Where Does The Time Go? Finding Wins on Your Extrusion Line

You don’t always know about downtime immediately.

It doesn’t always come with alarms, flashing lights, or burnt material smell.

Sometimes it’s just a moment: a sluggish feeder, a changeover that takes 42 minutes instead of 25, an operator wandering off during setup because no one showed him the right way to zero out the weigh scale when producing a day bin’s worth of dry-blend.

You don’t see those costs on a Work Order Traveler. But you pay them.

Every. Single. Shift.

The Plant That Couldn't Catch Up

We were running two good-sized profile lines with ‘sister parts’ in a medium-volume shop, one 3.5” single-screw and one 55mm twin. Both are pushing rigid PVC with about 15% reclaimed material.  Regrind into the single and pulverized is blended into the twin. Both were to have a similar matte finish. Looked good on paper.

But we couldn’t hit our throughput targets. Scrap was steadily accumulating, orders ran late, and downtime logs said we were saving in the low eighties percentile of the shifts. Only problem? Everyone on the floor felt that number was fiction.

Changeovers and re-starts dragged on a two,10-hour shift schedule. The Hopper Feeder surged at the start of almost every order. Operators were constantly “tuning the process” because the melt was wandering. Reclaim processing was chewing up more labor and power than it saved. And the scheduler? He was throwing darts every time he promised a ship date.

But none of that showed up in the daily metrics.

Why? Because our real-time drags weren’t being counted.

We weren’t tracking the cost of being "almost running."

Where Downtime Hides

We broke the shift into 15-minute chunks and started using real codes: set-up, running/saving, not-saving, purge, changeover, etc.

That’s when the time bleed showed up.

A short feeder delay? 15 minutes. A puller motor hiccup? 15 minutes. Changeovers that felt like 30 minutes were actually eating 75. And that “saving” time? We started catching on; some of it was out of spec or on the edge of control time, aka ‘near miss.’

Our reclaim setup was even worse. The time and labor to accumulate, cut up to fit into equipment throats, transport, shred, grind, pulverize, blend as necessary and reintroduce that quoted 15%? Nobody had actually costed it out for these profiles.

Turns out for short runs, we were spending more to make the reclaim work well than what we thought we were saving by using it.

It wasn’t killing us by the hour, it was grinding us down more slowly, like the relentless tic-tak, tic-tak, tic-tak of a die’s sealing surface being freshened up and ground away on the surface grinder –

Here’s Where the Fixes Started

1. Track Everything in Blocks, Then Call It Out

First step: block-based downtime reporting. Every operator filled in a shift log, 15-minute segments, coded by machine state. Nothing fancy. Just a basic standard shift report sheet and a pen to start.

We reviewed them in real time. That’s when the patterns emerged.

One operator’s line was "saving" 92% of the shift, but scrap said otherwise. Turns out, he was running hot and tweaking die pressure constantly to keep shape. That wasn’t uptime. That was “barely hanging on.”

You want throughput? You have to see what the machine is really doing, not just what the logbook says. 

2. Train to the Downtime, Not the Theory

We stopped guessing what training people needed.

Instead, we pulled the top downtime codes by shift and built quick, 10-minute huddles.

Feeder surge? Show them how to pre-prime. Changeovers slow? Walk through clamp sequencing, warm-up profiles, screw purges, all of it. Some guys had never seen a full teardown done right.

Training based on real downtime beats any SOP binder every time.

And pairing a seasoned hand with a newer operator? Worth more than any external seminar. 

3. Shift Maintenance from Firefighting to Pattern Recognition

Reactive maintenance feels heroic until you realize you’re just spinning plates; more keep being added, and some are now hitting the floor.

Our twin-screw had three unplanned stops in a month. All "minor": temp alarms, feeder stalls, haul-off alignment. Maintenance would clear the fault, log a 12-minute fix, and move on.

We cross-referenced the codes and saw the issue. Same motor drawing high amps on restart. Belt was slipping, but nobody had caught the trend.

We swapped it, realigned the drive, and haven’t seen a feeder fault since. Predictive maintenance starts with someone willing to ask: “Why does this keep happening on the same restarting schedule when operating two 10-hour shifts with a 4-hour break?”

The signs are always there. You just have to write them down; and wind back the clock. 

4. Stop Pretending Reclaim is Free

Reclaim always looks good on the balance sheet until you price in the energy, labor, and downtime to process it.

Our reclaim ratio was 15%. On paper, it was saving resin cost. But to reintroduce it? We ran a grinder, a shredder, and a pulverizer, using two part-time guys feeding and bagging or blending in day-bins.

And when it wasn’t preconditioned in staging in the winter properly? Ratholing and Feeder surge. Vent puke due to varying output based on time not loss in weight feed into starve fed screws, Blown shape. Extra startup scrap.

We ran the math. True cost per pound of reclaim before equipment upgrades was just 6% under virgin, and the hidden downtime from using it made it a net loss on any short run.

We cut reclaim to 5% for short orders and reserved 15% for long, stable campaigns. Cut waste. Cut downtime. Kept quality.

Turns out, reclaim isn’t the hero. Controlled inputs are.

The Bigger Payoff

Once we got honest with our data, things shifted. Not all at once, and not cleanly.

But changeovers sped up by more than a dozen and a half minutes on average. Feeder surge events dropped by almost two thirds. Scrap fell by 11%. And, most importantly, we could finally trust the line reports.

That meant scheduling stopped guessing. Maintenance stopped chasing ghosts. Operators started asking sharper questions because they could see what good looked like.

It wasn’t a Lean rollout. It wasn’t an ERP integration. It was just one hard look at how downtime hides in plain sight, and the decision to track what really matters.

Final Word

If you’re still measuring your extrusion line’s performance by “percent of shift saved,” you’re already paying the tax. You just haven’t totaled the invoice.

Want to boost throughput, cut waste, and tighten your quality?

Start on a line or two by tracking every 15 minutes and the WHY it was lost, like it costs you something, because it does.

Every unsaved block of time is lost margin. Every shortcut is compounding risk. Every “we’ll fix it later” moment is a crack in the foundation.

Start small. Get real. Ask better questions.

That’s where the time goes and the wins are hiding. 

#Manufacturing #WorkforceDevelopment #ProcessImprovement #ProductionEfficiency

Stick around, there’s a 3-part follow-up series in the wings that builds directly on “Where Does the Time Go?”, and together they outline more details for deeper accountability, sharper diagnostics, and smarter resource use in thermoplastic extrusion.