Same Screw, Different Rules

The equipment looks the same. The job isn't.

Move from industrial extrusion into medical tubing and the basic pieces don't change much, screw, barrel, die, puller, all stuff you already know. What changes is everything built on top of that. The materials you're allowed to use, how tight you have to hold things, how you control the process, the paperwork, even the room you're standing in.

Industrial extrusion is built around getting a lot of good parts out the door as cheap as possible. Medical is built around proving, every single time, that the part you made is the part you meant to make. That trade-off shows up everywhere. I've grouped it into three buckets: stuff that's different on the line, stuff that's different in the room, and stuff that's different in the paperwork.

On the Line

1. Micro pellets for micro screws. If you're running a screw an inch or under, you probably already know regular pellet size gives you fits. It bridges in that small feed throat, doesn't meter evenly, and starves the screw before it ever gets a good melt going. Most people figure this out the hard way, chasing what looks like a screw problem for half a shift before somebody checks the pellet size.

2. A lot of these lines skip the gear pump. On industrial equipment a gear pump is your answer when melt pressure isn't consistent. On the small medical lines, especially short runs, plenty of shops just don't bother. The added residence time, shear, and cleanout aren't worth it for a run that short. You end up controlling pressure and output through the screw and process settings instead.

3. The resin has to be dry, and I mean actually dry. Pebax, polycarbonate, stuff like that pulls moisture out of the air, and it needs real desiccant drying, sometimes four to six hours, before it's anywhere near the hopper. A lot of industrial resin is forgiving enough that you could get away with a quick pre-dry or skip it entirely. Skip it in medical and you'll see splay, brittleness, dimensional drift, the works, and you'll spend a whole shift chasing it before someone finally pulls a moisture reading.

4. No regrind. None. Industrial extrusion, you regrind your scrap and feed it back in, that's just normal cost control. In medical, it's virgin resin only. Full stop. Regrind means uncontrolled thermal history and traceability gaps, and that's not something you can have when the tubing might end up inside somebody. Losing that “free” material is one of the bigger mental adjustments guys make coming over.

5. Bump and taper is a puller trick, not a tooling trick. Bump and taper tubing, where the OD steps or tapers along one continuous piece, doesn't really exist in industrial work since you're always chasing one constant profile. First time I saw it I figured it had to be special die tooling. It's not. The die and melt stay exactly the same. It's the puller doing the work, running off a programmed speed profile that draws the tube down faster or slower to shift the diameter as it goes. Industrial never needs that because industrial never wants the diameter to change.

6. Durometer transitions along one shaft. Catheter shafts often need to go stiff at one end so they push well, and soft at the other so they don't tear anything up, with a smooth change in between instead of a hard seam. Getting that gradient usually takes a tandem or co-extrusion setup built specifically for blending durometers, and there's nothing like it on the industrial side. It's specialized enough that only a handful of equipment builders really do it well. Gimac, out of Italy, is one of the names that comes up a lot for this kind of micro-extrusion work. I haven't run their equipment myself, so take that as a starting point to look into rather than a recommendation.

7. Multi-lumen means juggling air in every channel at once. Extrude a multi-lumen tube and now you're controlling internal air or vacuum separately in every lumen, at the same time, because each one wants to collapse or balloon or drift differently as it cools. Single-lumen tubing never asks this of you. You've got one profile to manage, not four fighting each other.

8. Line speed is your real dimensional control. Industrial resin tends to be consistent enough lot to lot that you set your process and walk away. Medical resins like TPU and Pebax move around more than people expect, viscosity shifts, moisture sensitivity is higher, and your specs are tighter to begin with. So instead of trusting the resin to behave the same way every time, you're riding line speed, sometimes second to second, to hold your dimensions.

9. Some lines run straight up and down. For thin wall tubing, running the line vertically instead of horizontally cuts down on sag and keeps the wall more even. It's a real process reason, not a floor space thing, and it throws people the first time they walk up to a line running top to bottom instead of side to side.

10. Short runs don't give ultrasonic gauging much room to work. Ultrasonic wall thickness gauges need time to settle in, thermal stability, consistent coupling, a steady line. On a genuinely short medical run, that settling time can eat up a good chunk of the job before the gauge is even reading right. Doesn't mean the tech is bad, just means a short run might not give it enough runway, so plenty of shops fall back on manual checks, gravimetric, or laser micrometers instead.

11. Concentricity gets measured on its own. Medical tubing tracks how centered the ID is inside the OD as its own number, often in the ninety-plus percent range, tighter still for precision balloon tubing. An off-center wall can mean uneven burst strength or a device that just doesn't act right. On the industrial side, if OD and average wall are in spec, nobody's usually asking where the center sits.

12. Vision systems catch stuff industrial QC would wave through. Automated inspection runs at speed looking for gels, die lines, carbon specks, tiny scratches, things that would sail past an industrial QC check without anybody blinking. In medical those same flaws can affect how the device performs or how safe it is, so they get flagged.

In the Room

13. Somebody's watching the tanks for bacteria. Water tanks for sizing and cooling are a bioburden risk that industrial extrusion mostly ignores. In medical, tank cleaning, water quality, biofilm control, all of that is on a documented schedule. Nobody's waiting until the water looks cloudy to deal with it.

14. Clean rooms and gowning. Depending on classification, you might be running in an ISO-classified clean room, gowned head to toe, with controlled foot traffic and particulate monitoring going the whole shift. Industrial floors are loud and open and dusty by comparison. Walking into a gowning room before your shift for the first time is usually the moment it really hits you that you're not in Kansas anymore.

15. The air itself is spec'd, not just what you're wearing. Past the gowning, the room is held to its own spec, ISO class, HEPA filtered positive pressure air, particulate monitoring running constantly, tight temperature and humidity bands. Then you've got double-bagging just to move product out of the room. It's not only about what you've got on. The building itself is working around the clock too.

In the Paperwork

16. IQ, OQ, PQ. Installation, Operational, and Performance Qualification aren't a checklist you knock out once. IQ shows the equipment's installed right. OQ shows it runs within spec across its range. PQ shows it makes good parts consistently under real production conditions. Industrial has quality checks. Medical has a whole documented, auditable lifecycle behind every piece of equipment and every process.

17. Sample size isn't up to you. How many parts you pull and measure comes from a statistical rationale tied to your validation protocol and risk classification, not a gut feel. An auditor can ask you to justify that number, and “felt like enough” isn't an answer.

18. Protocol, verification, and validation aren't the same word. On an industrial floor guys use these words pretty loosely. In a regulated shop they mean specific, different things. A protocol is the plan you write before you do the work, laying out how you'll judge success. Verification means you've got objective evidence a requirement was actually met. Validation means the process consistently makes product that does what it's supposed to do. Mix those up in your documentation and that's an easy way to pick up a finding.

19. ISO and FDA aren't background noise. ISO 13485 and FDA's 21 CFR Part 820 shape how you document changes, handle nonconformances, manage your supply chain, how long you're keeping records. Industrial answers to a customer spec sheet and maybe ISO 9001. Medical answers to a regulator that can shut your line down.

20. DOE work is heavier than what you're used to. Design of experiments gets used a lot to understand how temperature, line speed, and screw RPM play off each other, so the process window you land on is backed by real data instead of “this is how we've always run it.” It's a bigger statistical lift than most industrial process development because that window has to hold up under a validation audit, not just make good parts today.

21. Biocompatibility testing is on the material, not the finished part. USP Class VI, ISO 10993 testing for things like cytotoxicity and sensitization, extractables and leachables studies, all of it is trying to answer one question: what could come off this tubing and into a patient. Industrial cares if the resin hits a mechanical spec. Medical cares what the resin gives off.

22. Every lot gets its own paperwork. Full traceability means every spool ties back to a specific resin lot, run date, operator, and process settings, with a Certificate of Analysis for that batch. Industrial might tie a run to a work order number and call it done. Medical builds a trail that's got to hold up years later if someone comes asking.

23. DHR and DMR are two different animals. The Device History Record is what actually happened on a given batch, who ran it, what the settings were, what got measured, any issues. The Device Master Record is what's supposed to happen, the specs, the methods, the inspection steps. Industrial usually doesn't separate the recipe from the proof you followed the recipe this cleanly. Medical keeps them as two separate, permanent records.

24. Cpk decides if you're even allowed to run. Most shops are targeting a Cpk around 1.33 or better, and 1.67 or better on anything critical, and that number can be the difference between a process being cleared for production or not. It's not about whether today's batch happened to come out fine. Control charts tracking drift over time become part of the daily routine, not something you glance at once a quarter.

None of this makes medical extrusion harder in the sense that running a screw and a die suddenly gets more technical. It's harder because the bar for “good enough” moves. On the industrial side, good enough means the part measures in spec. On the medical side, good enough means the part measures in spec and you can prove exactly why, every time, to somebody who wasn't even in the building.