Poolp Uses 3D Printing to Turn Scrap Into Parts

• 1
  The shred-to-print chain when Poolp uses 3D printing
• 2
  Which plastics survive the loop, and which do not
• 3
  Print settings that make or break a recycled part
• 4
  What a recycled part can and cannot do
• 5
  How to tell a good recycled part from a bad one
• 6
  Recycled filament vs virgin filament vs machined stock

The core idea is simple: take post-consumer or post-industrial scrap, reduce it to consistent granules, melt it, and re-form it as filament or pellets that a printer can consume. Nothing about the process is exotic. The difficulty is consistency, not chemistry.

Grinding comes first. A shredder cuts bottles, sprues or failed prints into flakes, then a granulator brings them down to roughly 3–5 mm. Flake size matters because the extruder screw feeds on volume, not weight. Mixed flake sizes cause surging, and surging shows up later as diameter drift in the filament.

Drying is the step most small setups skip. Most commodity plastics absorb moisture from air. PET and PA are the worst offenders. Wet pellets create steam inside the barrel, and steam creates bubbles and weak weld lines in the printed part. A 4-hour dry at 70–80 °C for PET, or 80 °C for PA, removes most of it.

Extrusion turns flake into filament. The screw melts the polymer, a filter screen catches unmelted bits and paper labels, and the die sets the final diameter. A puller and a laser gauge hold that diameter in a closed loop. Without the loop, you get spool-to-spool variation that no slicer setting can fix.

Not every plastic tolerates a second heat cycle. Each melt step shortens the polymer chains, so the material gets stiffer and more brittle. PLA might take two or three passes before print quality drops off. PET and PP behave differently, and PVC should be kept out of the loop entirely because it releases hydrogen chloride when overheated.

The practical split comes down to how the plastic was made and how clean it is. A single-polymer stream — one resin code, one color, no labels, no caps, no glue — recycles well. A mixed stream does not. If you cannot identify the resin, do not feed it. Unknown plastic is the fastest way to jam a nozzle and ruin a batch.

Additives change the picture too. Glass-filled or carbon-filled grades are abrasive. They wear a brass nozzle in a few kilograms. Switch to hardened steel or ruby if you plan to run them. Flame-retardant grades often contain fillers that degrade at normal print temperatures, so check the datasheet before you commit.

Color is the visible signal of degradation. Clear or light-colored scrap turns yellow after repeated heating. That yellowing tracks with chain scission, so it is a rough quality gauge. If the filament goes dark brown, the batch is done. Do not try to print through it — the part will crack at the layer lines.

Recycled filament prints hotter than virgin material of the same grade, because the melt flow index drifts upward after each pass. A typical bump is 5–15 °C above the virgin recommendation. Start at the top of the virgin range and adjust from there. If layers look glossy but do not bond, you are running too cold.

Flow rate needs calibration per spool, not per material. Measure the actual filament diameter in three places, average them, and enter that number into the slicer. On a 1.75 mm nominal filament, a 0.05 mm error changes the extruded volume by about 6%. That is enough to cause over-extrusion on the outer walls.

Layer adhesion is the weak point. Recycled polymer has shorter chains, so interlayer bonding is lower. Raising the nozzle temperature helps, and so does slowing the print. Dropping from 60 mm/s to 35–40 mm/s for the outer wall usually recovers most of the lost bond strength. Enclosing the printer for ABS or ASA also helps.

Cooling should be reduced, not increased. Fast cooling locks in stress and promotes warping. For a recycled part, run the part cooling fan at 30–50% of what you would use for virgin material, and keep the chamber warm if the printer allows it. The goal is slow, even solidification.

Recycled filament is not a drop-in replacement for virgin resin in a load-bearing application. Tensile strength typically drops, and elongation at break drops more. A part that needs to flex repeatedly will fail sooner. That rules out living hinges, snap fits with high strain, and anything that sees cyclic stress.

Where it works well is in non-critical geometry: jigs, fixtures, brackets, enclosures, cable guides, display stands and prototypes. These parts carry light loads, tolerate some brittleness, and do not see heat above 60 °C. Housing and furniture hardware are common uses, and so are shop-floor aids.

Surface finish also differs. Recycled parts often show more visible layer lines and occasional dark specks from contamination. That is cosmetic, not structural. If the part is customer-facing, plan on sanding, filler primer or a coating step. Budget for that in the process plan, not as an afterthought.

The honest boundary: if the part must meet a documented spec — a pressure rating, a flammability rating, a medical or aerospace requirement — do not print it from unknown recycled stock. Use certified virgin material, or move the geometry to CNC machining where the material cert is traceable. Poolp uses 3D printing for value, not for certified structural work.

You cannot inspect recycled parts the same way you inspect machined parts, because the variability is in the material, not just the geometry. Start with the filament. Measure diameter at the start, middle and end of the spool. If the spread is more than 0.05 mm, the spool is not ready to print.

Then check the print. Look for gaps at the top surface, which signal under-extrusion. Look for blobs and stringing, which signal moisture or too-hot melt. Snap a test bar and look at the break. A clean snap with a dull surface means the layers bonded. A break along layer lines means they did not.

Weight and dimension checks catch the rest. Weigh the finished part against the slicer estimate. A part that comes out 5% heavy usually has over-extrusion, which means the walls are thicker than the model. A part that comes out light has voids. Both are reasons to recalibrate before the next run.

For anything that matters, cut a cross-section and look at it under low magnification. Voids, unmelted flake and contamination all show up. This is a destructive test, but it is the only reliable way to confirm that a new scrap stream is actually printable before you commit a production run to it.