Three jobs. Three times the details got skimped on, complacency undermined verification.
The people in these stories are not bad technicians. They are experienced, capable people who knew their equipment. That is actually the whole point — because it means there is no clean excuse here. No missing information, no ambiguous marking, no procedure nobody had ever seen before. Just good people who let their attention slip or were overconfident at the wrong moment, on jobs that felt too familiar to warrant a second look.
Three incidents. Three times direction of flow was assumed rather than confirmed. Three times we paid for it in scrap and downtime that did not have to happen.
Incident One
A Brand New Pump That Shouldn’t Have Been Leaking
We were installing a melt pump. The flow direction arrow is stamped right into the casting — not hidden, not worn off, sitting there plain as anything. It was not checked.
The frustrating part is that gear pumps will run either way. They are positive displacement. The gears turn, material moves, head pressure builds. Standing at the control panel you would have had no reason to think anything was wrong. The line looked like it was running.
Then the seal started weeping. A brand new pump, straight out of the box, and it was leaking at the shaft seal area. As flow increased the leakage got worse. That made no sense to us — you do not expect a new pump to fail like that out of the gate. We went looking for a defect, a bad seal, a manufacturing issue. It took longer than I care to admit before someone thought to check the flow direction arrow on the casting and found the pump was installed backward in the line.
The shaft seal belongs on the low-pressure inlet side. We had it against full discharge pressure. Of course it was leaking. It had no choice.
Pump came out, went back in the right way, ran perfectly. What we lost was the time chasing a defect that did not exist, the time to correct the installation, and the material that ran while we were doing both. The installation was never verified.
Incident Two
Four Extruders. The Last One Had Crossed Cooling Lines in the Feed Section.
We were acid flushing feed sections on grooved feed extruders. Acid flush, lines off, clean, lines back on. Four extruders that day. The first three went without a hitch. By the time the crew got to the fourth one the job had become completely automatic. They had done it three times already. They knew the procedure. They were moving efficiently.
Three for three. Nobody was thinking about the fourth one the same way they thought about the first.
Inlet and outlet went back on swapped. The technician knew which port was which — that is not in question. The problem is that by the fourth machine the job did not feel like it needed checking. It felt done before it was finished.
The grooved feed section needs the coldest water arriving at the barrel end first. That is where the heat is greatest and where temperature governs solids conveying. Reversed connections meant warm water — already heated from its run through the jacket — was arriving at exactly the wrong place. |
Feed section temperature climbed. The compound started softening earlier than it should. Pellets stopped conveying cleanly, bridging started in the grooves, output went unstable. We generated real scrap before anyone looked at the cooling connections.
We caught it the same shift. Even a short run with crossed lines on that section produces waste you did not have to produce. The fix was a few minutes. The scrap was already sitting there. Three machines done perfectly, one undone by the comfort of repetition.
Incident Three
The Screw That Ran Backward
We replaced a motor and started the extruder without confirming rotation direction. The screw ran backward.
Every extruder technician knows what reversed phase connections do to a motor. The dry-run check — no material, no hopper loaded, confirm the screw is turning the right way before anything else happens — exists precisely because of this. It is not an obscure step. We skipped it.
Pellets do not feed into the barrel when the screw runs backward. They come back out.
The check takes thirty seconds. No load on the screw, nobody near the feed throat, confirm rotation before you introduce material or people into the area. We knew this. We did not do it. The material that came out of that feed throat was waste we created by skipping a half-minute verification on a job we had done before. |
Wiring corrected, controlled restart, line back in service. The material and time lost is gone.
The FIX
Knowing Is Not the Same as Checking
I have more stories, but these three are painful enough. What I keep coming back to is that they are all the same failure, the job felt known, so the confirmation step felt unnecessary.
That is the part of this work that is hard to accept, because you cannot fix it by training harder or knowing more. You fix it by building the verification in, so it does not depend on how the job feels that day.
Check the flow arrow on every pump before commissioning. Every one. The tenth install of the day gets the same check as the first.
Tag fluid connections before any flush or disassembly. Do not trust memory at the end of a long job on the fourth machine.
Confirm rotation direction with a no-load dry run after any drive work. No exceptions, no matter how familiar the job feels.
Put the verification on a checklist as a signed step. Remove the option to rely on confidence instead of confirmation.
Frustrating situations because the knowledge was there, the experience was there, the work was done, but it wasn't done right,,, because it was not verified.
A checklist for simple items may seem like overkill ... but it would have prevented these three avoidable moments of downtime.
Discipline should become habit.
Experience tells you what to check. Discipline is what makes you check it .
Had a job go wrong not because of missing knowledge but because familiarity switched off the attention? I would like to hear it. Drop it in the comments — these are the stories that actually teach us all something.
The Fix — One real problem. One real solution. Every week on Plastic Extrusion Problem Solving.