Who knew that terra firma would have been the culprit behind one client's PC connectivity problems? In the late 80s I had a customer with various PCs connected to an IBM mainframe controller via coax. We had a maintenance contract on all of the PC hardware that covered every component, and all but one of the PCs worked just fine. The 3270 emulation screen would just lock up and stop communicating very intermitte Who knew that terra firma would have been the culprit behind one client’s PC connectivity problems? In the late 80s I had a customer with various PCs connected to an IBM mainframe controller via coax. We had a maintenance contract on all of the PC hardware that covered every component, and all but one of the PCs worked just fine. The 3270 emulation screen would just lock up and stop communicating very intermittently to add a layer of complexity. Because this customer was a long drive away from our office, the contract dictated next day service, and so I had asked the dispatchers to let me know when it failed again so I could drop anything I was working on and go there immediately. The problem perplexed me but it had also become one that I was determined to figure out. The final time I was called out for this problem, the customer was asked to leave the PC alone in the failure state. When I arrived, sure enough, the emulation screen was locked up. I tried swapping the cable with another port on the controller – it still failed. Over time, I had swapped the 3270 coax card with just about every other brand on the market, still without success. New motherboard: failed. New power supply: failed. Every other removable PC part: failed. Static mats on the floor, under the PC and keyboard: no change. I was going to swap the whole PC with another brand of a newer model from our stock when I reached behind to start disconnecting cables. Zap! I got a strong shock from the BNC connector on the 3270 coax cable. A bit of research showed no difference from the ground in the outlet where the PC was connected to the BNC connector. I brought in an extension cord from the outlet where the 3174 controller was plugged in and measured again. 92 volts. Hmmm — a floating ground. As it turns out, the PC having the problem was in the original, old portion of the building. The newer part was built and the exteriors were all remodeled so that it looked like one new building. When I traced the outlet to the electrical panel and then from the panel to the exterior ground I could see the (formerly) long ground spike. I kicked it to unearth it. It was just about 6 inches off the original ground just resting vertically on the surface. When the sun would be out for a week or so, the ground would dry out enough that the old section of the building no longer had an earth ground. The PC would lock up and we would come out the next day. Of course, the night the call was placed, it rained. The ground was working enough so that it was not a problem when we would be on-site. We never saw the issue — just swapped hardware. An electrician was called, the ground spike replaced, and we never heard from them again. At least, not with that particular issue. Data Management