At 2:17 on a Saturday in late September 2022, my phone pulled me out of a deep sleep. The screen said “Control Room,” which is never a good sign at that hour.
“Boiler 2 tripped,” the shift supervisor said, without any small talk. “High drum level. The feedwater valve stopped responding before the protection did its job.”
I didn’t need to ask which valve. I drove to site in the dark knowing exactly what I would find — and knowing whose signature was on the purchase order that caused it.
That is probably the right place to admit who I am. I’ve looked after flow-control and rotating equipment at a chemical plant in the Midlands for 11 years. In that time, I’ve made four properly significant maintenance and procurement mistakes, totalling roughly £38,000 in wasted budget. I keep our team’s checklist now. This was the mistake that started the list.
How a boiler feedwater valve became a problem
Our site makes its own process steam. Two gas-fired boilers feed most of the plant, and if one trips, we can’t run at full rate. The drum level on each boiler is controlled by a globe valve, and both valves were original Flowserve units — installed around 2007, well maintained, still doing their job.
The weak point wasn’t the valves. It was the old analogue positioners bolted to them. They had no diagnostics, no remote feedback, and the manufacturer had discontinued the range. When they failed, they failed without warning, and we had no way to see trouble coming.
In early 2022, an engineering audit recommended replacing all four positioners across the boiler island. I agreed. What I didn’t agree on, at least not at first, was how to buy them.
One quote came through Flowserve Flow Control UK, the local service channel. They proposed two Flowserve Fastraq digital positioners, the correct mounting kits, and two days of onsite commissioning. Total: £18,920.
The other quote came from a controls house we had used for years. They proposed an equivalent digital positioner with what looked like the same specification, the same HART protocol, the same approvals. Total: £14,870.
I don’t remember exactly why the numbers stuck in my head — maybe four thousand pounds, maybe four thousand one hundred, give or take — but I remember the gap clearly enough. It was roughly £4,000, and it was the only number my manager wanted to talk about.
Why I signed the cheaper quote
I’d like to say I fought harder. The truth is, I didn’t have a good answer when my manager asked the obvious question: “If the specifications match, why are we paying four thousand pounds more?”
I pointed to service coverage. I pointed to the Fastraq diagnostics. I pointed to the fact that the Flowserve quote included a real engineer coming to site. But when you put all of that on a spreadsheet, it looked like soft value. The cheaper quote had the same protocol, the same input signal, the same IP rating, the same temperature range.
My gut said the spreadsheet was missing something. My gut couldn’t tell me what. Every piece of data in front of me said we were buying the same function for less money, so eventually I signed the purchase order.
I even remember the risk calculation I did. Worst realistic case: a valve misbehaves, we call the vendor, they send a replacement part under warranty. Cost: a few days of inconvenience. Best case: we save £4,000 and nobody remembers the difference by Christmas.
The calculation didn’t include one variable: what happens when the plant is down and the vendor’s support line treats us like a small account.
The breakdown
We installed the two budget positioners in July. For the first eight weeks, they behaved. Calibration was straightforward, the plant ran normally, and I started to relax.
Then, in the early hours of that Saturday in September, the operator reduced boiler load. The feedwater valve should have tracked the new setpoint smoothly. Instead, it started drifting. The control loop fought it for almost an hour. By the time the supervisor called me, the high drum-level trip had already done its job.
When our instrument technician stripped the valve down, he found the problem. The positioner’s feedback linkage — the mechanical connection that tells the positioner where the valve stem actually is — had worked loose. A small fastening in the mounting kit had backed out under vibration. The positioner was still sending signals; it just no longer had a reliable idea of what the valve was doing. It was flying blind.
That is the part I still replay. It wasn’t an electronics failure or a clever software fault. It was a bracket. A cheap, badly chosen bracket, fitted to a critical steam valve, on a machine that vibrates all day. The spreadsheet didn’t have a column for that.
The week that followed
We called the controls house first. To their credit, they answered quickly — but their technical support worked through a Europe-wide desk, and the person on the phone had never seen our site or our setup. They asked us to run tests we had already run. They suggested configuration changes that made no sense.
Eventually, they agreed to send replacement parts. Estimated delivery: five to ten working days. Not expedited. Not even particularly apologetic.
That was when I finally made the call I should have made in the first place.
The Flowserve Flow Control UK engineer answered at 8 a.m. on Monday. She listened to the story, asked three sensible questions, and said they had the Fastraq positioner in stock. She also said, very politely, that mounting kits are not something you should improvise.
By Wednesday afternoon, a Flowserve service engineer was on site. He didn’t tell me he had warned us. He didn’t need to. He fitted the new Fastraq, checked the rest of the loop, and showed my technician why the original mounting geometry mattered. He also looked at the second valve — the one we had “saved” money on — and recommended replacing it too before it failed the same way.
We did. Both valves are still running today with no repeat of the problem.
What the whole thing really cost
The finance team put together a post-incident number. I won’t pretend every penny of it was caused by my purchasing decision — some of it was just bad luck on a Saturday night. But the total unplanned cost of that trip, the lost production, the reduced steam capacity, the overtime, and the emergency repair came to just over £27,000.
So we saved £4,050 on the purchase order and lost £27,000 on the incident. That is not a good trade by any measure.
I’m not saying budget positioners can’t work. I’m sure plenty of them run fine in less demanding applications. What I learned is that on a critical boiler feedwater valve, the price on the quote is the smallest part of the cost. The real cost includes support, spare parts, mounting quality, local response time, and the ability to pick up the phone and reach someone who understands your plant.
Price is what the invoice says. Cost is what shows up in the incident report.
What I changed in our purchasing checklist
That incident is now item one in the maintenance checklist I keep for our team. We don’t make a significant flow-control purchase without asking four questions:
- If this fails on a Tuesday, who answers the phone and how fast can they be here?
- Are the mounting kits and mechanical parts designed for this specific valve and actuator, or are they generic?
- What does the expected downtime cost per day, and how does that compare with the price difference?
- Who will commission it and what diagnostic tools will they bring?
Since we introduced that list in late 2022, we’ve caught five potential mistakes — including one “equivalent” positioner that didn’t actually have the hazardous-area approval the supplier claimed. In my experience, that kind of check is worth more than any spreadsheet comparison.
These days, I specify Flowserve Fastraq on our critical valves and use Flowserve Flow Control UK for the service side. Not because the brand name makes me feel safer, but because I know exactly what happens when something goes wrong at 2 a.m. — and I’d rather pay for that certainty before the failure than after it.
Leave a reply