I worked for a company that developed an HAZLOC HMI (human-machine interface) - a ruggedized PC. Some were sold to a drilling rig systems integrator outfitting a drilling platform on Korea and when problems with the HMI were uncovered, I was sent there to troubleshoot and fix them.
As usual, the integrator was under extreme pressure: deadlines were approaching, milestones were missed and a lot of eyes were on the crew.
The HMI had a fiber optic serial link, used to communicate with other resources on the drill floor. The link was integral to the HMI's functions and the driller's primary station's link was intermittent. The integrating team had gone thru the usual voodoo: changing cards, reloading and rebooting, repositioning cables, trying to vary speeds and file size - anything to get a handle on the cause. No luck.
I came out to the rig, got the expected anxious speech and stories about the troubleshooting sessions and went off to follow the fiber routing from the HMI thru the system. The fiber was lashed to overhead cable trays as it ran to other computers and servo interfaces. Anyone who has ever been on a large vessel knows about these trays- all electrics get snugged down in them. Since it was a light pipe, I didn't worry about EMI at first. The first link run was about 100 feet thru a few rooms to a server. Just to get a feel for the tracing problem ahead, I grabbed a short step ladder and peeked at the tray bed. The orange cable was on top of the bundle, but the tie-downs were snugged so tight that the fiber sheath was actually creased in places! Way overkill, I thought, unless this platform does aerobatics.
A few clips with the side cutters and a trial run with a software team and all was well. I took the crew boss up to the next worst cable lash and showed him how severe the cable ties were tightened. He cut orders down the line and I caught a flight home.