By John Rezabek, Contributing Writer
Until a few years ago, one of the largest refineries on the continent kept its 1980-vintage DCS alive piece-by-piece and board-by-board. You sensed some pride from the group that was keeping the old system functioning reliably—we had sold them all the parts from a similar system we'd replaced decades ago.
I have to think these folks knew the system better than anyone back at the supplier company, which officially ceased to support its obsolete product years ago and had been going through rough financial times.
A few years ago, they began to install a present-generation system in its place, one full of Windows boxes and commercial-off-the-shelf (COTS) network hardware and third-party manufactured controls. Where were their old hands going to apply their highly developed skills?
I chat with our neighbor's analyzer maintenance specialist every now and then on the subject of troubleshooting. Each of us has individuals in our organization who appear to have a knack for troubleshooting, and a number who need more help. I always ask him, "Don't you think troubleshooting can be taught?" to which my neighbor flatly replies, "No." That's a little discouraging if you're trying to develop an organization whose positions haven't been filled on the basis of education, talent or ability. If gifted troubleshooters are born and not made, should maintenance supervisors just plan to be on call?
Strong troubleshooting skills can be application-specific, as well as physical or virtual (or both). I hope everyone has a gifted auto mechanic they trust, but some of those folks cease to be miracle-workers when faced with the heavily microprocessor-based and sensor-dependent machines of the past 20 years. Sometimes it seems easier to smash them into scrap than slog down the hard road of trial-and-error.
A significant portion of our everyday machines have followed the path of the automobile. Do you still have a corner TV repair shop? There's not a lot of troubleshooting when the whole TV consists of one wave-soldered motherboard. Hobbyists can make some guesses like: "These are the cheapest and lousiest capacitors on the board, so I'll start by replacing all of them."
Such practitioners enjoy a measure of success, but in the world of production-plant economies, it's often cheaper to just get a new monitor and relegate the old one to the electronics-disposal bin. Even if we're lucky enough to succeed 50% of the time, the labor costs can exceed the price of a new one, not to mention all the work that isn't getting done while you and your reports are wicking up solder in the shop.
In today's world, we're replacing entire boards or even an entire computer or logic solver, in large part because they've become so modular and inexpensive. But Windows-based boxes have some issues. If you have a Windows PC that's more than a few years old and it has a power supply problem, you could be surprised how hard it is to find one to replace it.
Even if you replace the entire PC, the new box probably has different memory, hard drives and optical drives with an entirely different interface (SATA), and potentially a new operating system. Have you tried to buy a copy of Windows XP lately?
Commercial-off-the-shelf was great for quick, cheap solutions, but some of us failed to foresee how dicey it could be to maintain and service them. Neither repair and refurbish nor rip and replace offer a smooth path back to a functioning system.
Diagnostics. Have you heard enough about diagnostics lately? We're getting a boatload of diagnostics from most of our microprocessor-based devices. How many times have you zipped up your Windows "Event" logs and sent them to your systems provider for analysis? Do you ever look at them?
The message here is that diagnostics are better than nothing, but more often than not they require extensive scrutiny by a thinking-brain human.
Visionaries depict our smart devices writing their own work orders in CMMS or ERP systems. For me, this conjures visions of thousands of inscrutable work orders. Some grand evolution is needed before I can believe we won't need gifted troubleshooters to make decisions about what's really broken.
Although there's a lot of rip and replace in our world today, we really can't afford to send mind-numb robots out to simply swap out failed boxes. Thinking brains, we need you to get on your yoke and resume slogging to a solution.