The Case of the Slow Cable

This entry was submitted by Jake Brodsky and edited by Rob Spiegel.

Starting as a field technician, Jake Brodsky has been building a career at the Washington Suburban Sanitary Commission for more than 25 years, and he's still building. He is a registered professional engineer of control systems and a member of several standards committees. Besides engineering, Brodsky stays busy with Ham radio, general aviation, shooting sports, and his three children.

It was the late 1980s and I was working on a brand new control system at a sewer treatment plant. We were using an RS-485 link to communicate with a PLC in a building about 2,000 feet away. We used the lowest loss twinaxial cable specified by the manual. It had a durable polyethylene jacket.

The operators complained that every time it rained, the PLC communications would get flaky. I got curious and took my truck loaded with test equipment to the site.

I figured there had to be a splice somewhere that was no good. I set up a function generator, made a crude TDR system on our scope, but I didn't see echoes. So I walked the length of cable. The conduits and trenches where it was laid were clearly under water in quite a few places. I found the foreman of that job, and he confirmed that he had purchased a 2,000ft roll of the cable and that it was a continuous run with no splices.

Not knowing what else to do, I simulated the 56kbit/s data speed with a function generator set to a 56kHz square wave on the cable. I then went to the far end of the cable to see what the signal looked like on my scope. The signal was weak and showed significant evidence of capacitance distortion.

I went back to the function generator at the other end, set it to 10kHz, and then returned to the scope. It looked acceptable. On that basis, I dropped the data rate to 9,600bit/s and behold, the PLC stayed online rock steady.

Later, I called the cable manufacturer to ask them about the cable that had been specified. "Let's see: The jacket is made of Polyethylene" he said, "and you should know that this material can soak up as much as 3 percent of its weight in water. It was never intended for direct burial."

"Well, it's under water" I said, "and the capacitance per foot is out of specification." "That sounds about right" he replied. "We didn't specify these conditions, so I guess you're on your own."

The link continued to work reliably at 9,600bit/s for several more years. During a later phase of plant upgrades, we made a change order to remove that cable and replace it with a direct burial fiber-optic system.

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