TPU Print Settings Basics: Starting Values for 95A
TPU offers rubber-like flexibility and impact resistance, but it can feel surprisingly tricky if you approach it with PLA habits. If you want to start printing TPU on an Ender or Bambu Lab machine, the least failure-prone approach is to dry your TPU 95A thoroughly, set print speed to 20mm/s, and keep retraction (the motion that pulls filament back from the nozzle) conservative. When I first printed a phone case and desk leg caps in TPU 95A, I ran the same speed and retraction settings I used for PLA. The result was stringing everywhere and a rough surface. Simply adding a drying step and lowering the speed stabilized everything dramatically. TPU is not some exotic specialty material -- adjust your starting assumptions and it becomes perfectly manageable. This article covers how to tell whether your printer is suited for TPU, the specific settings to start with, and how to address symptoms like stringing, surface blemishes, and jams with number-based adjustments. Rather than stumbling through PLA-based trial and error, getting the right starting point for TPU is the faster path.
What Is TPU Filament? How It Differs from PLA
TPU Is a Type of TPE
TPU belongs to the TPE (thermoplastic elastomer) family. TPE is a broad category of materials that can be shaped with heat and behave like rubber once cooled. Among them, TPU is the grade most widely available as 3D printer filament and the easiest to work with. Formlabs' TPU guide also categorizes TPU as a representative TPE known for its flexibility and resilience.
The relationship is straightforward: TPE is the umbrella category, and TPU is a specific material within it. Some products are labeled simply "TPE filament," but what actually gets used in FDM/FFF printing is mostly TPU around the 95A hardness range. You will often see Shore A hardness referenced here. The lower the number, the softer the material: 85A is quite pliable, 95A is a comfortable starting hardness for FDM, and 98A holds its shape noticeably well.
In my experience, 95A feels best described as "firm rubber." Press it with your finger and it gives slightly, but release and it snaps right back. A PLA case has that rigid shell feel when you grip it, while a TPU case conforms just slightly to your palm and distributes pressure differently. Even with identical geometry, the moment you pick it up the impression changes considerably. TPU is not just about appearance -- it changes how something feels in your hand, from the softness of contact to a subtle sense of security.

TPUを使った3Dプリント完全ガイド
TPU材を使った3Dプリント方法の比較、ラピッドプロトタイピングやカスタムの実製品用部品としてTPU製部品を3Dプリントする方法をご紹介します。
formlabs.comStrengths and Weaknesses of TPU
TPU's main strengths are flexibility, impact resistance, and abrasion resistance. It resists cracking from drops or compression and handles friction reasonably well, making it ideal for anything you touch constantly, bump into, or need to flex slightly. It also absorbs vibration, so it pairs well with cases, cushion feet, and snap-fit components. If PLA is "a rigid material that holds crisp shapes," TPU is "a functional material that absorbs motion and impact."
On the other hand, weaknesses become apparent when you treat TPU like PLA. Its softness means it flexes during extrusion, and print behavior differs significantly from rigid filaments. Bridges and overhangs in particular suffer because the material sags under its own weight, so they will not come out as sharp as with PLA. TPU also absorbs moisture readily -- when it does, expect surface blemishes, increased stringing, and a bubbly roughness.
Narrowing the PLA-vs-TPU comparison to the points that matter during printing:
- Flexibility: PLA is rigid and holds shape easily; TPU bends and returns
- Impact resistance: PLA cracks under heavy impact; TPU absorbs it
- Abrasion resistance: TPU outperforms PLA in most friction scenarios
- Bridges and overhangs: PLA has the advantage; TPU sags and roughens more easily
- Moisture sensitivity: TPU demands more care; drying quality shows directly in the finish
Standard TPU benefits from low speeds, and Flashforge's documentation also recommends starting around 20mm/s. High-speed variants are a separate category -- products like PolyFlex TPU95-HF claim 100mm/s and above, but these high-flow types should be considered independently from standard TPU.
Moisture makes a significant difference with TPU as well. Drying guidelines commonly suggest 50-60C for 4-6 hours, though some products like BigRep TPU 98A specify 80C for 4-6 hours. There is no one-size-fits-all number -- follow the manufacturer's recommended drying conditions for each filament.
For a quick visual comparison:
| Attribute | PLA | TPU |
|---|---|---|
| Flexibility | Low | High |
| Impact resistance | Medium | High |
| Abrasion resistance | Low-Medium | High |
| Moisture sensitivity | Low-Medium | High |
| Print difficulty | Low | Medium-High |
Use Cases: Where TPU Shines and Where It Struggles
TPU excels with parts that should not be rigid and parts that will take hits. Gaskets, caps, protective cases, cable bushings, and furniture cushion feet are classic examples. With some wall thickness, it works for shoe sole prototypes, and the combination of rebound and wear resistance makes it popular for RC tires. Grip components where you want a softer feel and covers that should not scratch what they protect are also strong fits.
What makes TPU useful in these applications is not just softness -- it is the rebound. It recovers after compression and tolerates repeated deformation to a degree, functioning somewhere between a cushioning material and a structural part. I have found that simply swapping PLA for TPU on desk leg caps or small cases eliminates that hard, clacky contact and makes the finished piece noticeably more pleasant to use. From a presentation standpoint, the gentle feel on first touch also adds credibility to prototypes.
Conversely, TPU struggles with taut, precise geometry. Long bridges, steep overhangs, thin tall decorative features, and tight-tolerance snap fits are better served by PLA or PETG. Support removal can also be difficult -- especially on the underside of broad flat surfaces, where designing for post-processing becomes important. Rather than being universally capable because it is soft, TPU delivers its best value in shapes where flexing itself is the function.
Hardness differences translate directly to use cases. 85A is quite soft, suited to protective cases and cushion parts. 90A works where you need a balance of softness and shape retention -- shoe soles, RC tires. 95A is the easiest all-purpose flexible baseline. 98A leans toward high-load and wear-resistant parts. Softer grades feel nicer to touch but are harder to print, so for practical projects, working backward from "how much softness does this actually need" prevents a lot of failures.
Three Things to Check Before Starting with TPU
Extruder Type Compatibility
The most common first stumble with TPU is the extruder type. Soft filaments behave best in systems where pushing force transmits directly to the nozzle, and direct drive extruders fit that description -- the short path from gear to nozzle keeps the filament responsive. For TPU 95A and similar grades, direct drive is the most approachable starting point. Machines like Bambu Lab's X1 series, which have documented TPU workflows, make entry even easier.
Bowden setups, by contrast, route filament through a longer tube from the extruder to the nozzle. That extra distance gives soft filament room to flex and wander. Minor resistance or play that goes unnoticed with rigid PLA becomes a source of buckling and feed inconsistency with TPU. A Bowden setup is not an automatic disqualifier -- people do print TPU on them -- but dialing in settings is noticeably more demanding than on direct drive, and the gap widens as you move toward softer grades like 85A or 90A.
When I ran 95A through an Ender-series Bowden setup for the first time, I assumed slowing down would be enough. Instead, I got intermittent under-extrusion that felt less like a nozzle issue and more like the filament was struggling partway along the path. Looking closer, I found a small step at the filament guide entrance where the soft material was catching and partially jamming. Adding a spacer to eliminate that step and dropping the speed further brought the feed to a much more stable state. With TPU, it pays to look beyond "direct drive vs. Bowden" and evaluate how clean the path actually is within your setup.
One more thing worth noting: automatic material systems deserve the same scrutiny. Bambu Lab's AMS, for example, is a convenient system, but it is listed as not recommended for very soft filaments, including some TPU. If you are accustomed to multi-color workflows through the AMS, this is easy to overlook. Whether your printer supports TPU and whether the feed system can handle it are two separate questions.
Filament Path Gap Check
When TPU feed problems appear, the temptation is to look at nozzle temperature or retraction. In practice, gaps in the filament path are often the real culprit. What you need to inspect is the continuity from the extruder gear through the guide hole, into the PTFE tube entrance, and down to the heat break. Any step or lateral play in that path gives soft TPU room to veer sideways instead of moving straight forward. Clearances that rigid filament pushes right past become escape routes for TPU.
The most critical spot is the small space immediately after the drive gear. If there is any distance where compressed filament can wander before entering the next guide, it tends to bulge sideways rather than advance. This leads to biting, shaving, and feed slippage. Even a small step matters significantly with TPU -- the path needs to be not just "passable" but "inescapable."
When I jammed 95A on my Ender-series Bowden machine, this was exactly the culprit. A tiny step at the filament guide entrance -- completely harmless with PLA -- caused the TPU tip to deflect and bind. After adding a spacer to align the entrance and reducing speed, the feed felt noticeably lighter and the issue stopped recurring. When TPU gives you "intermittent extrusion failure" or "suddenly goes thin mid-print," check the path before adjusting temperature.
A useful illustration here would be a close-up showing the structural difference between direct drive and Bowden, plus a detail of the gap between gear and guide in a problematic setup. Beginners tend to focus on the nozzle side, but with TPU, the battle is often won or lost at the feed entrance.
Drying Check
With TPU, drying state translates directly to print quality. Printing with absorbed moisture produces surface blemishes, bubbly roughness, and increased stringing, and the surface finish will not stabilize. PLA sometimes muddles through even when slightly damp, but TPU does not forgive that. The material is already challenging to control during extrusion -- add moisture and the adjustment window shrinks dramatically.
Drying conditions vary by product. Common TPU guidelines suggest 50-60C for 4-6 hours, while BigRep TPU 98A specifies 80C for 4-6 hours. Rather than memorizing numbers, the principle is to prioritize the manufacturer's recommended drying conditions for your specific filament. Even with generic TPU 95A, running a drying cycle before printing often reduces stringing and surface roughness in one step.
💡 Tip
With TPU, "dry first, then tune settings" is far more efficient than the reverse. Trying to adjust speed and temperature on damp filament scatters the variables and makes it hard to tell what is actually helping.
I now dry TPU as a default step before any phone case print. With dry filament, the extrusion sounds cleaner, the surface has a visible sheen, and setting changes produce predictable results. Damp spools, on the other hand, add not just stringing but a subtle roughness that drags down the overall finish. If you want TPU to look good, drying matters more than temperature tuning.
Recommended Starting Settings for TPU
Prerequisites
This section focuses on initial values designed to get TPU 95A through a first print successfully. The goal is not fine-tuned quality -- it is completing a short test model without stringing or feed failures stopping you mid-print. 95A is the most approachable hardness in FDM-grade TPU, holding shape better than 85A or 90A, which makes it a solid baseline for a starting recipe.
These settings work across major slicers (Ultimaker Cura, OrcaSlicer, etc.), though menu names and layout vary by version and preset. Refer to your slicer's documentation for exact navigation. The focus here is on what to look for and what each setting affects. These are example starting points -- always prioritize your filament manufacturer's recommended temperature range for nozzle temp and fan speed.
Quick checklist for a first run:
- Drying: 50-60C for 4-6 hours (prioritize your filament's recommendation)
- Nozzle temperature: e.g. 220C (use your filament's recommended range)
- Bed temperature: 50C
- Layer height: 0.2mm
- Perimeter/inner wall/infill speed: 20-25mm/s
- Fan: 70% (adjust within 50-80% depending on material and geometry)
- Retraction: Direct drive -- start around 1.0mm distance / 20mm/s speed; Bowden -- start at 0-1.0mm
After the first print with these values, there are not many knobs to turn. For TPU 95A, adjustments work best in small increments: temperature in 5C steps, retraction distance in 0.5mm steps, speed in 5mm/s steps. A useful mental model is to picture the starting recipe at the center, with arrows pointing to +/-5C for temperature, +/-0.5mm for retraction, and +/-5mm/s for speed.
If stringing is heavy, the straightforward move on direct drive is to increase retraction distance by 0.5mm at a time. If that does not help enough, lower nozzle temperature by 5C. When I was troubleshooting a stringy 95A print, the filament left thin membrane-like threads between travel moves at 220C. Dropping to 215C and bumping retraction from 1.0mm to 1.5mm cleared most of the webbing, and post-processing effort dropped noticeably. With TPU, small increments produce more readable results than large changes.
If you see under-extrusion, go the opposite direction: raise nozzle temperature by 5C first. The urge to increase flow rate immediately is strong, but pushing more volume through insufficiently melted material can make outer walls look thicker without actually filling properly. If walls still look thin or gaps persist, then fine-tune flow rate within the 95-105% range.
Bed temperature targets 40-60C. If adhesion is weak, raise it; if parts stick too hard, lower by 5C at a time or change your adhesion aid or build surface. Fan speed starts at 50-80% mid-range. When layers sag or overhangs collapse, increase fan slightly while dropping speed another notch. TPU does not universally benefit from maximum cooling, but the combination of fan and speed is effective against sagging soft geometry.
On Bowden setups, shift your tuning focus from retraction to speed and path stability. When stringing bothers you, the instinct is to increase retraction distance, but with TPU that frequently triggers jams instead. Restraining retraction is itself a setting decision. Rather than trying to clean everything up at once, establish a stable low-speed orbit first, then tighten temperature, retraction, and flow rate in that order.
How Settings Change Between Direct Drive and Bowden
The Direct Drive Approach
One reason TPU settings trip people up is not the material itself but carrying over the wrong mental model from one extruder type to another. Direct drive systems keep the path from extruder to nozzle short, so the response when pushing soft filament is direct and predictable. Standard grades like TPU 95A print comfortably, and even softer formulations become feasible. The range of "workable" settings is wider, which means finding a good combination takes less iteration.
When I first ran 95A on a Bambu Lab direct drive machine, I honestly expected more of a fight. Instead, it printed more normally than anticipated. Not PLA-level carefree, but feed lag and nozzle-side deflection were minimal, and small adjustments to temperature and retraction brought the surface quality together quickly. Bambu Lab's own TPU guide separates TPU presets from maximum volumetric flow considerations, which underscores how extrusion path stability directly determines how much speed headroom you have.
The key awareness with direct drive is that "easy to run" does not mean "crank up the speed." The wider tolerance for speed and retraction is an advantage, but the tuning goal remains balancing stringing against surface quality. TPU reacts visibly to small parameter shifts -- sheen, edge sharpness, and corner definition all change. On direct drive, small refinements outperform large corrections.
A comparison for quick reference:
| Attribute | Direct Drive | Bowden |
|---|---|---|
| TPU suitability | High | Medium-Low |
| Recommended retraction range | 0.5-2.0mm | Start at 0-1.0mm, adjust carefully |
| Speed tolerance | Relatively wide | Stable at lower speeds |
| Primary failure mode | Stringing tuning shortfall | Buckling, feed inconsistency, jams |
| Beginner-friendliness | High | Somewhat low |
The gap widens with softer TPU. Bambu Lab offers 85A and 90A as separate product lines partly because as softness increases, shape retention and feed behavior change meaningfully. 95A is the accessible starting point, but even there, the direct drive advantage is unmistakable.
The Bowden Approach
With Bowden setups, priorities shift. The longer path to the nozzle means TPU can flex inside the tube, and pushing force does not arrive at the tip as predictably. Applying direct drive logic here -- increasing retraction to fight stringing, or raising speed for a better finish -- tends to produce buckling and feed failures before it produces improvement.
Ender-series machines are the textbook case. Particularly on a classic Ender 3 Bowden configuration, the effective strategy for stabilizing TPU is quite clear. In my own experience, while 95A was cooperative on the Bambu setup, the Ender 3 responded best to locking in low speed and keeping retraction minimal. At 20mm/s with retraction near the minimum, some stringing remained but the print ran to completion without collapsing. That difference is not about tuning skill -- it is about the physics of the extrusion system.
The priority on Bowden is stabilizing the filament path, not chasing cosmetic improvements. High PTFE tube friction, steps at the tube entrance, or play on the extruder side -- these small issues escalate quickly with TPU, causing jams and deflection. Tube and guide upgrades, or modifications to reduce path play, are effective because they address the mechanical root cause. The popularity of direct drive conversions for Ender-series machines is not a trend -- it is an acknowledgment that the physical difference is obvious.
That said, Bowden does not mean TPU is impossible. Some Creality machines explicitly list TPU 95A compatibility, and even older Bowden setups can produce clean prints when you maintain three conditions: low speed, minimal retraction, and a clean path. For relatively simple shapes -- phone cases, leg caps -- avoiding aggressive speed and travel-heavy toolpaths makes a real difference in outcome.
The Bowden mindset is less "dial in the perfect settings" and more "converge on conditions where nothing goes wrong." When you do want to improve finish quality, the first thing to check is not retraction but whether the slicer is generating excessive travel moves, whether the path is flowing smoothly, and whether speed is still reasonable. Following that order eliminates a lot of machine-specific variation.
What Not to Do, by Extruder Type
The common direct drive mistake is getting comfortable with stable extrusion and jumping straight to high speed. Unless you are using a high-speed formulation like TPU 95A-HF, pushing standard TPU the same way produces stringing and soft corners before you get a quality gain. Direct drive has more headroom, but "it extrudes" and "it extrudes well" are different thresholds, and TPU makes the gap obvious.
The Bowden mistake to avoid is increasing retraction to eliminate stringing. This backfires with surprising frequency. Pulling soft filament back through a long tube increases deformation risk inside the path, which leads to jams and feed failures down the line. The reputation Ender 3 has for being "difficult with TPU" largely stems from this cycle of chasing stringing with retraction and creating new problems.
Both types should avoid leaving excessive travel moves unchecked. With TPU, every non-extruding travel move is an opportunity for thin strings to form in mid-air. Models with many small islands, toolpaths that jump to distant positions, and frequent back-and-forth moves are stringing magnets. Rather than trying to eliminate stringing through print settings alone, optimizing travel in the slicer or reorienting the model to shorten travel distances is often more effective.
💡 Tip
Frame TPU tuning as "establish reliable feed first" before "eliminate stringing." Direct drive excels at finish refinement; Bowden needs stability as the foundation.
One more overlooked issue: copying settings between extruder types. Numbers that work well on a Bambu direct drive will not produce similar results on an Ender Bowden. The reverse is also true -- importing conservative Ender settings into a direct drive machine leaves performance on the table. TPU settings are material-specific, yes, but in practice they are better understood as extruder-type-specific recipes.
Choosing by Shore Hardness: 85A, 90A, 95A, and 98A
Hardness Basics and Selection
When in doubt, start with 95A. It is the most common FDM-grade TPU hardness and offers the best balance between flexibility and printability. From there, 90A adds softness, 85A goes significantly softer (and harder to print), while 98A trades flexibility for shape retention and easier handling.
Shore hardness numbers are simple: lower means softer, higher means firmer. With TPU, this directly determines feel, shape retention, and how easily the filament feeds through your printer. Bambu Lab's own comparison between 85A and 90A positions 85A as softer and more impact-absorbing, with 90A offering slightly better shape retention. At 95A, versatility jumps -- phone cases, leg caps, light vibration-dampening parts all work in the "bends enough but does not collapse" range. 98A feels more like a solid rubber component.
The tactile difference becomes obvious when you handle the same model in different hardnesses. Comparing 90A and 95A side by side, 90A gives noticeably when you press the edge -- it has a cushion-like quality. 95A pushes back. Rather than just being soft, it has a springy rebound that 90A lacks. For projects where the finished piece needs to look intentional, that difference in recovery behavior matters more than you might expect.
Summary by hardness:
| Attribute | TPU 85A | TPU 90A | TPU 95A | TPU 98A |
|---|---|---|---|---|
| Softness | Very high | High | Moderate | Firm |
| Shape retention | Low | Medium | Medium-High | High |
| Typical uses | Protective cases, cushion parts | Shoe soles, RC tires, elastic functional parts | General-purpose flexible parts | High-load, wear-resistant parts |
| Print difficulty | High | Somewhat high | Approachable | Relatively easy |
An illustration comparing the tactile feel at each hardness would work well here: 85A as "soft enough to squish with a finger," 95A as "firm rubber," and 98A as "solid, holds its shape." That turns spec numbers into something tangible.
Why 95A Is the Default Recommendation
The reason 95A is the standard starting point is straightforward: you clearly feel the flexibility, yet the material does not fight you during printing. Softer grades like 85A and 90A are interesting, but a first spool needs to do more than just feel nice. It needs to hit dimensional targets, keep outer walls intact, and produce a case or cap that looks clean when finished. 95A delivers that balance better than anything else in the TPU range.
In practice, 95A is remarkably capable as a flexible filament entry point. A phone case in 95A flexes during installation but does not sag when sitting on a table. A desk leg cap cushions the contact point without going limp. Moving down to 90A softens the feel but introduces a "floppy" quality in thin-walled designs that requires rethinking wall thickness and ribbing. At 85A, the effect is even more pronounced.
The selection logic is clear: start with 95A; go to 90A if softness is a priority; reserve 85A for when you specifically need extreme flexibility. 85A is appealing but raises print difficulty, so I think of it as "the hardness you move to after you understand TPU behavior." The finished feel might draw you toward softer options, but 95A as a first purchase minimizes failure and gives you the clearest picture of where to expand next.
💡 Tip
Softer does not automatically mean more premium. On thin-walled parts, 95A often produces tighter edges and a more product-like finish than softer grades.
98A Characteristics and Considerations
98A sits at the firm end of the TPU range. The feel is "flexible material, but noticeably solid" -- more like resilient, tough rubber than something that bends freely. Shape retention is high, which also makes it one of the more stable grades to print. If you want TPU's properties without excessive softness, 98A is a clear choice.
Use cases shift accordingly. Where 95A covers general-purpose flexible parts, 98A targets high-load and wear-resistant applications. Parts that slide or rub, contact surfaces that need to hold firm, and protective components where deformation should be limited all benefit from this hardness. Some products like BigRep TPU 98A also advertise heat resistance up to 100C, making it suitable for applications beyond just mechanical durability.
However, 98A is not an upgrade over 95A -- it is a different tool. If impact absorption or grip comfort is the priority, 85A, 90A, or 95A will serve better. Even for phone cases, if you value corner compliance on drops and easy removal, 95A is the better balance; 98A will feel noticeably rigid. Tactilely, if 95A is firm rubber, 98A is a step firmer still.
On the handling side, "harder means less fussy" is not quite right -- harder means more predictable characteristics is closer. BigRep TPU 98A specifies drying at 80C for 4-6 hours, which is higher than typical TPU. When choosing 98A, the clearest framing is: pick it when shape retention, wear resistance, or heat tolerance matters more than softness.
Drying and Storage: Building a Pre- and Post-Print Routine
Drying Guidelines and Exceptions
With TPU, suspect moisture before you suspect settings. Chasing surface roughness or stringing through temperature adjustments alone will not work if the filament has absorbed water. In my own testing, printing with undried TPU 95A on a humid day produced a finely textured outer wall with a slightly clouded appearance. After running the same spool through a filament dryer at 60C for about 5 hours and reprinting the identical model, the surface calmed down noticeably -- smooth enough to feel "softly slick" under the fingertip. The difference with TPU is substantial.
As a baseline, common TPU drying recommendations are 50-60C for 4-6 hours, consistent with guidance from sources like SK Honpo. Higher-hardness products can require different conditions -- BigRep TPU 98A calls for 80C for 4-6 hours. The important takeaway is not to treat all TPU identically. Generic 95A and high-hardness 98A can have different drying requirements. Rather than memorizing numbers, default to the manufacturer's recommendation and you will stay on track.
From an appearance standpoint, drying delivers visible results. TPU's softness means even slight surface roughness can make a print look sloppy. Properly dried material, by contrast, produces edges that look tighter, and for hand-held parts like phone cases or caps, the finish quality jumps noticeably. A before-and-after surface comparison photo would communicate this powerfully -- roughness versus recovered sheen on the same model.
Recognizing Moisture Symptoms
TPU moisture shows up clearly in both sound and appearance. The most telltale sign is a popping or crackling sound during extrusion -- small snapping noises near the nozzle that persist through the print. Visually, look for surface blemishes (small bumps), a bubbly or pockmarked texture, and on light or translucent colors, a whitish haze or clouding.
Some symptoms only become apparent after the print finishes. The outer wall may look slightly powdery, or layers that appeared to stack fine during printing feel weak at the layer interfaces when you flex the part. TPU normally has a tenacious, stretchy quality when bent. If it feels dry and crumbly, or layers separate too easily, moisture is worth investigating.
When multiple symptoms appear together, adjusting temperature or retraction is likely wasted effort. Re-dry the filament first, then reassess. I have caught myself more than once tuning settings for what I thought was a stringing issue, only to discover moisture was the primary cause. TPU's moisture symptoms and setting problems can look similar, so evaluating sound, surface texture, and layer adhesion as a set makes diagnosis more reliable.
💡 Tip
Quiet extrusion, reduced surface roughness, and restored flexibility when bending. When all three are present after drying, you can be confident it worked.
Storage and Workflow Habits
Dried TPU needs proper storage to stay that way. The basic approach is a sealed container with desiccant. This is not just for long-term storage between projects -- treating your default spool location as sealed storage keeps subsequent prints starting clean. Leaving a spool on the desk in open air may not look different, but extrusion quality degrades gradually.
For active printing, feeding directly from a dry box is highly compatible with TPU. Even after a thorough drying cycle, extended exposure to ambient air during a long print can shift material behavior. A setup where the storage container doubles as the feed source eliminates that variable. The benefit is most noticeable on cases and thin-walled parts where surface uniformity is the priority. A diagram showing dry box placement -- spool, desiccant, feed port, and printer positioning -- would lower the setup barrier considerably.
The daily workflow is simple: check filament condition before printing, dry if anything seems off, and return to sealed storage immediately after. Locking in this routine dramatically reduces the "TPU is being difficult today" experience. I prioritize appearance in my projects, so my honest assessment is that storage discipline matters more than settings refinement for TPU. Especially with a staple like 95A, proper drying and storage alone let the material's strengths come through consistently.
Common TPU Failures and How to Fix Them
Stringing
Stringing -- thin threads stretching between travel moves -- is the most frequent TPU issue. Rather than dismissing it as "soft material, unavoidable," separating it into too-high temperature, absorbed moisture, and incorrect retraction and addressing them in order gets results faster. Jumping straight to aggressive retraction often destabilizes the feed and creates a different problem.
Start by lowering nozzle temperature in 5C steps. Higher temperatures make TPU more prone to oozing, and strings form during every travel move. If stringing persists, adjust retraction in 0.5mm increments. The practical working range for TPU is roughly 0.5-2.0mm, but more retraction does not always mean less stringing. On Bowden systems especially, extra retraction destabilizes the filament in the tube and can come back as jams or feed inconsistency.
Stringing and moisture-related roughness frequently appear together. If stringing suddenly increases, re-dry the filament before changing settings. Dry TPU behaves noticeably more predictably at the same temperature. As a diagnostic sequence: lower temperature slightly, adjust retraction slightly, re-dry and retest. That order keeps variables manageable.
Surface Blemishes, Bubbles, and Whitening
Fine bumps across the surface, bubbly extrusion marks, and a whitish haze on translucent or light-colored prints. This is moisture with high probability. TPU makes absorbed water visible on the surface, and symptoms that look like temperature or flow problems are frequently moisture-driven at the root. On parts where surface uniformity matters -- phone cases, covers -- these defects directly impact the finished quality.
In one case, I reprinted a blemished model after re-drying at 60C for 6 hours without changing any print settings. The outer wall texture improved dramatically, becoming much more uniform. The improvement was large enough without any setting changes that moisture was clearly the dominant factor. The whitish cloudiness also receded, returning to TPU's characteristic slightly matte, smooth appearance.
The fix is straightforward: re-dry, improve storage, and dry the feed path. Even if the spool was dried before printing, leaving it exposed to air during a long print can reintroduce enough moisture to shift surface quality. Moving to sealed-box feeding addresses that. When symptoms are blemish-type, return to moisture management before touching settings.
Jams and Buckling
Sudden extrusion stop at the nozzle, filament bending or kinking near the extruder, extrusion motor running but print going thin. Jams and buckling in TPU typically stem from excessive speed and filament path problems. Soft material that cannot push through redirects force into path gaps, bulging sideways and buckling.
The highest-impact fix is dropping print speed to 15-20mm/s. Even if you are already printing slowly, reducing speed further at the point where jams occur can restore stability. Bowden setups are especially sensitive -- any play in the long path gives TPU room to misbehave, and "slightly too fast" converts to failure more readily than on direct drive. The reason low speed is so effective on Ender-series Bowden machines printing TPU 95A comes down to this extrusion compliance.
The other culprits are path gaps and excessive retraction. If the guide from extruder to hot end has loose spots, soft TPU will escape sideways immediately. Reducing the space where filament can flex between the drive gear and the heat break often helps. Retraction works the same way -- too much pull-back stresses the material and promotes buckling. For jam-type symptoms, reducing retraction is frequently the productive direction.
Overhang Roughness
Rough underside on overhangs, rounded corners where sharp edges should be, surfaces riding on supports that are suddenly much rougher. TPU's overhang issues come from material sag combined with insufficient cooling. Angles that PLA handles comfortably push TPU past its ability to bridge cleanly, because the extruded bead droops under its own weight.
The baseline fix is slower speed plus higher fan. TPU stays soft longer at printing temperature, so overhanging sections benefit from a "place and solidify" approach rather than continuous fast deposition. When the underside remains rough despite those adjustments, design-side changes -- adding fillets, reorienting the model, or splitting the part -- typically deliver a larger improvement than any further setting change. For product-oriented small items, that one design tweak often jumps quality dramatically.
TPU also has support-specific quirks. Thin support contacts are prone to detaching mid-print, while broad flat interfaces tend to bond too strongly for clean removal. In other words, too little contact and the support fails; too much and it will not come off. When contact conditions need careful control, specialized materials like PolyDissolve S1 water-soluble support can make a meaningful difference. A brief soak lets the support separate gently, which changes the post-processing experience substantially for TPU compared to mechanical removal.
For quick reference, a symptom-to-fix table is practical here. Stringing: lower temperature 5C, adjust retraction 0.5mm. Blemishes/bubbles/whitening: re-dry, improve feed path dryness. Jams/buckling: drop to 15-20mm/s, reduce retraction. Overhang roughness: slow down, increase fan, reconsider supports and geometry.
Supports, Bridges, and Design Tricks to Reduce Failures
Support Marks and Removal
TPU produces rougher contact surfaces from support than PLA under identical settings. On visible undersides or case interiors where fingers touch, support marks directly diminish the finished quality. The complicating factor is that "weaker contact = easier removal" does not hold universally. Too-weak support detaches mid-print, while broad flat support bonds like a solid plate and resists clean separation.
When I printed a part with a large top surface, the support adhered across the entire face and came off as a single plate rather than crumbling away -- extremely difficult to remove, and it pulled the surface rough in the process. After adjusting to 0.3mm contact distance with 4 interface layers, the same broad contact area released cleanly in some cases. Not a universal magic number, but a worthwhile starting point when TPU support removal is giving you trouble.
These settings work because TPU's softness means it conforms to the interface -- and the interface design determines where the separation boundary forms. Too few interface layers and the support basically fuses with the part. Enough interface layers and the support side peels away as a unit. For appearance-critical parts, think less about support density and more about where the separation plane sits.
Support quality is better evaluated by contact spread than by contact point count. Broad, flush contact provides strong support but fights removal. Contact that approximates lines or points separates more cleanly, even with TPU.
Water-Soluble Support: Thinking in Terms of Compatibility
When support removal becomes a persistent struggle, switching to dedicated support material can be faster than endless setting adjustments. PolyDissolve S1 water-soluble support, for example, is noted for good compatibility with TPU 95-series filaments. For complex cavities or internal features where tools cannot reach, the ability to dissolve away support opens up design freedom that TPU-only support cannot match.
TPU's softness means that forcing support off mechanically can flex the part itself, and instead of a clean break you get surface damage. Water-soluble support eliminates that force. PolyDissolve S1 documentation suggests roughly 2-3 minutes of water immersion to initiate separation in some cases. Just being able to start the peel without prying makes a tangible difference with TPU.
The practical consideration is system-level compatibility, not just the support material itself. Extrusion temperature ranges, filament path requirements, and multi-material feed constraints all need to align. On systems like Bambu Lab's AMS where soft material transport is restricted, both the TPU and the support material need to feed reliably. TPU 95A + PolyDissolve S1 is a promising pairing, but verify against your specific machine's compatibility documentation.
For appearance-focused projects, engineering the removal process to be gentle -- rather than scraping and cutting aggressively -- protects the final surface. Especially on parts with internal geometry where tools cannot reach, water-soluble support is less a luxury and more a practical necessity for a clean result.
Design-Side Avoidance Techniques
Here is something that does not get enough attention: TPU support problems frequently shrink dramatically with geometry changes rather than setting changes. Before I start fine-tuning support parameters on a prototype, I ask whether the part can be shaped so broad surfaces do not contact support in the first place. That often fixes removal and appearance simultaneously.
Pocket features (material removal from flat undersides) are highly effective. Breaking up a continuous flat bottom into smaller contact zones prevents support from bonding as a monolithic plate, making separation much easier.
Chamfers and fillets also help significantly. A 90-degree overhang sits directly on support with full face contact, leaving a strong mark on TPU. Even a small chamfer changes how support engages the edge, and underside roughness can drop noticeably. The added visual softness also works well for product-oriented parts, so structural benefit and aesthetics align.
Splitting the model to shorten bridges is another powerful tool. Rather than printing a single large top surface that demands extensive support, dividing the part and assembling afterward can eliminate whole support regions. TPU's adhesion and fit behavior add complexity to splits, so this is not universally applicable, but for protecting visible surfaces it is very effective.
Finally, tilting flat surfaces even slightly can transform support behavior. A perfectly horizontal face makes full contact with support across its entire area, but adding a few degrees of angle turns that contact toward lines. Sometimes this change alone makes removal dramatically easier, without touching a single print setting. With TPU, designing not just "does this need support" but "how will support touch this" prevents a lot of problems before they start.
In cross-section, the failure case is a broad horizontal surface with support bonded flat underneath. The success case adds chamfers, tilts, or pocket features to break up that contact. Diagrams would communicate this far better than text alone. Photos of the water-soak-then-peel workflow would also help, showing readers how far to let the material soften before starting removal.
When in Doubt: Conclusions for TPU Beginners
The Shortest Path to Getting Started
If you are unsure, your first spool should be TPU 95A. The softer 85A and 90A are interesting materials but make first-time success harder. Start with 95A to understand how the material behaves, always dry before printing, and use low speed with minimal retraction as your baseline. This makes failure causes much easier to trace.
On the printer side, direct drive is the easier starting point, but Bowden can absolutely work. The difference is "how carefully you need to begin," not "whether it is possible." Bowden setups have a longer filament path where soft material can misbehave, so checking for unnecessary gaps in the path before your first print saves a lot of downstream adjustment. Automatic feed systems like the AMS may not support soft filaments -- check your hardware's compatibility documentation before routing TPU through them.
I did not jump straight to a phone case for my first TPU print. Instead, I started with a small cable bushing -- simple geometry, quick feedback. Getting a feel for stringing behavior and extrusion characteristics on a small part made the transition to larger prints feel remarkably smoother.
Setting Adjustment Priority
The starting philosophy is simple: dry 95A, slow speed, conservative retraction. Temperature begins around 220C, with adjustments in +/-5C increments. Retraction starts at roughly 1.0mm on direct drive and 0-1.0mm on Bowden.
Fix the adjustment sequence and decision-making becomes easier. Persistent stringing? Do not change everything at once. Check drying, lower temperature slightly, add a bit of retraction -- in that order. Jams or buckling? The first suspects are excessive speed and filament path issues. TPU responds better to isolating variables one at a time than to sweeping changes.
Complex geometry can wait until settings are stable. Keep initial tests to small, simple shapes so you can focus on temperature, feed, and stringing without other variables complicating the picture.
Next Steps
From here, the action items are few:
- Confirm whether your printer uses direct drive or Bowden
- Get TPU 95A and dry it before use
- Print a small, simple test model and adjust incrementally based on symptoms
TPU is not a material that rewards trying to nail the perfect print on the first attempt. It is a material that becomes surprisingly manageable once you pick the right entry point. Start with dry 95A at low speed, verify your filament path is clean, and build from there. Internal links to related guides (e.g., slicer settings, material storage) will be added to this article's "Next Steps" section and throughout once those articles are published (currently on hold pending companion content).
Related Articles
How to Choose 3D Printer Filament: Materials Compared by Use Case
Choosing the right 3D printer filament works best when you focus on what you're making and what your printer can handle, rather than memorizing material names. I started with PLA and had great early wins printing small items -- until a print warped in a hot car, and I realized practical parts demand a different approach.
10 Best PLA Filaments for Beginners: How to Choose Your First Spool
If you are buying PLA filament for the first time, starting with a standard PLA or a well-documented PLA+ is the fastest path to good prints. Confirm whether your printer uses 1.75mm or 2.85mm filament before ordering to avoid a costly mistake.
PLA vs ABS vs PETG — Comparing the Big Three Filaments
If you're stuck choosing between common 3D printing filaments, a simple rule works well: start with PLA, move to PETG for functional parts, and reach for ABS when heat resistance matters. This guide breaks down the real-world differences between PLA, PETG, and ABS for anyone picking their first filament or trying to stop reprinting parts made from the wrong material.
PETG Filament Guide: Properties, Settings, and Tips for Your First Roll
Outgrown PLA but not ready for the demands of ABS? PETG sits right in between -- tougher, more heat-resistant, and still approachable. After switching car-dashboard trinkets from PLA to PETG, the difference in rigidity at 70-80 C (~158-176 F) was immediately obvious.