Flexible filament has a reputation problem, and it is not entirely deserved. The material is blamed for failing in machines where the real problem is the path the filament travels. Understanding the difference between a durometer rating and an extrusion system is what turns TPU from a fight into a routine print.
This guide covers how shore hardness actually translates into printed behaviour, why some printers handle flexible filament and others cannot, and the settings that matter once the hardware question is settled.
Shore Hardness Is a Range, Not a Number
The first thing to understand is that a shore rating describes a measurement method as much as a material. The durometer scale comes in several variants, and two of them matter for filament.
Shore A is used for softer materials and is measured with a blunted indenter under a lighter spring load. Shore D is used for harder materials with a sharp indenter and heavier load. The scales overlap slightly between roughly 90 A and 60 D, which is why you occasionally see the same material described both ways.
Flexible 3D printing filament lives almost entirely on the shore A scale, and the useful working range runs from about 85 A down to roughly 70 A. Below that, the material stops behaving like a filament and starts behaving like a wet noodle, which is a hardware problem rather than a settings problem.
What Each Range Feels Like
| Shore rating | Character | Typical use | Printer demand |
|---|---|---|---|
| 98 A – 95 A | Firm, slightly compliant; bends without cracking | Phone cases, wear pads, vibration dampers, living hinges | Easy — prints on most direct-drive machines |
| 95 A – 90 A | Genuinely rubbery; returns to shape after compression | Gaskets, seals, grips, flexible hinges, shoe soles | Moderate — direct drive strongly preferred |
| 90 A – 85 A | Soft, deforms under finger pressure | Shock absorbers, soft grips, prosthetic liners, drone bumpers | Demanding — short direct path essential |
| 85 A – 80 A | Very soft, drapes under its own weight | Squishy props, medical models, soft robotics prototypes | Hard — needs a short, well-guided path and slow speeds |
The single most common mistake is choosing the softest material available on the assumption that softer means more flexible. Softness is not a value judgement. A gasket that compresses too easily will not seal; a phone case that deforms under normal handling is worse than a rigid one. Pick the rating that matches the force the part needs to resist, then let the printer dictate whether that rating is practical.
Why Flexible Filament Fails in Some Machines
This is where most of the frustration originates, and it is a mechanical problem with a clean explanation.
Every extruder pushes filament by gripping it between a drive gear and an idler and forcing it down a constrained path into the melt zone. With rigid filament, the material itself acts as a pushrod: it transmits the drive force from the gear to the nozzle with almost no loss, because it will not buckle under the compressive load.
Flexible filament buckles. When the drive gear pushes, a soft filament deforms laterally before it moves forward, and the deformation has to go somewhere. In a long or loose path, it goes sideways into the gap between gear and guide, forming a loop or a coil that jams everything behind it.
Two geometry variables decide whether this happens:
- Path length. The longer the distance from the drive gear to the melt zone, the more opportunities the filament has to deflect. A Bowden tube setup with a metre of PTFE between extruder and hotend is close to worst case. A direct-drive extruder sitting directly on the toolhead with a few millimetres of guided travel is close to best case.
- Path constraint. A filament that is fully captured in a tube of matched internal diameter cannot buckle sideways, because there is no room to go. A path with gaps, unsupported spans, or an oversized guide bore gives it somewhere to escape.
This is why the same spool of 95 A filament prints cleanly on one machine and produces a jam on another with no settings changed. The material did not change; the escape route did.
The Retraction Trap
Retraction deserves separate attention because it attacks flexible filament from the opposite direction. Retraction works by pulling the filament back to relieve pressure in the melt zone. With rigid filament, that pull releases stored compression and the material springs back into place cleanly.
With soft filament, the pull has to overcome friction along the entire path, and a soft filament stretches rather than sliding. The material thins out, the drive gear loses grip on a stretched section, and the filament grinds. Worse, the stretched section then relaxes unevenly, producing inconsistent extrusion on the next few moves.
The practical response is to treat retraction as a last resort:
| Setting | Rigid filament | Flexible filament | Reason |
|---|---|---|---|
| Retraction distance | 0.8–2 mm (direct) / 3–7 mm (Bowden) | 0.2–0.8 mm, or off entirely | Avoid stretching the filament |
| Retraction speed | 35–60 mm/s | 10–25 mm/s | Slow pull reduces grinding |
| Print speed | 60–200 mm/s | 15–30 mm/s | Back pressure rises faster than grip |
| Flow rate | 100 % | 95–105 %, calibrate per spool | Soft material over-extrudes easily |
| Travel moves | Retraction-based | Minimise count in the slicer | Fewer retractions, fewer failure points |
On a well-tuned direct-drive machine, disabling retraction entirely and accepting a small amount of stringing is often the more reliable choice than fighting for a clean surface. Stringing on a flexible part is easy to trim; a mid-print jam is not recoverable.
Design Constraints That Follow From the Material
Flexible parts need to be designed differently, because the printing constraints and the service loads are both different.
Wall Count Beats Infill
On a rigid part, walls and infill contribute to stiffness in roughly comparable terms, and infill percentage is the familiar knob. On a flexible part, that relationship inverts. Perimeter walls form a continuous skin that resists localised deformation, while infill is a lattice that bends easily. A part with four walls and 10 % infill feels considerably firmer than the same part with two walls and 40 % infill.
Design the feel of a flexible part with wall count first. Use infill to fine-tune, and keep in mind that sparse infill also lets the part compress more before it bottoms out, which is what you want in a bumper or a damper.
Compression Needs Room to Go
A gasket or seal only functions if the material has somewhere to displace to when compressed. If the cavity it sits in is exactly its own volume, the material becomes incompressible, and the load transfers straight through instead of sealing. Design a deliberate compression allowance, typically 15–30 % of the gasket cross-section for the working range of the joint.
Thin features also matter. Very thin flexible sections tear at layer lines under repeated flexing, because the bond between layers is the weak plane. Anything that will cycle in service should be at least three or four perimeters thick, so the load is carried by continuous extrusion rather than by inter-layer adhesion.
Layer Orientation Is a Wear Decision
A part that flexes repeatedly will eventually crack along the layer boundary, because that is where the material is discontinuous. Orient the part so the flexing direction runs within a layer rather than across layers. For a living hinge, that means printing the hinge so its bending axis lies flat on the plate and the extrusion lines run along the length of the hinge. Printing the same hinge standing up puts the entire bending load across the weakest plane in the part, and it will fail.
Solving Adhesion Without Fighting the Part
Flexible materials bond to build surfaces differently from rigid ones, and the usual remedies make things worse. A smooth surface that grips rigid filament aggressively will do the same to a flexible part, and the removal force required is higher because you cannot pry a soft part the way you pry a stiff one.
The approach that works is to reduce adhesion deliberately rather than increase it:
- Use a textured or smooth PEI surface at moderate temperature. Flex plates release a soft part by bending, which peels the print off gradually instead of requiring a lever.
- Skip the adhesive for most parts. Flexible filament is tacky enough on its own. Adding glue stick on top turns release into a genuine problem, and it also contaminates the first layer.
- Widen the first layer slightly. A 105–110 % first layer width compensates for the material spreading under nozzle pressure and gives a more consistent bond.
- Slow the first layer considerably. 10–15 mm/s for the first layer prevents the nozzle from dragging soft material and gives the bond time to form.
- Let the part cool fully before removal. Warm flexible parts tear easily at thin sections. Patience is cheaper than a reprint.
If parts are still lifting, the cause is usually contamination on the plate rather than insufficient adhesion. Cleaning with warm water and a mild detergent, then handling the plate only by its edges, fixes more lift problems than any adhesive.
Where Flexible Material Is the Only Correct Answer
It is worth being clear about when the added process difficulty is justified, because flexible filament is not a general-purpose upgrade.
| Application | Why flexibility is required | Practical rating |
|---|---|---|
| Gaskets, seals, O-ring replacements | Must compress to fill an irregular gap and recover | 85 A – 90 A |
| Vibration dampers and feet | Absorbs energy rather than transmitting it | 90 A – 95 A |
| Grips and wear surfaces | High friction and impact resistance at contact points | 95 A |
| Living hinges in an assembly | Survives thousands of flex cycles where rigid material cracks | 90 A – 95 A |
| Impact bumpers and guards | Deflects and returns rather than fracturing | 85 A – 90 A |
| Wearable and prosthetic interfaces | Comfort requires compliance against skin | 80 A – 85 A |
Where flexible material is the wrong answer: anything structural, anything that must hold a precise dimension under load, and anything with fine detail that a slow print speed would make impractical. For those jobs a rigid material with a compliant design feature is almost always the better engineering choice than a soft material doing a rigid job.
Hardware Choices That Change the Difficulty
If flexible printing is going to be a recurring part of your work rather than a one-off, the hardware decisions matter more than the settings. Three of them are worth thinking through.
The feed path. A short, fully constrained path from drive gear to melt zone is the single highest-impact factor. If your machine is Bowden, flexible filament below about 95 A will be difficult regardless of tuning, and the honest answer is to print it on a different machine.
The nozzle and hotend. Flexible material needs a pressure-stable melt zone more than it needs high flow. A worn or oversized nozzle lets soft filament spread unpredictably, and abrasive-filled flexible grades will wear a brass nozzle quickly. A hardened tungsten nozzle holds its bore far longer under abrasive filament, which keeps extrusion consistent across a long spool.
Filament conditioning. Flexible materials are not as moisture-hungry as PETG, but they are not immune. Moisture in a soft filament turns into steam in the melt zone, which produces inconsistent extrusion that is easy to misdiagnose as a mechanical problem. Drying before a long print removes the variable, and a drying cabinet also keeps the spool conditioned between sessions, which matters more for flexible material because it tends to sit in the open longer than a fast-consuming PLA spool.
If you are building up a flexible capability from scratch, the efficient order is to start with a firm grade that prints easily and learn the adhesion and speed behaviour on a material that will not punish small mistakes. A 95A high-speed grade is a reasonable starting point because its higher stiffness tolerates a wider range of setups, and its formulation allows higher print speeds than standard TPU. Once the workflow is stable, moving down to an 85A grade becomes a settings adjustment rather than a new skill. For a softer result, a lightweight 90A grade lowers the density while keeping the print process manageable.
Where the flexible part has to interface with a rigid one, surface preparation matters more than usual. Soft material held against a smooth plate tends to slide and release unpredictably, so a cold PEI plate that holds firmly at low bed temperature gives a more consistent first layer on soft grades without requiring adhesive. And when a flexible part is going into an assembly with rigid components, the feed path is not the only thing that needs support. A filament buffer and auto feeder keeps tension constant on a soft spool, which prevents the spool from unspooling loosely and tangling between prints, a failure mode that is far more annoying with flexible material than with rigid.
The Workflow, Start to Finish
Pulling the pieces together, a reliable flexible print follows a sequence:
- Match the rating to the load. Decide what force the part must resist before choosing material.
- Confirm the machine can feed it. Direct drive and a short path, or accept that sub-95 A will be difficult.
- Dry the spool. Remove moisture as a variable before touching settings.
- Cut the speed hard. 15–30 mm/s. Flexible filament does not reward rushing.
- Reduce or disable retraction. Accept some stringing over the risk of a jam.
- Add walls, remove infill. Control firmness with perimeters, not density.
- Orient for flex direction. Keep bending loads inside layers wherever possible.
- Remove the part while it is still warm. Soft parts peel off flexible plates best before they stiffen.
Do that and flexible filament stops being a material people warn each other about. It becomes what it actually is: a specific engineering material with a specific set of process requirements, no more mysterious than any other polymer once the feed path is understood. The prints that come off the plate are parts nothing else in FDM can make, and that is the reason to learn it properly rather than to keep avoiding it.