Flexible filaments and TPU: Shore hardness, extruder type and feeding limits

Shore hardness: what the number on the spool means

Flexible filaments are rated on the Shore hardness scales. Bambu Lab’s TPU guide explains the convention: Shore A covers softer materials and Shore D harder ones, and a higher number means a harder, less flexible filament. Its ordering runs from 77D at the hard end through 95A, 90A and 85A down to 70A ^[3].

Hardness decides how easy the filament is to push. Prusa’s flexible-materials article says the softer the filament, the harder it is to print ^[1]. Bambu Lab sets a floor: it recommends 85A or harder, explains that softer TPU bends during feeding and causes buckling and tangling, and says its printers do not currently support TPU below 85A ^[3].

The rating on the label is a product name, and measured values can differ. Polymaker reports 95A for PolyFlex TPU95 under ISO 7619-1, while Prusa Polymers’ data sheet for Prusament TPU 95A lists a measured Shore A of 92 (and Shore D of 39) under its own method ^{[7, 9]}. Polymaker describes its TPU95-HF as closer to 98A than to 95A ^[8].

Extruder type and the feed path

The distance between the drive gears and the nozzle is the main hardware factor. Polymaker’s wiki says the further that gap, the more difficult TPU becomes and the slower it must print ^[8]. For its softer TPU90 it strongly recommends a direct extruder rather than a Bowden setup; TPU95 prints on most printers but runs faster on direct drive ^[8].

Extruder tension matters as well. Prusa recommends loosening the idler so the gears push the filament only gently, because high pressure can tangle flexible filament in the gears ^[1]. Prusa’s TPU 95A guide lists 1.5 turns looser on the idler screws as the fix when filament tangles ^[2].

Nozzle choice is part of the feed path on Bambu Lab printers. The TPU guide says 0.2 mm nozzles are not recommended for TPU, that 85A TPU cannot be printed with a 0.4 mm nozzle, and that a nozzle previously used with carbon- or glass-fiber filament can leave residue that increases resistance, so several cold pulls are advised first ^[3].

AMS, MMU and filament sensors

Automatic material systems push filament through long tubes and gears, and filament that is too soft cannot pass smoothly through them ^[4].

• Bambu Lab AMS family. The material table lists TPU as not compatible with the AMS and AMS lite and names soft TPU among filaments that cause loading failures and clogs ^[4]. The TPU guide adds that the AMS 2 Pro is also not compatible, and that the AMS HT can hold TPU only as a sealed drying box, feeding through a dedicated TPU outlet with automatic feeding unused ^[3]. Bambu’s TPU 95A HF product page states that all AMS series are not compatible ^[6].
• TPU for AMS. Bambu Lab sells a TPU grade that the guide says loads and prints through the standard feeding of all AMS models ^[3].
• Prusa printers. The flexible-materials article warns that automatic filament change might not work properly with flexible filaments ^[1]. On the Original Prusa XL, Prusa suggests bypassing the side filament sensor and feeding TPU directly into the tube; the CORE One+ and CORE One L have a latch that loosens the side sensor lever for flexible filament ^{[1, 2]}.

Bambu Lab also asks you to turn off flow dynamics calibration before printing TPU ^[3].

Speed and volumetric flow

Standard TPU prints slowly. Prusa gives about 20 mm/s as a typical speed and 30–40 mm/s as the maximum; faster printing risks clogs or filament wrapping around the gears ^[1]. Prusament TPU 95A lists up to 30 mm/s, and PolyFlex TPU95 lists 30–50 mm/s for a 0.4 mm nozzle ^{[9, 7]}.

High-flow TPU grades change those numbers. Bambu Lab’s product page compares its TPU 95A HF at 12 mm³/s (147 mm/s) with its standard TPU 95A at 3.6 mm³/s (44 mm/s), both at a 230 °C nozzle ^[6]. Polymaker suggests 40–100 mm/s for TPU95-HF, slower on long feed paths ^[8]. The volumetric speed calculator converts between mm³/s and mm/s for a given layer height and line width.

Retraction, stringing and moisture

Retraction recommendations range widely. Prusa suggests lower retraction to reduce clogs and tangling, and says it is fine to start with retraction off and raise it while watching the result ^[1]. Bambu Lab’s TPU 95A HF data sheet lists 0.8–1.4 mm, while Polymaker’s TPU95 data sheet lists 3–6 mm at 40–60 mm/s ^{[5, 7]}. OrcaSlicer’s retraction tower defaults to 0–2 mm for direct drive and suggests 1–6 mm for Bowden ^[10]; see the retraction calibration.

For oozing, Bambu Lab’s guide lists four fixes: print one object at a time to cut travel moves, lower the nozzle temperature by about 5 °C and the max volumetric speed by 0.4 mm³/s, increase retraction by 0.2 mm, and dry the filament ^[3]. Polymaker warns that TPU strings by nature and that removing every hair is difficult ^[8].

Moisture guidance differs between makers. Bambu Lab calls TPU highly hygroscopic and recommends storage below 20% RH ^[3]. Its material table lists drying as required ^[4]. Prusa says Prusament TPU 95A usually does not need drying unless it was stored poorly ^[2]. Bubbles and stringing are the symptoms both name; see wet filament.

Plates, cooling and supports

TPU generally adheres strongly to the build plate ^[3]. Prusa says a powder-coated (textured) sheet needs no separation layer, while smooth or satin sheets need glue stick or Kapton tape ^[1]. Its TPU 95A guide marks the smooth PEI sheet as not recommended without glue, because removal can damage it ^[2]. Bambu Lab says every plate except the Cool Plate SuperTack works, and suggests isopropyl alcohol in the gap to peel stuck parts ^[3].

Supports are difficult: Polymaker says TPU’s strong layer adhesion makes same-material supports very hard to remove, and suggests designing parts to avoid them ^[8]. Prusa recommends a 0.3 mm Z gap and an XY separation of at least 100% ^[1]. Bambu Lab recommends PLA as the support material under TPU ^[3].