A very large power generation plant purchased an engineered fire protection system for the additional coal/gas burning turbines that had recently been installed.
The project got a green light and was then passed on to the company's diverse, or should I say motley, gang of crack engineers. We've got all kinds: mechanical, fire, an acoustical engineer, an architect, and me, the only electrical guy.
Then came the usual rounds of coffee and meetings, revisions, another revision, revising the last revision, hurling blame back and forth, and too much coffee, until project completion.
Parts and plans are shipped/emailed to the installation contractor after the customer approved some ridiculously high rev lettered design. A few months pass and I get a message from the boss: "The contractor is having trouble with the fire control system network, call them."
I did call them, and their story that went something like this:
The contracting crew had pulled approximately 4,000m of twisted pair to link all the gas detectors, flame detectors, heat detectors, cut-outs, shut-offs, low level switches, high level switches, solenoid valves, bells horns, and flashing lights to the FACP (fire alarm control panel). The twisted pair starts and ends at the FACP. After all this effort, the network would not work. They thought, "Hey, this is an electrically noisy environment, let's use signal extenders," and indeed, they did. They even went so far as to use more than the maximum number of network extenders allowed by the manufacturer (eight, if I remember correctly). Even all this signal repeating horsepower couldn't take care of the problem.
I did not have any explanation. Nothing about the design was pushing any sort of limit. When I designed the network, I didn't say, "I don't know about this one, guys." Tensions were getting high in the office, and the blame for this nonfunctioning network was swinging in my direction. Then I thought, "What kind of wire did they use? Surely, they used what was specified." I felt like I was reaching, but at this point exploring the more unlikely explanations was all I had.
It turns out they did not use the twisted pair specified. The contactor was given a choice between three different makes of Belden wire, all chosen specifically for their compatibility with the network. What they used instead was speaker wire! Yes, they pulled 4km or 2.485 miles, of speaker wire.
In the email hailstorm which ensued, I asked why they would do something so unbelievably stupid. They responded with this: "It looked the same as the other wire." Meaning that because it was twisted, unshielded wire of the proper gauge, they thought it would work the same as what I had specified. I told them to pull the wire out and replace it. They refused, told me I was wrong, and demanded an alternate solution. I told them that there was nothing I or anybody else could do. Silence filled the air for a bit. Then they emailed me a release of some kind that shifted liability for the electrical portion of the install from them, to me.
Of course, I didn't sign that release, and the contractor eventually had to replace all of that wire, and when they did, the network functioned flawlessly.
- Botched Construction Job Leads to Finger Pointing
- CNC Control in Electrically Noisy Environments
- Glue & Sealant Gunk Up the Line
- Disastrous Automotive Soldering Connections
- If I Only Knew Then...
- Circuit Protection Can't Be an Afterthought
- Teamwork is the Key to Innovation, Author tells NIWeek Attendees
- Automated Requirements Traceability & Verification of Software: Who Cares? I Do