When I was working as a field engineer for an instrumentation and control contractor, the company was starting up a new municipal surface-water treatment plant. The engineer who designed and oversaw the installation of the I&C system left for another company before the installation was complete. The job of completing the commissioning of the plant fell to me.

One irritating item on the punch list was a pair of radar-level transmitters on the raw water and finished water tanks. They were located a distance from the control room. The transmitters worked just fine for days on end, then suddenly one or both would begin to send skewed, erratic signals to the PLC. Grounding issues had already required some attention before my arrival, so I checked the grounding. All of the grounds seemed OK, so I turned my attention to the devices themselves.

I went through the setup and calibration procedures step-by-step as the manual directed in the installation instructions. Again, the transmitters would work just fine for days, then one or both would act up again. I tried to discern a pattern -- time-of-day, temperature variation, humidity variation, etc. The time-of-day the devices acted up was roughly the same for both transmitters, usually in the last few hours before sunrise. I often received an early-morning call saying "Get to the plant, the tanks have overfilled."

Eventually, I brought in the sales engineer from the device supplier. He and I spent a day onsite working through the problem while on a conference call with the engineers at the factory. Still, the problem remained. As a last ditch effort, the representative brought out a replacement device. Same problem. We were all frustrated and baffled.

Then one morning I received yet another early call to come to the plant. They were experiencing another overfilling episode. I had no idea what else to look at. I was fresh out of ideas. As the shift operator and I stood outside by the tank discussing the situation, I was looking at the ground and swinging my foot through the dew-covered grass. It suddenly struck me what the problem might be.

Here was my theory: The geology of this area was a porous limestone in a very arid region of central Texas. In the morning when dew forms, the otherwise high-resistance soil could lose enough resistance long enough for a ground loop to initiate. That could create a difference in ground potential between the device and the PLC. After a couple of hours in the sun, the ground would dry out, the resistance would rise, and the device would behave properly. It was acting like a giant semiconductor!

I got out the installation and operation manual and read it from front to back. Buried in an appendix at the end of the manual (not in the installation procedures, where it should have been) was a single sentence, indicating that devices with cable that runs more than 150 feet should be grounded at the device, not in the control panel. I checked the grounding, and sure enough, it was wrong. I corrected the grounding configuration, feeling fairly sure that this would solve the problem. It did. We never had another skewed signal. This was proven out by reviewing the trend history of the SCADA system. I can't tell you how relieved we all were that this pesky problem was off of our back.

This entry was submitted by Russ Hancock and edited by Rob Spiegel.

Tell us your experience in solving a knotty engineering problem. Send to Rob Spiegel for Sherlock Ohms.