tacoma Narrows in Washington that failed months after its completion. That was followed by some discussion of resonance, and then we went on to other things. There was never any kind of "post mortem" discussion of design projects (successful or failed) to learn from by example.   Dave's presentation seeks to correct this educational deficiency through a group exploration of various design efforts, from automatic toll collectors to aircraft anti-collision systems, where he was called in to help rescue a failing project. But Dave's not a wisdom-from-Olympus kind of guy. He doesn't lecture, but instead follows a Socratic discussion approach. He gives the symptoms, and then invites the attendees to respond with their thoughts and ideas.   "How NOT to do..." is a serious effort at learning from past mistakes, but not a grim one. "There are some comical interludes," he promises, "as well as a way-too exciting video of what happens without collision avoidance." Attendees can expect to walk away with some valuable insights, and perhaps a smile.   Rich Quinnell EE Times

In business school, they use case studies to see what works and what doesn't. In law school, they study old court arguments. But it's not often that an engineering school uses failure as a teaching tool.   I had this point brought home to me when speaking with Dave Nadler, a consulting engineer whose presentation, How NOT To Do Embedded Development! Practical Lessons From Real Projects That Almost Went Off A Cliff is scheduled for May 7 at ESC Boston. "In any of a number of professions," Dave told me, "you study failures to learn how to recognize impending disaster and avoid it. But in engineering we don't. That's a peculiar thing to our field." Yet in his experience most engineering projects that are trying to do something original don't end up well, for very unoriginal reasons.   Reflecting back on my own engineering schooling I can see his point. I have had dozens of courses on engineering theory, a few involving hands-on design, and only one that even touched on the topic of learning from past failures. It was the one-credit "Introduction to Engineering" course my freshman year at Virginia Tech, and in one of the lectures the instructor showed us a video of " Galloping Gertie ," the suspension bridge over the Tacoma Narrows in Washington that failed months after its completion. That was followed by some discussion of resonance, and then we went on to other things. There was never any kind of "post mortem" discussion of design projects (successful or failed) to learn from by example.   Dave's presentation seeks to correct this educational deficiency through a group exploration of various design efforts, from automatic toll collectors to aircraft anti-collision systems, where he was called in to help rescue a failing project. But Dave's not a wisdom-from-Olympus kind of guy. He doesn't lecture, but instead follows a Socratic discussion approach. He gives the symptoms, and then invites the attendees to respond with their thoughts and ideas.   "How NOT to do..." is a serious effort at learning from past mistakes, but not a grim one. "There are some comical interludes," he promises, "as well as a way-too exciting video of what happens without collision avoidance." Attendees can expect to walk away with some valuable insights, and perhaps a smile.   Rich Quinnell EE Times

The environment during the assembly process is quite sensitive. Heat, humidity, and cleanliness should be strictly controlled. Depending on the product, a controlled clean room must be used to ensure cleanliness. For standard fine-pitch assembly and for consumer or non-critical business use, more standard conditions can be used. However, the cleanliness should exclude large contamination. Some dust might not matter, but large pieces of material can bridge conductors and cause problems. In this article I am going to spotlight two cleanliness issues, the first example shows an assembly area that does not have the expected cleanliness. The second illustrates an environmental issue that caused the product to fail. Figures 1 and 2 show the inside of a failing product. The product has a sealed enclosure, butÿthere are ants beneath the polyimide tape that was used as a protective layer as well as scattered on the PCB assembly (PCBA). You can only imagine the number of ants that must have been crawling around the assembly area to have two crawling on the PCBA surface so that they could be trapped with the small piece of tape. This manufacturing plant certainly needed an exterminator! Figure 1: Ant on the PCBA surface Figure 2: Ants beneath the polyimide tape. The second example comes from a product that failed in the field after two years of operation. This product was used in a janitorial closet (wardrobe?) in a school building. The outside environment was hot and humid, and it was located near a seacoast. Figures 3, 4 and 5 show some of the dust and debris found within the cabinet. Look at the number of mosquitoes (Figure 5) that had collected in the chassis over the two years! Figure 3: Dust and debris on memory ICs Figure 4: Dust and debris on headers. Header pins were also corroded Figure 5: Mosquitoes found in the chassis   In the second example, the product was thoroughly cleaned and tested. It was then fully operational, showing that it was the buildup of dust and debris that induced the failure. The dust and debris had absorbed enough humidity to become conductive. This conduction was enough to disrupt the signals to and from the memory devices. In this case, the corrective action was to install additional filtration to filter out some of the dust and to suggest that a janitorial closet was not the best place to house the equipment. What kind of gross contamination have you found in your products returned from the field? If you have some good PCB contamination photos, share them in the comments section with a brief explanation/caption. Michelle Woolley

Heat, humidity, and cleanliness should be strictly controlled during the assembly process. Depending on the product, a controlled clean room must be used to ensure cleanliness. For standard fine-pitch assembly and for consumer or non-critical business use, more standard conditions can be used. However, the cleanliness should exclude large contamination. Some dust might not matter, but large pieces of material can bridge conductors and cause problems. In this article I am going to spotlight two cleanliness issues, the first example shows an assembly area that does not have the expected cleanliness. The second illustrates an environmental issue that caused the product to fail. Figures 1 and 2 show the inside of a failing product. The product has a sealed enclosure, butÿthere are ants beneath the polyimide tape that was used as a protective layer as well as scattered on the PCB assembly (PCBA). You can only imagine the number of ants that must have been crawling around the assembly area to have two crawling on the PCBA surface so that they could be trapped with the small piece of tape. This manufacturing plant certainly needed an exterminator! Figure 1: Ant on the PCBA surface Figure 2: Ants beneath the polyimide tape. The second example comes from a product that failed in the field after two years of operation. This product was used in a janitorial closet (wardrobe?) in a school building. The outside environment was hot and humid, and it was located near a seacoast. Figures 3, 4 and 5 show some of the dust and debris found within the cabinet. Look at the number of mosquitoes (Figure 5) that had collected in the chassis over the two years! Figure 3: Dust and debris on memory ICs Figure 4: Dust and debris on headers. Header pins were also corroded Figure 5: Mosquitoes found in the chassis   In the second example, the product was thoroughly cleaned and tested. It was then fully operational, showing that it was the buildup of dust and debris that induced the failure. The dust and debris had absorbed enough humidity to become conductive. This conduction was enough to disrupt the signals to and from the memory devices. In this case, the corrective action was to install additional filtration to filter out some of the dust and to suggest that a janitorial closet was not the best place to house the equipment. What kind of gross contamination have you found in your products returned from the field? If you have some good PCB contamination photos, share them in the comments section with a brief explanation/caption. Michelle Woolley